Home

PGI User's Guide

1. Fortran Type C C Type Size bytes CHARACTER n x char x n n REAL x float x 4 REAL 4 x float x 4 REAL 8 x double x 8 DOUBLE PRECISION x double x 8 INTEGER x int x 4 INTEGER 1 x signed char x 1 INTEGER 2 x short x 2 INTEGER 4 x int x 4 INTEGER 8 x long long x 8 LOGICAL x int x 4 LOGICAL 1 x char x 1 LOGICAL 2 x short x 2 LOGICAL 4 x int x 4 LOGICAL 8 x long long x 8 Table A 18 Fortran and C C Representation of the COMPLEX Type Fortran Type C C Type Size bytes COMPLEX x struct float r I x 8 COMPLEX 8 x struct float r I x 8 COMPLEX 16 x struct double dr di x 16 DOUBLE COMPLEX x struct double dr di x 16 C C arrays and Fortran arrays use different default initial array index values By default C C Run time Environment 243 arrays start at 0 and Fortran arrays start at 1 A Fortran array can be declared to start at zero Another difference between Fortran and C C arrays is the storage method used Fortran uses column major order and C C use row major order For one dimensional arrays this poses no problems For two dimensional arrays where there are an equal number of rows and columns row and column indexes can simply be reversed Inter language function mixing is not recommended for arrays other than single dimensional arrays and square two dimensional arrays Structures Unions Maps and Derived Types Fields within
2. rcx i xmm0 lt ignored gt 32 11 rdx lt ignored gt xmml f 40 1 r8 s1 i s1 f xmm2 lt ignored gt r9 lt ignored gt xmm3 d Implementing a Stack In general compilers and programmers must maintain a software stack The stack pointer register Yorsp is set by the operating system for the application when the program is started The stack must grow downwards from high addresses A separate frame pointer enables calls to routines that change the stack pointer to allocate space on the stack at run time e g alloca Some languages can also return values from a routine allocated on stack space below the original top of stack pointer Such a routine prevents the calling function from using Yorsp relative addressing to get at values on the stack If the compiler does not call routines that leave rsp in an altered state when they return a frame pointer is not needed and is not used if the compiler option Mnoframe is specified Run time Environment 237 The stack must always be 16 byte aligned except within the prolog and within leaf functions Variable Length Parameter Lists Parameter passing in registers can handle a variable number of parameters The C language uses a special method to access variable count parameters The stdarg h and varargs h files define several functions to access these parameters A C routine with variable parameters must use the va_start macro to set up a data structure bef
3. Creates a relocatable object file Command Line Options 41 Option Description r4 and r8 Interpret DOUBLE PRECISION variables as REAL r4 and r8 Interpret REAL variables as DOUBLE PRECISION rc file Specifies the name of the driver s startup file S Stops after the compiling phase and saves the assembly language code in filename s s Strips the symbol table information from the object file shared Linux only Passed to the linker Instructs the linker to generate a shared object file Implies fpic show Display driver s configuration parameters after startup silent Do not print warning messages time Print execution times for the various compilation steps tp lt target gt Specify the type of the target processor U lt symbol gt Undefine a preprocessor macro u lt symbol gt Initializes the symbol table with lt symbol gt which is undefined for the linker An undefined symbol triggers loading of the first member of an archive library V release_number Displays the version messages and other information or allows invo cation of a version of the compiler other than the default V Displays the compiler assembler and linker phase invocations W Passes arguments to a specific phase W Do not print warning messages There are a large number of compiler options specific to the PGCC and PGC compilers especially PGC
4. pragma global novector float a 100 100 b 100 100 c 100 100 int time maxtime n i j maxtime 10 n 100 for time 0 time lt maxtime time for j 0 j lt n j for i 0 i lt n i c i j a i j b i j Optimization Directives and Pragmas 155 In this version the compiler does not perform vectorization for the entire source file Another use of the pragma scoping mechanism turns an option on or off locally either for a specific procedure or for a specific loop The following example shows the use of a loop scoped pragma main float a 100 100 b 100 100 c 100 100 int time maxtime n i j maxtime 10 n 100 pragma loop novector for time 0 time lt maxtime time for j 0 j lt n j for i 0 i lt n i c i j a i j b i j Loop level scoping does not apply to nested loops That is the pragma only applies to the following loop In this example the pragma turns off vector transformations for the top level loop If the outer loop were a timing loop this would be a practical use for a loop scoped pragma The following example shows routine pragma scope include math h func1 pragma routine novector float a 100 100 b 100 100 float c 100 100 d 100 100 int i j for i 0 1 lt 100 i for j 0 j lt 100 4 alil j alil j bii IJTI celil jl efil J c i j b i j d i j func2 float a 200 200
5. This many array alignments were propagated by interprocedural analysis I318 IPA routine distribution formats propagated This many array distribution formats were propagated by interprocedural analysis I319 IPA routine distribution targets propagated This many array distribution targets were propagated by interprocedural analysis I320 IPA routine common blocks optimized This many mapped common blocks were optimized by interprocedural analysis 276 Messages I321 IPA routine common blocks not optimized This many mapped common blocks were not optimized by interprocedural analysis either because they were declared differently in different routines or they did not appear in the main program I322 IPA analyzing main program Interprocedural analysis is building the call graph and propagating information with the named main program I323 IPA collecting information for Interprocedural analysis is saving information for the current subprogram for subsequent analysis and propagation W324 IPA file appears to be out of date W325 IPA file is for wrong subprogram W326 Unable to open file to propagate IPA information to I327 IPA subprograms analyzed I328 IPA dummy arguments replaced by constants I329 IPA INTENT IN dummy arguments should be INTENT INOUT I330 IPA dummy arguments changed to INTENT IN I331 IPA inherited array alignments replaced I332 IPA t
6. filename F filename FOR filename F95 filename f90 filename f95 filename hpf filename c filename i filename C filename cc filename s filename o filename obj filename a filename lib filename so filename dll indicates a Fortran source file that can contain macros and preprocessor directives to be preprocessed indicates a Fortran source file that can contain macros and preprocessor directives to be preprocessed indicates a Fortran 90 95 source file that can contain macros and preprocessor directives to be preprocessed indicates a Fortran 90 95 source file that is in freeform format indicates a Fortran 90 95 source file that is in freeform format indicates an HPF source file indicates a C source file that can contain macros and preprocessor directives to be preprocessed indicates a pre processed C or C source file indicates a C source file that can contain macros and preprocessor directives to be preprocessed indicates a C source file that can contain macros and preprocessor directives to be preprocessed indicates an assembly language file Linux systems only indicates an object file Windows systems only indicates an object file Linux systems only indicates a library of object files Windows systems only indicates a library of object files Linux systems only indicates a library of shared object files Windows systems only indicates a library of shared obje
7. suffix lt suff gt use lt suff gt as the suffix of the output file containing makefile dependencies Windows only link with the DLL versions of the runtime libraries This flag is required when linking with any DLL built by any of The Portland Group compilers match the behavior of certain gcc bugs Minfo option option 80 instructs the compiler to produce information on standard error where option is one of the following all instructs the compiler to produce all available Minfo information inline instructs the compiler to display information about extracted or inlined functions This option is not useful without either the Mextract or Minline option ipa instructs the compiler to display information about interprocedural optimizations loop instructs the compiler to display information about loops such as information on vectorization Command Line Options opt mp time unroll instructs the compiler to display information about optimization instructs the compiler to display information about parallelization instructs the compiler to display compilation Statistics instructs the compiler to display information about loop unrolling Mneginfo option option neginfo instructs the compiler to produce information on standard error where option is one of the following Minform level Mkeepasm Command Line Options all concur loop instructs the comp
8. tp flags Displays driver options on the standard output Use this option with v to list options that are recognized and ignored as well as the valid options Cross reference D v fpic Linux only Generate position independent code suitable for inclusion in shared object dynamically linked library files Cross reference shared G R fPIC Linux only Equivalent to fpic Provided for compatibility with other compilers Cross reference fpic shared G R Command Line Options 49 G Linux only Passed to the linker Instructs the linker to produce a shared object file Cross reference fpic shared R The g option instructs the compiler to include symbolic debugging information in the object module Debuggers such as PGDBG require symbolic debugging information in the object module to display and manipulate program variables and source code Note that including symbolic debugging information increases the size of the object module If you specify the g option on the command line the compiler sets the optimization level to O0 zero unless you specify the O option For more information on the interaction between the g and O options see the O entry Symbolic debugging may give confusing results if an optimization level other than zero is selected Default The compiler does not put debugging information into the object module Usage In the follo
9. 108 OpenMP Directives for Fortran region redundantly These threads share work by means of work sharing constructs such as parallel DO loops see below The number of threads in the team is controlled by the OMP_NUM_THREADS environment variable If OMP_NUM_THREADS is not defined the program will execute parallel regions using only one processor Branching into or out of a parallel region is not supported All other shared memory parallelization directives must occur within the scope of a parallel region Nested PARALLEL END PARALLEL directive pairs are not supported and are ignored The END PARALLEL directive denotes the end of the parallel region and is an implicit barrier When all threads have completed execution of the parallel region a single thread resumes execution of the statements that follow NOTE By default there is no work distribution in a parallel region Each active thread executes the entire region redundantly until it encounters a directive that specifies work distribution For work distribution see the DO PARALLEL DO or DOACROSS directives PROGRAM WHICH PROCESSOR AN I INTEGER A 0 1 INTEGER omp_get_thread_num A 0 1 A 1 1 SOMP PARALLEL A omp_get_thread_num omp_get_thread_num SOMP END PARALLEL PRINT A 0O A 0 A 1 A 1 END The variables specified in a PRIVATE list are private to each thread in a team In effect the compiler creates a separate copy of each of these vari
10. A 3 4 2 Fortran Naming Conventions By default all globally visible Fortran symbol names subroutines functions common blocks are converted to lower case In addition an underscore is appended to Fortran global names to distinguish the Fortran name space from the C C name space A 3 4 3 Fortran Argument Passing and Return Conventions Arguments are passed by reference i e the address of the argument is passed rather than the argument itself In contrast C C arguments are passed by value When passing an argument declared as Fortran type CHARACTER an argument representing the length of the CHARACTER argument is also passed to the function This length argument is a four byte integer passed by value and is passed at the end of the parameter list following the other formal arguments A length argument is passed for each CHARACTER argument the length arguments are passed in the same order as their respective CHARACTER arguments A Fortran function returning a value of type CHARACTER adds two arguments to the beginning of its argument list The first additional argument is the address of the area created by the caller for the return value the second additional argument is the length of the return value If a Fortran function is declared to return a character value of constant length for example CHARACTER 4 FUNCTION CHF the second extra parameter representing the length of the return value must still be supplied A Fortran comp
11. Do not print warning messages Usage In the following example the driver does not display warning messages pgf95 silent myprog f Cross reference v V w time Print execution times for various compilation steps Usage In the following example pgf95 prints the execution times for the various compilation steps pgf95 time myprog f Cross reference tp Set the target architecture By default the PGI compilers produce code specifically targeted to the type of processor on which the compilation is performed In particular the default is to use all supported instructions wherever possible when compiling on a given system As a result executables created on a given system may not be useable on previous generation systems for example executables created on a Pentium 4 may fail to execute on a Pentium III or Pentium II Processor specific optimizations can be specified or limited explicitly by using the tp option In this way it is possible to create executables that are usable on previous generation systems With the exception of k8 64 k8 64e p7 64 and x64 any of these sub options are valid on any x86 or x64 processor based system The k8 64 k8 64e p7 64 and x64 options are valid only on x64 processor based systems The tp x64 option is used to generate unified binary object and executable files The mm k8 64 and tp k8 64e options result in generation of code supported on and optimized for AMD x64 proc
12. MPSTKZ 8M export MPSTKZ in bash sh or ksh If your program is still failing you may be encountering the hard 8 MB limit on main process stack sizes in Linux You can work around the problem by issuing the command limit stacksize unlimited in csh or ulimit s unlimited in bash sh or ksh Getting Started 9 1 6 Using the PGI Compilers on Windows 1 6 1 BASH Shell Environment On Windows platforms the tools that ship with the PGI Workstation or PGI Server command level compilers include a full featured shell command environment After installation you should have a PGI icon on your Windows desktop Double left click on this icon to cause an instance of the BASH command shell to appear on your screen Working within BASH is very much like working within the sh or ksh shells on a Linux system but in addition BASH has a command history feature similar to csh and several other unique features Shell programming is fully supported A complete BASH User s Guide is available through the PGI online manual set Select PGI Workstation under Start gt Programs and double left click on the documentation icon to see the online manual set You must have a web browser installed on your system in order to read the online manuals The BASH shell window is pre initialized for usage of the PGI compilers and tools so there is no need to set environment variables or modify your command path when the command window comes up In addition t
13. PGIS hello hello PGIS 1 2 1 Command line Syntax The command line syntax using pgf95 as an example is pgf95 options path filename Where options is one or more command line options all of which are described in detail in Chapter 3 Command Line Options Case is significant for options and their arguments The compiler drivers recognize characters preceded by a hyphen as command line options For example the Mlist option specifies that the compiler creates a listing file in the text of this manual we show command line options using a dash instead of a hyphen for example Mlist In addition the pgCC command recognizes a group of characters preceded by a plus sign as command line options The order of options and the filename is not fixed That is you can place options before and after the filename argument on the command line However the placement of some options is significant for example the 1 option 2 Getting Started Note If two or more options contradict each other the last one in the command line takes precedence path is the pathname to the directory containing the file named by filename If you do not specify path for a filename the compiler uses the current directory You must specify path separately for each filename not in the current directory filename is the name of a source file assembly language file object file or library to be processed by the compilation syste
14. The Fortran directive which tells the compiler that for a given loop the last value computed for all scalars make it safe to parallelize the loop is cpgi l safe_lastval In addition a command line option Msafe_lastval provides this information for all loops within the routines being compiled essentially providing global scope 2 7 Inter Procedural Analysis and Optimization using Mipa The PGI Fortran C and C compilers use interprocedural analysis IPA that results in minimal changes to makefiles and the standard edit build run application development cycle Other than adding Mipa to the command line no other changes are required For reference and background the process of building a program without IPA is described below followed by the minor modifications required to use IPA with the PGI compilers While the PGCC compiler is used here to show how IPA works similar capabilities apply to each of the PGI Fortran C and C compilers Note that the examples use Linux file naming conventions On Windows o files would be obt files and a out files would be exe files 2 7 1 Building a Program Without IPA Single Step Using the PGCC command level compiler driver three for example source files can be compiled and linked into a single executable with one command o pgcc o a out filel c file2 c file3 c In actuality the pgcc driver executes several steps to produce the assembly code and object file
15. and math libraries 8 1 Using builtin Math Functions in C C The name of the math header file is math h Include the math header file in all of your source files that use a math library routine as in the following example which calculates the inverse cosine of p 3 include lt math h gt define PI 3 1415926535 main double x y x PI 3 0 y acos x Including math h will cause PGCC C and C to use builtin functions which are much more efficient than library calls In particular the following intrinsics calls will be processed using builtins if you include math h atan2 cos fabs exp atan sin log pow tan sqrt log10 8 2 Creating and Using Shared Object Files on Linux All of the PGI Fortran C and C compilers support creation of shared object files Unlike statically linked object and library files shared object files link and resolve references with an executable at runtime via a dynamic linker supplied with your operating system The PGI compilers must generate position independent code to support creation of shared objects by the linker This is not the default use the steps that follow to create object files with position independent code and shared object files that are to include them The following steps describe how to create and use a shared object file 159 160 To create an object file with position independent code compile it with the appropriate PGI compiler using the fpic option the fPIC
16. e The base class name may be omitted in a base class initializer if there is only one immediate base class e Assignment to this in constructors and destructors is allowed This is allowed only if anachronisms are enabled and the assignment to this configuration parameter is enabled e A bound function pointer a pointer to a member function for a given object can be cast to a pointer to a function e A nested class name may be used as a non nested class name provided no other class of that name has been declared This anachronism is not applied to template classes e A reference to a non const type may be initialized from a value of a different type A temporary is created it is initialized from the converted initial value and the reference is set to the temporary e A reference to a non const class type may be initialized from an rvalue of the class type or a derived class thereof No additional temporary is used 295 e A function with old style parameter declarations is allowed and may participate in function overloading as though it was prototyped Default argument promotion is not applied to parameter types of such functions when the check for compatibility is done so that the following declares the overloading of two functions named f int f int int f x char x return x It will be noted that in C this code is legal but has a different meaning a tentative declaration of f is followed by its definition e When n
17. ebp The frame pointer holds a base address for the current stack frame Consequently a function has registers pointing to both ends of its frame Incoming arguments reside in the previous frame referenced as positive offsets from ebp while local variables reside in the current frame referenced as negative offsets from ebp A function must preserve this register value for its caller Run time Environment 211 Joeax esi edi Joecx Yoedx st 0 st 1 Gerti EFLAGS Integral and pointer return values appear in eax A function that returns a structure or union value places the address of the result in eax Otherwise this is a scratch register These local registers have no specified role in the standard calling sequence Functions must preserve their values for the caller Scratch registers have no specified role in the standard calling sequence Functions do not have to preserve their values for the caller Floating point return values appear on the top of the floating point register stack there is no difference in the representation of single or double precision values in floating point registers If the function does not return a floating point value then the stack must be empty Floating point scratch registers have no specified role in the standard calling sequence These registers must be empty before entry and upon exit from a function The flags register contains the system flags such as the direction f
18. firstprivate list reduction operator list copyin list if scalar_expression num_threads scalar_integer_expression This pragma declares a region of parallel execution It directs the compiler to create an executable in which the statements within the following C C structured block are executed by multiple lightweight threads The code that lies within the structured block is called a parallel region The OpenMP parallelization pragmas support a fork join execution model in which a single thread executes all statements until a parallel region is encountered At the entrance to the parallel region a system dependent number of symmetric parallel threads begin executing all statements in the parallel region redundantly These threads share work by means of work sharing constructs such as parallel for loops see the next example The number of threads in the team is controlled by the OMP_NUM_THREADS environment variable If OMP_NUM_THREADS is not defined the program will execute parallel regions using only one processor Branching into or out of a parallel region is not supported All other shared memory parallelization pragmas must occur within the scope of a parallel region Nested omp parallel pragmas are not supported and are ignored There is an implicit barrier at the end of a parallel region When all threads have completed execution of the parallel region a single thread resumes execution of the statements that follow 126 Ope
19. for loop to be executed in parallel gt Clauses private list shared list default shared none firstprivate list lastprivate list reduction operator list copyin list if scalar_expression ordered schedule kind chunk num_threads scalar_integer_expression The semantics of the omp parallel for pragma are identical to those of a parallel region containing only a single parallel for loop and pragma The available clauses are as defined in Section 6 2 omp parallel and Section 6 6 omp for 6 9 omp sections The omp sections pragma uses the following syntax pragma omp sections Clauses pragma omp section lt C C structured block executed by processor i gt pragma omp section lt C C structured block executed by processor j gt OpenMP Pragmas for C and C 133 Clauses private list firstprivate list lastprivate list reduction operator list nowait The omp sections pragma defines a non iterative work sharing construct within a parallel region Each section is executed by a single thread If there are more threads than sections some threads will have no work and will jump to the implied barrier at the end of the construct If there are more sections than threads one or more threads will execute more than one section An omp section pragma may only appear within the lexical extent of the enclosing omp sections pragma In addition the code within the omp se
20. function observe the following return value convention e When the C or C function returns a value call it from Fortran as a function and otherwise call it as a subroutine For a C C program calling a Fortran function the call should return a similar type Table 10 1 Fortran and C C Data Type Compatibility lists compatible types If the call is to a Fortran subroutine a Fortran CHARACTER function or a Fortran COMPLEX function call it from C C as a function that returns void The exception to this convention is when a Fortran subroutine has alternate returns call such a subroutine from C C as a function returning int whose value is the value of the integer expression specified in the alternate RETURN statement 188 Inter language Calling 10 4 Upper and Lower Case Conventions Underscores By default on Linux systems all Fortran symbol names are converted to lower case C and C are case sensitive so upper case function names stay upper case When you use inter language calling you can either name your C C functions with lower case names or invoke the Fortran compiler command with the option Mupcase in which case it will not convert symbol names to lower case When programs are compiled using one of the PGI Fortran compilers on Linux systems an underscore is appended to Fortran global names names of functions subroutines and common blocks This mechanism distinguishes Fortran name space from C C name sp
21. r8 home 104 rbp rdx home 112 rbp rcx home 120 rbp return address current 128 rbp caller s Yorbp 0 rsp variable size Key points concerning the stack frame The parameter area at the bottom of the stack must contain enough space to hold all the parameters needed by any function call Space must be set aside for the four register parameters to be homed to the stack even if there are less than four register parameters used in a given call Sixteen byte alignment of the stack is required except within a function s prolog and within leaf functions All registers on an x64 system are global and thus visible to both a calling and a called function Registers rbx rsp rbp rsi rdi r12 r13 r14 and r15 are non volatile Therefore a called function must preserve these registers values for its caller Remaining registers are scratch If a calling function wants to preserve such a register value across a function call it must save a value in its local stack frame Registers are used in the standard calling sequence The first four arguments are passed in registers Integral and pointer arguments are passed in these general purpose registers listed in order rcx rdx r8 r9 Floating point arguments are passed in the first four XMM registers xmm0 Gamm xmm2 xmm3 Registers are assigned using the argument s ordinal position in the argument list For example if a function s
22. safe_lastval In the case where a scalar is used after a loop but is not defined on every iteration of the loop the compiler does not by default parallelize the loop However this option tells the compiler it safe to parallelize the loop For a given loop the last value computed for all scalars make it safe to par allelize the loop Code Generation no nosave determines whether the compiler assumes that all Fortran Language local variables are subject to the SAVE statement pgf77 pgf95 and pghpf only no scalarsse do don t use SSE SSE2 instructions to perform Optimization scalar floating point arithmetic schar specifies signed char for characters pgcc and C C Language pgCC only also see uchar no second_underscore do don t add the second underscore to the name of a Fortran global if its name already contains an underscore pgf77 pgf95 and pghpf only Code Generation no signextend do don t extend the sign bit if it is set Code Generation nosingle do don t convert float parameters to double parameter characters pgcc and pgCC only C C Language no smart do don t enable optional AMD64 specific post pass assembly optimizer Optimization no smartalloc add a call to the routine mallopt in the main rou tine To be effective this switch must be specified when compiling the file containing the Fortran C or C main program Environment
23. standard causes the compiler to flag source code that does not conform to the ANSI standard pgf77 pgf95 and pghpf only Fortran Language nostrideO the compiler generates does not generate alter nate code for a loop that contains an induction variable whose increment may be zero pgf77 pgf95 and pghpf only Code Generation uchar specifies unsigned char for characters pgcc and pgCC only also see schar C C Language 60 Command Line Options pgflag Description Category unix uses UNIX calling and naming conventions for CodeGeneration Fortran subprograms pgf77 pgf95 and pghpf for Win32 only noJnounixlogical determines whether logical TRUE and FALSE Fortran Language are determined by non zero TRUE and zero FALSE values for unixlogical With nounixlogi cal the default 1 values are TRUE and 0 values are FALSE pgf77 pgf95 and pghpf only noJunroll controls loop unrolling Optimization noupcase determines whether the compiler allows upper Fortran Language case letters in identifiers pgf77 pgf95 and pghpf only varargs force Fortran program units to assume calls are to Code Generation C functions with a varargs type interface pgf77 and pgf95 only no vect do don t invoke the code vectorizer Optimization Following are detailed descriptions of several but not all of the M lt pgflag gt
24. If the operator portion of the reduction clause is omitted the default reduction operator is addition e The copyin clause applies only to threadprivate variables In the presence of the copyin clause data from the master thread s copy of the threadprivate variable is copied to the thread private copies upon entry to the parallel region e In the presence of an if clause the parallel region will be executed in parallel only if the corresponding scalar_expression evaluates to a non zero value Otherwise the code within the region will be executed by a single processor regardless of the value of the environment variable OMP_NUM_THREADS e If the num_threads clause is present the corresponding scalar_integer_expression must evaluate to a positive integer value This value sets the maximum number of threads used during execution of the parallel region A num_threads clause overrides either a previous call to the library routine omp_set_num_threads or the setting of the OMP_NUM_THREADS environment variable 6 3 omp critical The OpenMP omp critical pragma uses the following syntax pragma omp critical name lt C C structured block gt Within a parallel region there may exist subregions of code that will not execute properly when executed by multiple threads simultaneously This is often due to a shared variable that is written and then read again 128 OpenMP Pragmas for C and C The omp critical pragma defines a
25. Integral and Pointer Arguments As mentioned a function receives all its arguments through the stack the last argument is pushed first In the standard calling sequence the first argument is at offset 8 ebp the second argument is at offset 12 ebp etc as previously shown in Table A 3 Stack Contents for Functions Returning struct union Functions pass all integer valued arguments as words expanding or padding signed or unsigned bytes and halfwords as needed 214 Run time Environment Table A 4 Integral and Pointer Arguments Call Argument Stack Address g 1 2 3 void 0 1 8 ebp 2 12 ebp 3 16 ebp void 0 20 ebp Floating Point Arguments The stack also holds floating point arguments single precision values use one word and double precision use two The example below uses only double precision arguments Table A 5 Floating point Arguments Call Argument Stack Address h 1 414 1 2 998e10 word 0 1 414 8 Yebp word 1 1 414 12 ebp 1 16 ebp word 0 2 998e10 20 ebp word 1 2 998e10 24 ebp Structure and Union Arguments Structures and unions can have byte halfword or word alignment depending on the constituents An argument s size is increased if necessary to make it a multiple of words This may require tail padding depending on the size of the argument Structure and union arguments are pushed
26. Kpic and KPIC options are supported for compatibility with other systems you may have used and are equivalent to fpic For example use the following command to create an object file with position independent code using pgf95 o pgf95 c fpic tobeshared f To produce a shared object file use the appropriate PGI compiler to invoke the linker supplied with your system It is customary to name such files using a so filename extension On Linux this is done by passing the shared option to the linker o pgf95 shared o tobeshared so tobeshared o Note that compilation and generation of the shared object can be performed in one step using both the fpic option and the appropriate option for generation of a shared object file To use a shared object file compile and link the program which will reference functions or subroutines in the shared object file using the appropriate PGI compiler and listing the shared object on the link line o pgf95 o myprog myprof f tobeshared so You now have an executable myprog which does not include any code from functions or subroutines in tobeshared so but which can be executed and dynamically linked to that code By default when the program is linked to produce myprog no assumptions are made on the location of tobeshared so In order for myprog to execute correctly you must initialize the environment variable LD_LIBRARY_PATH to include the directory containing tobeshared so If LD_LIBRA
27. Mconcur 2 Unoptimized code compiled using the option O0 can be significantly slower than code generated at other optimization levels Like the Mvect option the Munroll option sets the optimization level to level 2 if no O or g options are supplied For more information on optimization see Optimization amp Parallelization Usage In the following example since no optimization level is specified and a O option is specified the compiler sets the optimization to level 2 pgf95 O myprog f Cross reference g M lt pgflag gt 0 Names the executable file Use the o option to specify the filename of the compiler object file The final output is the result of linking Command Line Options 85 Syntax o filename Where filename is the name of the file for the compilation output The filename must not have a f extension Default The compiler creates executable filenames as needed If you do not specify the o option the default filename is the linker output file a out Usage In the following example the executable file is myprog instead of the default a out pgf95 myprog f o myprog Cross reference c E F S pc tp px p5 p6 piii targets only The pc option can be used to control the precision of operations performed using the x87 floating point unit and their representation on the x87 floating point stack Syntax pc 321641 80 The x87 architecture i
28. Mnoopenmp when used in combination with the mp option causes the compiler to ignore OpenMP parallelization directives or pragmas but still process SGI style parallelization directives or pragmas Mnosgimp when used in combination with the mp option causes the compiler to ignore SGI style parallelization directives or pragmas but still process OpenMP parallelization directives or pragmas Mnopgdllmain Windows only do not link the module containing the default DlIMain Q into the DLL This flag applies to building DLLs with the PGF95 and PGHPF compilers If you want to replace the default DI Main Q routine with a custom DllMain use this flag and add the object containing the custom DllMain to the link line The latest version of the default DIIMain used by PGF95 and PGHPF is included in the Release Notes for each release the PGF95 and PGHPF specific code in this routine must be incorporated into the custom version of DllMain to ensure the appropriate function of your DLL Mpreprocess perform cpp like pre processing on assembly and Fortran input source files Default For arguments that you do not specify the default miscellaneous options are as follows inform warn nolist nobounds Usage In the following example the compiler includes Fortran source code with the assembly code pgf95 Manno S myprog f In the following example the compiler displays information about inlined functions with fewer than approximate
29. Mvect in Chapter 3 Command Line Options for a detailed description of all available sub options 2 5 2 Vectorization Example Using SSE SSE2 Instructions One of the most important vectorization options is Mvect sse This section contains an example of the use and potential effects of Mvect sse When the compiler switch Mvect sse is used the vectorizer in the PGI Workstation compilers automatically uses SSE and SSE2 instructions where possible when targeting processors where these are supported This capability is supported by all of the PGI Fortran C and C compilers See table P 2 for a complete specification of which x86 and x64 processors support SSE and SSE2 instructions Using Mvect sse performance improvements of up to two times over equivalent scalar code sequences are possible In the program in Example 2 3 Vector operation using SSE instructions the vectorizer recognizes the vector operation in subroutine loop when the compiler switch Mvect sse is used This example shows the compilation informational messages and runtime results using the SSE instructions on an AMD Opteron processor based system along with issues that affect SSE performance First note that the arrays in Example 2 3 Vector operation using SSE instructions are single precision and that the vector operation is done using a unit stride loop Thus this loop can potentially be vectorized using SSE instructions on any processor that
30. TRUE FALSE letterl numintl1 numint2 numfloatl numdoub1 numshor1 To compile this Fortran subroutine and C program use the following commands pgcc c cmain f pgf95 Mnomain cmain o forts f Executing the resulting a out file should produce the following output TRUE v 11 44 39 6 902 299 Inter language Calling 195 10 10 Example C Calling C Example 10 7 Simple C Function cfunc c void cfunc numl num2 res int numl num2 res printf func a Sd b d ptr c x n num1 num2 res res numl1 num2 printf fune res d n res Example 10 8 C Main Program cpmain C Calling a C Function xtern C void cfunc int n int m int p include lt iostream gt main int a b c a 8 b 2 cout lt lt main a lt lt a lt lt b lt lt b lt lt ptr c lt lt amp c lt lt endl cfunc a b amp c cout lt lt main res lt lt c lt lt endl To compile this C function and C main program use the following commands pgcc c csub c pgCC cpmain C csub o Executing the resulting a out file should produce the following output main a func a b 2 ptr c DEE C oxbffffb94 bffffb94 8 8 func res main res eeo 196 Inter language Calling 10 11 Example C Calling C Example 10 9 Simple C Function cpfunc C with Extern C include lt iostream gt extern C void cpfunc int numl int num2 int res cou
31. e The THREADPRIVATE directive must appear after every declaration of a thread private common block Only named common blocks can be made thread private e It is illegal for a THREADPRIVATE common block or its constituent variables to appear in any clause other than a COPYIN clause e A variable can appear in a THREADRIVATE directive only in the scope in which it is declared It must not be an element of a common block or be declared in an EQUIVALENCE statement e A variable that appears in a THREADPRIVATE directive and is not declared in the scope of a module must have the SAVE attribute OpenMP Directives for Fortran 121 5 18 Run time Library Routines User callable functions are available to the Fortran programmer to query and alter the parallel execution environment integer omp_get_num threads returns the number of threads in the team executing the parallel region from which it is called When called from a serial region this function returns 1 A nested parallel region is the same as a single parallel region By default the value returned by this function is equal to the value of the environment variable OMP_NUM_THREADS or to the value set by the last previous call to the omp_set_num_threads subroutine defined below subroutine omp_set_num_threads scalar_integer_ exp sets the number of threads to use for the next parallel region This subroutine can only be called from a serial region of code If it is called from withi
32. hidden first argument in rdi This address will also be returned in rax A 2 3 Argument Passing Integral and Pointer Arguments Integral and pointer arguments are passed to a function using the next available register of the sequence rdi rsi rdx Yrcx r8 r9 After this list of registers has been exhausted all remaining integral and pointer arguments are passed to the function via the stack Floating Point Arguments Float and double arguments are passed to a function using the next available XMM register taken in the order from xmm0 to xmm7 After this list of registers has been exhausted all remaining float and double arguments are passed to the function via the stack Structure and Union Arguments Structure and union arguments can be passed to a function in either registers or on the stack The size and type of the structure or union determine how it is passed If a structure or union is larger than 16 bytes it is passed to the function in memory 222 Run time Environment To determine whether a 16 byte or smaller structure or union can be passed to a function in one or two registers examine the first eight bytes of the structure or union The type or types of the structure or union s fields making up these eight bytes determine how these eight bytes will be passed If the eight bytes contain at least one integral type the eight bytes will be passed in the first available general purpose register of the sequence r
33. ing at runtime C Stops after the assembly phase and saves the object code in file name o D lt args gt Defines a preprocessor macro dryrun Show but do not execute driver commands E Stops after the preprocessing phase and displays the preprocessed file on the standard output F Stops after the preprocessing phase and saves the preprocessed file in filename f this option is only valid for the PGI Fortran compilers fast Generally optimal set of flags for the target fastsse Generally optimal set of flags for targets that include SSE SSE2 capa bility flags Display valid driver options fpic Linux only Generate position independent code fPIC Linux only Equivalent to fpic g Includes debugging information in the object module g77libs Linux only Allow object files generated by g77 to be linked into PGI main programs gopt Includes debugging information in the object module but forces assembly code generation identical to that obtained when is not present on the command line help Display driver help message I lt dirname gt Adds a directory to the search path for include files 40 Command Line Options Option Description i2 i4 and i8 Treat INTEGER variables as 2 bytes i2 14 and i8 Treat INTEGER variables as 4 bytes i2 i4 and i8 Treat INTEGER and LOGICAL variables as 8 bytes and use 64 bits for INTEGER 8 operations K lt flag gt Requests special comp
34. object2 f function funf2 real funf2 funf2 2 0 return end prog1 f 162 program test external subf1 real funf2 val integer n call subtil n Libraries and Environment Variables val funf2 write val val stop end 2s Create the DLL obj12 dIl and its import library obj12 lib using the following series of commands pgf95 c object1l f object2 f pgf95 objectl o object2 o Mmakedll o obj12 dll out implib obj12 lib 3 Compile the main program pgf95 Mdll o progl progl f L lobj12 The Mdll switch causes the compiler to link against the PGI runtime DLLs instead of the PGI runtime static libraries The Mdll switch is required when linking against any PGI compiled DLL such as obj12 dll The l switch is used to specify that obj12 lib the DLL s import library will be used to resolve the calls to subf1 and funf2 in prog1 f 4 Ensure that obj12 dll is in your path then run the executable prog1 to determine if the DLL was successfully created and linked progl n 1 val 2 000000 FORTRAN STOP Should you wish to change obj12 d11 without changing the subroutine or function interfaces no rebuilding of prog is necessary Just recreate obj12 dll and the new obj12 dll will be loaded at runtime Example 8 2 Build two DLLs when each DLL is dependent on the other and use them to build the main program In the following source files object3 c makes calls to routines de
35. t ignore potential data dependen depchk Drg cies eqvchk noeqvchk Do don t check EQUIVALENCE s for data eqvchk Drg dependencies invarif noinvarif Do don t remove invariant if constructs invarif Drg from loops ivdep Ignore potential data dependencies depchk Drg Istval nolstval Do don t compute last values Istval Drg opt Select optimization level N A lg safe_lastval Parallelize when loop contains a scalar not enabled dl used outside of loop unroll nounroll Do don t unroll loops nounroll Drg vector novector Do don t perform vectorizations vector Drg vintr novintr Do don t recognize vector intrinsics vintr Drg In the case of the vector novector directive the scope is the code following the directive until the end of the routine for r scoped directives as opposed to the entire routine or until the end of the file for g scoped directives as opposed to the entire file altcode noaltcode Optimization Directives and Pragmas 143 Instructs the compiler to generate alternate code for vectorized or parallelized loops The noaltcode directive disables generation of alternate code This directive affects the compiler only when Mvect or Mconcur is enabled on the command line cpgi altcode cpgi altcode alignment cpgi altcode n concur cpgi altcode n concurreduction cpgi altcode n nontemporal cpgi altcode n nopeel cpgi altcode n vector 144 Enables a
36. 1 pgCC only Enable compilation of C with compatibility with cfront version 3 0 create_pch filename pgCC only Create a precompiled header file with the name filename dependencies see M pgCC only Print makefile dependencies to stdout dependencies_to_file filename diag_error tag pgCC only Print makefile dependencies to file filename pgCC only Override the normal error severity of the spec ified diagnostic messages diag_remark tag pgCC only Override the normal error severity of the specified diagnostic messages Command Line Options 43 Option Description diag_suppress tag pgCC only Override the normal error severity of the spec ified diagnostic messages diag_warning tag pgCC only Override the normal error severity of the spec ified diagnostic messages display_error_number pgCC only Display the error message number in any diag nostic messages that are generated e lt number gt pgCC only Set the C front end error limit to the speci fied lt number gt no_ exceptions pgCC only Disable enable exception handling support The default is exceptions gnu_extensions pgCC only Allow GNU extensions like include next which are required to compile Linux system header files no llalign pgCC only Do don t align long long integers on integer boundaries Th
37. 120 DOAGROSS eiretadas EE 116 Function Inlining inlining and makefiles AAA 103 inlining examples secen 104 inlining restrictions s e 105 l Inter language Calling c ceceeeeeeeeeeeeees 187 KEE 192 arguments and return values eeeee 191 array INGICES 0 ee eeeeeeeeteeeeeeeeteneeeeeneeeeees 193 Calling CA 197 Ctteadling KEE 196 C Calling Fortran sessen 199 character case conventions c eee 189 character return values A 192 compatible data types 189 Fortran calling C A 193 Fortran calling Ca 197 UNderSC rE S acia eege ec deer deu 189 L Language Options 68 Libraries BLAS heech tacit ee AEE E A A T LAPAGK AATE EE shared object files lk EE Header Files Parallelization LIStING FIIGS aie anit nee LOOP UNPOILNG ee eeeeeeeeeeneeeeseeeeeeeeeeeneeeeeaees 19 Index Loops failed auto paralleltzatton cesses 27 Ed ue EE 27 Del 28 ul ue EE T 27 N Command line Options SU lte ET 98 alternative_tokensS sssneeeeseee neern 95 SEL EE 96 S OXCOPUOMS viecisssicccnsiceessssesessedenaucessescesnsesnes 97 PCN_MESSAQES eeeeeeseeeeeneeeeeneeeeeneeeesaees 99 FUSING SIA WEE 99 oO OpenMP C C Pragmas sseseeseeeeeeeeeeeee 125 e UE 135 Darrier ranen i Bas a os 132 CHIUGAl EE 128 UE EE 136 TOM EE 130 MASON EE 129 ordered EE 135 Parallel ccs ssiesdessecegs ge inanin aseesasaieaices 126 parallel fli s sc ecseesteeeeeteeseepescenetege cece 133 pa
38. 97 ClAG_SUPPISS eseeceeseseeeesrereseneeeeeees 97 ag warmmimg o e 97 display_error number seeen 97 S IER EECH 48 er ebe iter ety tied 48 E 48 PAST SEE A AESA E EEE AA 48 49 flag EE 49 APG EA EAA 49 STIG is E E oetoapies est dye ceosereel acgecepeesaess 49 Gin ian tree nantes 50 SO rvee testes Sege dE at geen anne E 50 SO7 TNS A E E tt 50 S ll 50 Nel Divisione vita E 51 e BEE 51 AQ Se 1B EE 52 MIT AG NEE 52 tks ivi aided Sided ian ata ates 53 Eege E ee 54 ageet Bongen eine EE 98 MANNO rtrd capa irane aguek SERA 79 Masmkeyword AAA 68 Mbackalach eraen 66 MBOUNGS i ccccsceccsevess desdedeciseiendavaiscocteeseeseaats 79 MbyteSwapio eeeeeeseeeeseeeeeeneeeeeneeeeeaees 79 Mcache align eeesccceseceeeseseeesseeteeeees 70 Mchkipstky e s san teniron senie 79 chMebktfp e a dE naen ieee 79 AEI EPEE EAT AAT 79 mcTmmodel medium ER Mconcur Mcray MD 9 Mdaz Mdee k wists vvacrcdacisvasetavnarsnencaiae e aa Mdefaultunit Mdepchk Mdee see e r E EKGs Gd UR chgioll t iirinn deeg ENER due deet Mdwarf1 Mdwarf2 Mdwarf3 Mextend Mextract Mfcon Mfixed gl UE Mfprelaxed Miree re Mfunc32 Mgccbugs AMAR Zeg ed eas hk Be MINOR EE dl lo le SOllelvull EEEE 67 MID EE 72 MK O PaSIM cccescceeeeeereseeeteseeeeessneeesecees 81 Mlarge array Sneinen ennan 62 Mrs 64 Mlist 82 Stee 74 Mmakedll nen 82
39. A common block occurs in more than one subprogram of a source file and its size is not identical The maximum size is chosen The message is applicable only when an assembly file is the output of the compiler 168 Incompatible size of common block A common block occurs in more than one subprogram of a source file and is initialized in one subprogram Its initialized size was found to be less than its size in the other subprogram s The message is applicable only when an assembly file is the output of the compiler W169 Multiple data initializations of common block A common block is initialized in more than one subprogram of a source file Only the first set of initializations apply The message is applicable only when an assembly file is the output of the compiler Messages 265 W170 F90 extension Use of a nonstandard feature A description of the feature is provided W171 F90 extension nonstandard statement type W172 F90 extension numeric initialization of CHARACTER A CHARACTER 1 variable or array element was initialized with a numeric value W173 W174 W175 W176 W177 W178 W179 W180 w181 W182 F90 F90 F90 F90 F90 extension nonstandard use of data type length specifier extension type declaration contains data initialization extension IMPLICIT range contains nonalpha characters extension nonstandard operator extension nonstandard use of keyword argument
40. Command Line Options 83 Default The compiler places mod files in the current working directory and searches only in the current working directory for pre compiled intermediate mod files Usage The following command line requests that any intermediate module file produced during compilation of myprog f be placed in the directory mymods in particular the file mymods myprog mod will be used pgf95 module mymods myprog f mp align no Jnuma Use the mp option to instruct the compiler to interpret user inserted OpenMP shared memory parallel programming directives and generate an executable file which will utilize multiple processors in a shared memory parallel system See OpenMP Directives for Fortranand OpenMP Pragmas for C and C for a detailed description of this programming model and the associated directives and pragmas The align sub option forces loop iterations to be allocated to OpenMP processes using an algorithm that maximizes alignment of vector sub sections in loops that are both parallelized and vectorized for SSE This can improve performance in program units that include many such loops It can result in load balancing problems that significantly decrease performance in program units with relatively short loops that contain a large amount of work in each iteration The numa suboption uses libnuma on systems where it is available Default The compiler ignores user inserted shared memory parallel programming directive
41. DO loop DO END DO directive pairs may not be nested Branching into or out of a OMP DO loop is not supported By default there is an implicit barrier after the end of the parallel loop the first thread to complete its portion of the work will wait until the other threads have finished their portion of work If NOWAIT is specified the threads will not synchronize at the end of the parallel loop Other items to note about OMP DO loops The DO loop index variable is always private 114 OpenMP Directives for Fortran e S OMP DO loops must be executed by all threads participating in the parallel region or none at all The END DO directive is optional but if it is present it must appear immediately after the end of the enclosed DO loop PROGRAM DO USE REAL A 1000 B 1000 DO I 1 1000 B I FLOAT I ENDDO SOMP PARALLEL SOMP DO DO I 1 1000 A I SQRT B I ENDDO SOMP END PARALLEL END The SCHEDULE clause specifies how iterations of the DO loop are divided up between processors Given a SCHEDULE type chunk clause type can be STATIC DYNAMIC GUIDED or RUNTIME These are defined as follows e When SCHEDULE STATIC chunk is specified iterations are allocated in contiguous blocks of size chunk The blocks of iterations are statically assigned to threads in a round robin fashion in order of the thread ID numbers The chunk must be a scalar integer expression If chunk is not specified a default c
42. Detection during INliniNng ssssesssssrrsssssrrsssrirrssriirrrnserrnsssrennnnnnt 104 Examples enoni srie aa aE EE EEA irii 104 Restrictions op MIMMO AANEREN 105 OpenMP Directives for Fortran cccsssseeeeeeseneeeeeeeeeeeeeeeeeeees 107 Parallelization Directives ccccccesccccceeeeeccceeeeseseeceeeeseeeceeeeeeneeaeeeeeenenees 107 PARALLEL END PARALLEL aipania aa ia aaa aaa 108 CRITICAL END CRITICA D arrea auhena ee atasan aeaiia aaea ainar aaia 111 MASTER END MASTER ccccececcceeseeeceeeeeeeeeeeseeeeeeeaaeeseeeeesaaeeeeenees 112 SINGLE END SINGL E aaarnas ed a aantig aaia aaaea aa aiim 113 DOr END DO ete ee se Hae ties ev E e a ere Ee 113 WORKSHARE END WORKSHARE nsiisssesseesseesiresisserrrssrnserrnsernsne 116 BAR PIER A E RA A AT 116 Contents 5 9 5 10 5 11 5 12 5 13 5 14 5 15 5 16 5 17 5 18 5 19 Chapter 6 6 16 Chapter 7 Chapter 8 8 1 8 2 8 3 8 4 Contents PARADE ed Oe coe tert en ees eg Oe ee ae lee EE 117 PARALLEL WECKER eege angueerd Tee ANE as teaceg aaaea 118 SECTIONS END SECTIONS oo ecccceceeeeeeeeeeeeeeeeeeeeeeseaeeeeeeeeesenaeeeeaes 118 PARALLEL SECTIONS A eis eebe SE ah beatin ee re ht aa 119 EIERE ee Ehe e 120 ATOMIC defteg gaere Sauer e e o aaran erer ial eed a araia ERSS 120 PES EE 121 THREADPRIVATE buede AEN E 121 Run time Library Routines c ce cceeeeeceeeeeeeeneeeeeeeeaeeeeeeeeaeeeeeeenaeeeeeeeeaa 122 Environment EE 124 OpenMP Pragmas f
43. EEN 20 77 inline lIDrarieS sensira 102 INSWUCIONS EE 18 Inter Procedural Anahvsie AA 14 LEE 14 W local Optimization ceccseeseeeeteeeteeeees 13 Win32 Calling Conventions loop OPTIMIZATION cceeeecseseesetecseteeeeeeees 14 EE 200 202 loop unrolling eceeeeeeeeeeeeeeeeeeeeeeeeees 14 19 Defaults aaran ea iee aia deeg 200 202 ee TD cis cava daacectess cries beni teties acne 200 202 rO A E ENEE tees symbol name Construchon 201 Se UNIX style ENEE 200 202 OT eege dee Ee Eege he SI enge Ee ed OI level eer ees dE parallelization PEO ees eet een E hes tte DOIDIEES gees d prefetching profile feedback PFO sseeeseseeseeeeee 35 Profile Feedback Optimization 14 vechortzation 14 20 306 Index
44. Environment The logical constants TRUE and FALSE are defined to be the four byte values 1 and 0 respectively A logical expression is defined to be TRUE if its least significant bit is 1 and FALSE otherwise Note that the value of a character is not automatically NULL terminated A 2 4 2 Naming Conventions By default all globally visible Fortran symbol names subroutines functions common blocks are converted to lower case In addition an underscore is appended to Fortran global names to distinguish the Fortran name space from the C C name space A 2 4 3 Argument Passing and Return Conventions Arguments are passed by reference i e the address of the argument is passed rather than the argument itself In contrast C C arguments are passed by value When passing an argument declared as Fortran type CHARACTER an argument representing the length of the CHARACTER argument is also passed to the function This length argument is a four byte integer passed by value and is passed at the end of the parameter list following the other formal arguments A length argument is passed for each CHARACTER argument the length arguments are passed in the same order as their respective CHARACTER arguments A Fortran function returning a value of type CHARACTER adds two arguments to the beginning of its argument list The first additional argument is the address of the area created by the caller for the return value the second additio
45. Fortran structures and derived types and multiple map declarations within a Fortran union conform to the same alignment requirements used by C structures Common Blocks A named Fortran common block can be represented in C C by a structure whose members correspond to the members of the common block The name of the structure in C C must have the added underscore For example the Fortran common block INTEGER I J COMPLEX C DOUBLE COMPLEX CD DOUBLE PRECISION D COMMON COM i j c cd d is represented in C with the following equivalent extern struct int i int j struct float real imag c struct double real imag cd double d com and in C with the following equivalent extern C struct int i in J struct float real imag c struct double real imag cd double d com_ Note that the compiler provided name of the BLANK COMMON block is implementation specific 244 Run time Environment Calling Fortran COMPLEX and CHARACTER functions from C C is not as straightforward as calling other types of Fortran functions Additional arguments must be passed to the Fortran function by the C C caller A Fortran COMPLEX function returns its value in memory the first argument passed to the function must contain the address of the storage for this value A Fortran CHARACTER function adds two arguments to the beginning of its argument list The following example of calling a Fortran CHARACTER fun
46. INTEGER 263 to 263 1 LOGICAL same as INTEGER true or false LOGICAL 1 8 bit value true or false LOGICAL 2 16 bit value true or false LOGICAL 4 same as INTEGER true or false LOGICAL 8 same as INTEGER true or false BYTE 2 s complement 128 to 127 REAL Single precision floating point 10 37 to 1038 1 REAL 4 Single precision floating point 10 37 to 1038 1 REAL 8 Double precision floating point 10 307 to 10308 1 DOUBLE PRECISION Double precision floating point 10 307 to 10308 1 COMPLEX See REAL See REAL DOUBLE COMPLEX See DOUBLE PRECISION See DOUBLE PRECI SION COMPLEX 16 Same as above Same as above CHARACTER n Sequence of n bytes 1 Approximate value 176 Fortran C and C Data Types The logical constants TRUE and FALSE are all ones and all zeroes respectively Internally the value of a logical variable is true if the least significant bit is one and false otherwise When the option Munixlogical is set a logical variable with a non zero value is true and with a zero value is false Fortran C and C Data Types 177 Table 9 2 Real Data Type Ranges Data Type Binary Range Decimal Range Digits of Precision REAL 2 126 to 2128 10 37 to 1038 7 8 REAL 8 2 1022 to 21024 10 307 to 10308 15 16 Table 9 3 Scalar Type Alignment Type Is Aligned ona LOGICAL 1 1 byte boundary LOGICAL 2 2 byte boundary LOGICAL 4 4 byte bou
47. Instruction Counting As this chapter shows once you have a program that compiles executes and gives correct results you may optimize your code for execution efficiency Selecting the correct optimization level requires some thought and may require that you compare several optimization levels before arriving at the best solution To compare optimization levels you need to measure the execution time for your program There are several approaches you can take for timing execution You can use shell commands that provide execution time statistics you can include function calls in your code that provides timing information or you can profile sections of code Timing functions available with the PGI compilers include 3F timing routines the SECNDS pre declared function in PGF77 or PGF95 or the SYSTEM_CLOCK or CPU_CLOCK intrinsics in PGF95 or PGHPF In general when timing a program one should try to eliminate or reduce the amount of system level activities such as program loading I O and task switching The following example shows a fragment that indicates how to use SYSTEM_CLOCK effectively within either an HPF or F90 F95 program unit Example 2 4 Using SYSTEM_CLOCK integer nprocs hz clock0O clockl real time integer allocatable t botz distribute t cyclic if defined HPF allocate t number of processors elif defined _OPENMP allocate t OMP_GET NUM THREADS else allocate t 1 endif call sys
48. Register Allocation for Example AA 237 Win64 Fortran Fundamental Types cccececcccceeeeeecneeeeeeeenaeeeeeeeenaeeeeeee 240 Fortran and C C Data Type Compatbiltv nesre 243 Fortran and C C Representation of the COMPLEX Type 243 Figures Figure 9 1 Internal Padding in a Structure seeeeeeeeeeeenesen esn nsesrnsnnrsnnsrnnsssrnesrnnses Figure 9 2 Tail Padding in a Structure Preface This guide is part of a set of manuals that describe how to use The Portland Group PGI Fortran C and C compilers and program development tools In particular these include the PGF77 PGF95 PGHPF PGC and PGCC ANSI C compilers the PGPROF profiler and the PGDBG debugger These compilers and tools work in conjunction with an x86 or x64 assembler and linker You can use the PGI compilers and tools to compile debug optimize and profile serial and parallel applications for x86 Intel Pentium IJ TII 4 M Intel Centrino Intel Xeon AMD Athlon XP MP or x64 AMD Athlon64 Opteron Turion Intel EM64T processor based systems This PGI User s Guide provides operating instructions for the command level compilation environment and general information about PGI s implementation of the Fortran C and C languages This guide does not teach the Fortran C or C programming languages Audience Description This guide is intended for scientists and engineers using the PGI compilers To use these compilers you should be aware of the rol
49. Static libraries and DLLs also handle global data differently Global data in static libraries is automatically accessible to other objects linked into an executable Global data in a DLL can only be accessed from outside the DLL if the DLL exports the data and the image that uses the data imports it To this end the C compilers support the Microsoft storage class extensions __declspec dllimport and __declspec dllexport These extensions may appear as storage class modifiers and enable functions and data to be imported and exported extern int _ declspec dllimport intfunc float declspec dllexport fdata The Fortran compilers support the DEC ATTRIBUTES extensions DLLIMPORT and DLLEXPORT cDEC ATTRIBUTES DLLEXPORT object object CDECS ATTRIBUTES DLLIMPORT object object cis one of C c or object is the name of the subprogram or common block that is exported or imported Note that common block names are enclosed within slashes In example cDEC ATTRIBUTES DLLIMPORT intfunc DECS ATTRIBUTES DLLEXPORT fdata The Examples in this section further illustrate the use of these extensions To create a DLL from the command line use the Mmakedll option The following switches apply to making and using DLLs with the PGI compilers Mdll Link with the DLL version of the runtime libraries This flag is required when linking with any DLL built by the PGI compilers Mmakedll Generate a dynamic link library or
50. Table A 8 Standard Stack Frame Position Contents Frame 8n 16 rbp argument eightbyte n previous 16 rbp argument eightbyte 0 8 rbp return address current 0 rbp caller s rbp current 8 rbp unspecified 0 rsp variable size 128 rsp red zone Key points concerning the stack frame e The end of the input argument area is aligned on a 16 byte boundary The 128 byte area beyond the location of rsp is called the red zone and can be used for temporary local data storage This area is not modified by signal or interrupt handlers e A call instruction pushes the address of the next instruction the return address onto the stack The return instruction pops the address off the stack and effectively continues execution at the next instruction after the call instruction A function must preserve non volatile registers described below Additionally the called function must remove the return address from the stack leaving the stack pointer rsp with the value it had before the call instruction was executed All registers on an x64 system are global and thus visible to both a calling and a called function Registers rbx rsp rbp r12 r13 r14 and r15 are non volatile across function calls Therefore a function must preserve these registers values for its caller Remaining registers are volatile scratch If a calling function wants to preserve such a register value ac
51. The next table lists several of these options but is not exhaustive For a complete list of available options including an exhaustive list of PGC options use the help command line option For further detail on a given option use help and specify the option explicitly 42 Command Line Options Table 3 2 C and C specific Compiler Options Option Description A pgCC only Accept proposed ANSI C no_ alternative_tokens pgCC only Enable disable recognition of alternative tokens These are tokens that make it possible to write C without the use of the amp and and characters The alternative tokens include the operator keywords e g and bitand etc and digraphs The default is no_alternative_tokens Allow C comments using in C source pgCC only Compile with cfront 2 1 compatibility This accepts constructs and a version of C that is not part of the language definition but is accepted by cfront EDG option b3 pgCC only Compile with cfront 3 0 compatibility See b above no_ bool pgCC only Enable or disable recognition of bool The default value is bool I no Do don t compile with math subroutine builtin support which causes selected math library routines to be inlined The default is builtin cfront_2 1 cfront_3 0 pgCC only Enable compilation of C with compatibility with cfront version 2
52. The sum after conversion f n f return File sum 3 s Computes paral para2 para3 text 238 Run time Environment align 16 globl sum 3 sum_3 pushq srbp leaq 128 rsp srbp cvtsi2ss ecx xmm0 addss xmml xmm0O cvtss2sd xmm0 xmm0 addsd Sxmm2 xmm0O cvtsd2ss xmm0 xmm0 popg srbp ret type sum_3 function size sum_3 sum_3 AAA Win64 Fortran Supplement Sections A3 4 1 through A3 4 4 define the Fortran supplement to the AMD64 Software Conventions for Win64 The register usage conventions set forth in that document remain the same for Fortran Run time Environment 239 AAA Fortran Fundamental Types Table A 16 Win64 Fortran Fundamental Types Fortran Type aes Ee Sa INTEGER 4 4 INTEGER 1 1 1 INTEGER 2 2 2 INTEGER 4 4 4 INTEGER 8 8 8 LOGICAL 4 4 LOGICAL 1 1 1 LOGICAL 2 2 2 LOGICAL 4 4 4 LOGICAL 8 8 8 BYTE 1 1 CHARACTER n n 1 REAL 4 4 REAL 4 4 4 REAL 8 8 8 DOUBLE PRECISION 8 8 COMPLEX 8 4 COMPLEX 8 8 4 COMPLEX 16 16 8 DOUBLE COMPLEX 16 8 A logical constant is one of TRUE FALSE 240 Run time Environment The logical constants TRUE and FALSE are defined to be the four byte values 1 and 0 respectively A logical expression is defined to be TRUE if its least significant bit is 1 and FALSE otherwise Note that the value of a character is not automatically NULL terminated
53. a previous novector directive The directives vector and novector only apply if Mvect has been selected on the command line vintr novintr The directive novintr directs the vectorizer to disable recognition of vector intrinsics The directive vintr is used to re enable recognition of vector intrinsics after a previous novintr directive The directives vintr and novintr only apply if Mvect has been selected on the command line 7 3 Scope of Directives and Command Line options This section presents several examples showing the effect of directives and the scope of directives Remember that during compilation the effect of a directive may be to either turn an option on or turn an option off Directives apply to the section of code following the directive corresponding to the specified scope that is the following loop the following routine or the rest of the program Consider the following code integer maxtime time parameter n 1000 maxtime 10 double precision a n n b n n c n n do time 1 maxtime do 4 1 T do j 1 n 148 Optimization Directives and Pragmas c i j a i j b i j enddo enddo enddo end When compiled with Mvect both interior loops are interchanged with the outer loop pgf95 Mvect dirvectl f Directives alter this behavior either globally or on a routine or loop by loop basis To assure that vectorization is not applied use the novector directive with global scope c
54. a separate copy of the variable exists on each processor e Variables declared in a firstprivate list are private and in addition are initialized from the original object existing before the construct e Variables declared in a lastprivate list are private and in addition the thread that executes the sequentially last iteration updates the version of the object that existed before the construct The reduction clause is as described in Section 6 2 omp parallel The schedule clause is explained below e If ordered code blocks are contained in the dynamic extent of the for directive the ordered clause must be present See Section 6 11 omp ordered for more information on ordered code blocks The omp for pragma directs the compiler to distribute the iterative for loop immediately following across the threads available to the program The for loop is executed in parallel by the team that was started by an enclosing parallel region Branching into or out of an omp for loop is not supported and omp for pragmas may not be nested By default there is an implicit barrier after the end of the parallel loop the first thread to complete its portion of the work will wait until the other threads have finished their portion of work If nowait is specified the threads will not synchronize at the end of the parallel loop Other items to note about omp for loops e The for loop index variable is always private and must be a signed integer e
55. and are not preserved Floating Point Control Word The control word contains the floating point flags such as the rounding mode and exception masking This register is initialized at process initialization time and its value must be preserved Signals can interrupt processes Functions called during signal handling have no unusual restriction on their use of registers Moreover if a signal handling function returns the process resumes its original execution path with registers restored to their original values Thus programs and compilers may freely use all registers without danger of signal handlers changing their values A 2 2 Function Return Values Functions Returning Scalars or No Value e A function that returns an integral or pointer value places its result in the next available register of the sequence rax rdx e A function that returns a floating point value that fits in the XMM registers returns this value in the next available XMM register of the sequence xmm0 xmm1 e An X87 floating point return value appears on the top of the floating point stack in st 0 as an 80 bit X87 number If this X87 return value is a complex number the real part of the value is returned in st 0 and the imaginary part in st 1 e A function that returns a value in memory also returns the address of this memory in rax e Functions that return no value also called procedures or void functions put no particular value in any regist
56. and pghpf only noopenmp when used in combination with the mp option Miscellaneous causes the compiler to ignore OpenMP parallel ization directives or pragmas but still process SGI style parallelization directives or pragmas nopgdllmain do not link the module containing the default DII Miscellaneous Main into the DLL Windows only norpath On Linux do not add rpath paths to the link line Miscellaneous nosgimp when used in combination with the mp option Miscellaneous causes the compiler to ignore SGI style parallel ization directives or pragmas but still process OpenMP directives or pragmas nostartup do not link in the standard startup routine pgf77 Environment pgf95 and pghpf only nostddef instructs the compiler to not recognize the stan Environment dard preprocessor macros nostdinc instructs the compiler to not search the standard Environment location for include files nostdlib instructs the linker to not link in the standard Environment libraries noonetrip determines whether each DO loop executes at Language least once pgf77 pgf95 and pghpf only novintr disable idiom recognition and generation of calls Optimization to optimized vector functions 58 Command Line Options pgflag Description Category pfi instrument the generated code and link in libraries for dynamic collection of profile and data infor mation at runtime Optimization pfo read a pgfi out trace fil
57. annotate the generated assembly code with source code when either the S or Mkeepasm options are used enables array bounds checking If an array is an assumed size array the bounds checking only applies to the lower bound If an array bounds violation occurs during execution an error message describing the error is printed and the program terminates The text of the error message includes the name of the array the location where the error occurred the source file and the line number in the source and information about the out of bounds subscript its value its lower and upper bounds and its dimension For example PGFTN F Subscript out of range for array a a f 2 subscript 3 lower bound 1 upper bound 2 dimension 2 disables array bounds checking swap byte order from big endian to little endian or vice versa upon input output of Fortran unformatted data files instructs the compiler to check for internal consistency of the x87 floating point stack in the prologue of a function and after returning from a function or subroutine call Floating point stack corruption may occur in many ways one of which is Fortran code calling floating point functions as subroutines Oe with the CALL statement If the PGI_CONTINUE environment variable is set upon execution of a program compiled with Mchkfpstk the stack will be automatically cleaned up and execution will continue There is a performance penalty associated with the stack cleanu
58. as not found See the online man page for Idd for more information on options and usage Libraries and Environment Variables 8 3 Creating and Using Dynamic Link Libraries on Win32 Some of the PGI compiler runtime libraries are available in both static library and dynamic link library DLL form for Windows There are several differences between these two types of libraries Both libraries are used when resolving external references when linking an executable but the process differs for each type of library When linking with a static library the code needed from the library is incorporated into the executable Once the executable has been built the library is no longer needed the executable does not rely on the static library at runtime When linking with a DLL external references are resolved using the DLL s import library not the DLL itself The code in the DLL associated with the external references does not become a part of the executable The DLL is loaded when the executable that needs it is run For the DLL to be loaded in this manner the DLL must be in your path Windows directory or Windows systems directory Static libraries and DLLs also handle global data differently If two static libraries contain global data with the same name and both libraries are linked to the executable the global data item in the libraries will be resolved to the same memory location If this situation occurs with two DLLs however the globa
59. as a type It s not necessary to mangle the name of the local class itself unless it s also a nested class 11 2 4 Template Class Name Mangling Template classes have mangled names that encode the arguments of the template template lt class T1 class T2 gt class abc abe lt int int gt x abc pt 3 ii This describes two template arguments of type int with the total length of template argument list string including the underscore and a fixed string indicates parameterized type as well the name of the class template C Name Mangling 207 208 C Name Mangling Appendix A Run time Environment This appendix describes the programming model supported for compiler code generation including register conventions and calling conventions for x86 and x64 processor based systems Section Al addresses these conventions for processors running linux86 or Win32 operating systems section A2 for processors running linux86 64 operating systems and section A3 for processors running Win64 operating systems Al Linux86 and Win32 Programming Model This section defines compiler and assembly language conventions for the use of certain aspects of an x86 processor running a linux86 or Win32 operating system These standards must be followed to guarantee that compilers application programs and operating systems written by different people and organizations will work together The conventions supported by the PGCC ANSI C compiler implement the appl
60. chunk is specified a default chunk size is chosen equal to 1 When schedule guided chunk is specified the chunk size is reduced in an exponentially decreasing manner with each dispatched piece of the iteration space Chunk specifies the minimum number of iterations to dispatch each time except when there are less than chunk iterations remaining to be processed at which point all remaining iterations are assigned If no chunk is specified a default chunk size is chosen equal to 1 When schedule runtime is specified the decision regarding iteration scheduling is deferred until runtime The schedule type and chunk size can be chosen at runtime by setting the OMP_SCHEDULE environment variable If this environment variable is not set the resulting schedule is equivalent to schedule static omp barrier The OpenMP omp barrier pragma uses the following syntax 132 OpenMP Pragmas for C and C pragma omp barrier There may be occasions in a parallel region when it is necessary that all threads complete work to that point before any thread is allowed to continue The omp barrier pragma synchronizes all threads at such a point in a program Multiple barrier points are allowed within a parallel region The omp barrier pragma must either be executed by all threads executing the parallel region or by none of them 6 8 omp parallel for The omp parallel for pragma uses the following syntax pragma omp parallel for clauses lt C C
61. complex classes or structures All classes are laid out in the same general way using the following pattern in the sequence indicated e First storage for all of the direct base classes which implicitly includes storage for non virtual indirect base classes as well e When the direct base class is also virtual only enough space is set aside for a pointer to the actual storage which appears later e In the case of a non virtual direct base class enough storage is set aside for its own non virtual base classes its virtual base class pointers its own fields and its virtual function information but no space is allocated for its virtual base classes Next storage for the class s own fields e Next storage for virtual function information typically a pointer to a virtual function table e Finally storage for its virtual base classes with space enough in each case for its own non virtual base classes virtual base class pointers fields and virtual function information 9 2 4 Aggregate Alignment The alignment of an array a structure or union an aggregate affects how much space the object occupies and how efficiently the processor can address members Arrays use the alignment of their members Arrays align according to the alignment of the array elements For example an array of short data type aligns on a 2 byte boundary Structures and Unions align according to the most restrictive alignment of the enclosing me
62. data files produced by most RISC workstations and high end servers to the little endian format used on x86 or x64 systems on the fly during file reads writes This option assumes that the record layouts of unformatted sequential access and direct access files are the same on the systems Also the assumption is that the IEEE representation is used for floating point numbers In particular the format of unformatted data files produced by PGI Fortran compilers is identical to the format used on Sun and SGI workstations that allows you to read and write unformatted Fortran data files produced on those platforms from a program compiled for an x86 or x64 platform using the byteswapio option Default The compiler does not byte swap data on input output Usage The following command line requests byte swapping are performed on input output pgf95 byteswapio myprog f C Enables array bounds checking If an array is an assumed size array the bounds checking only applies to the lower bound If an array bounds violation occurs during execution an error message describing the error is printed and the program terminates The text of the error message includes the name of the array the location where the error occurred the source file and the line number in the source and information about the out of bounds subscript its value its lower and upper bounds and its dimension Default The compiler does not enable array bounds checking Usage
63. data item while val refers to the value of that data item Subroutine and function names are converted into symbol names according to the rules outlined in table 10 3 Consider the following subroutine call call work ERR a b n Inter language Calling 201 where a is a double precision scalar b is a real vector of size n and n is an integer Default The symbol name for the subroutine is constructed by pre pending an underscore converting to all upper case and appending an sign followed by an integer indicating the total number of bytes occupied by the argument list Byte counts for character arguments appear immediately following the corresponding argument in the argument list The following is an example of the pseudo code for the above call using Default conventions call _WORK 20 addr ERR 3 addr a addr b addr n STDCALL The symbol name for the subroutine is constructed by pre pending an underscore converting to all lower case and appending an sign followed by an integer indicating the total number of bytes occupied by the argument list Character strings are truncated to the first character in the string which is passed by value as the first byte in a 4 byte word The following is an example of the pseudo code for the above call using STDCALL conventions call _work 20 val E val a addr b val n Note that in this case there are still 20 bytes in the argument list
64. different IPA information will be regenerated which saves on compile time 2 7 7 Building a Program with IPA Using Make As in the previous two sections programs can be built with IPA using the make utility just by adding the Mipa command line switch OPT Mipa fast a out filel o file2 o file3 o pgce OPT o a out filel o file2 o file3 o filel o filel c pgce OPT c filel c file2 o file2 c pgcce OPT c file2 c file3 o0 file3 c pgce OPT c file3 c The single command o make will invoke the compiler to generate any object files that are out of date then invoke pgcc to link the objects into the executable at link time pgcc will call the IPA linker to regenerate any stale or invalid IPA optimized objects 2 7 8 Questions about IPA Why is the object file so large An object file created with Mipa contains several additional sections One is the summary information used to drive the interprocedural analysis In addition the object file contains the compiler internal representation of the source file so the file can be recompiled at link time with interprocedural optimizations There may be additional information when inlining is enabled The total size of the object file may be 5 10 times its original size The extra sections are not added to the final executable What if I compile with Mipa and link without Mipa The PGI compilers generate a legal object file even when the source file is compil
65. ele UL DEE 81 Mnoasmkevword AA 69 Mnobackelach s nnnnnnnnnennsennsnnnnnnnrenssssnnnnnne 67 Mnobounde oonnnnnnnnnnenssesnnnnnnnnnnnnesssnnnnnne 79 MNO AZ eina a iana 61 MnodclChk 22cccccccceceeeeeeeseseeeeeeeeeeeeeeees 67 Mnodeiaultuntt 67 MnoOde pen isise 72 MiNOGIINGS 3 a vancscdacvesicscdesdoaseds cecversbiadddoves saad 67 SAPALO EE 62 Mnofprelaxed 72 Mnoframe 14 MNOIA4 EE 75 Mnoiomutenx 67 Mnolarge_arrays nsee 62 S Vulle E GE 82 AMNMOIG iis och des iao Sela coves 74 SMMOMAIN Sisto ss scan lechatass tan Segerstng ENEE 62 303 Mnontemporal AAA 62 MNOONCUI Div aria eriin eres 67 Mnoopenmp 82 Mnopgdllmain A 82 Mnoprefetch ccescceceseeeeeeesereteeeeeeees 75 Elle secs ieee A Nes 75 MNOFBintriNSiCs cceceeeeeteeeeteeeteeeeeees 76 Mnorecursive cceeeeeeeeeeeeeeteeeeeeeeeeeeeeeees 63 Mnoreentrant ceceeeeceeeeeteeeeeeeeeeteeeeeees 63 Mnoref evtemals eenen eneee 63 MNOSAVC seien egen 67 Mnoscalarsese reene 76 Mnosecond_underscore sssseeeseeseeeeee 63 MMOSQIMD haehae 82 Mnosignextend AAA 63 MNOSINGIC 55 csececc ccs cceceeescececttce seen eesti eee coe 69 Bleu 76 e Mnostartlp chi aca naive 64 Mnosdet AA 64 HMnostell Din eis bs efor a posses cia teases earar kE aeni 65 Mnostridet 64 Mnoumntsloocel AAA 68 Mnounroll ia rerien aeii 77 Mnoupcase 68 Eder 78 SC LL 78 Seel 83 SMOMOU BEE 67 AUP EE 84 MD fic dee
66. example the compiler inlines functions with fewer than approximately 100 statements in the source file myprog f and writes the executable code in the default output file a out pgf95 Minline size 100 myprog f Refer to Chapter 3 Command Line Options for more information on the Minline options 4 1 1 Using an Inline Library If you specify one or more inline libraries on the command line with the Minline option the compiler does not perform an initial extract pass The compiler selects functions to inline from the specified inline library If you also specify a size or function name all functions in the inline library meeting the selection criteria are selected for inline expansion at points in the source text where they are called If you do not specify a function name or a size limitation for the Minline option the compiler inlines every function in the inline library that matches a function in the source text In the following example the compiler inlines the function proc from the inline library lib il and writes the executable code in the default output file a out pgf95 Minline name proc lib lib il myprog f The following command line is equivalent to the line above the only difference in this example is that the name and lib inline keywords are not used The keywords are provided so you can avoid name conflicts if you use an inline library name that does not contain a period Otherwise without the keywords
67. extended precision values This can lead to answers when rounded that do not match expected results For example if the argument to sin is the result of previous calculations performed on the floating point stack then an 80 bit value used instead of a 64 bit value can result in slight discrepancies Results can even change sign due to the sin curve being too close to an x intercept value when evaluated To maintain consistency in this case you can assure that the compiler generates code that calls a function According to the x86 ABI a function call must push its arguments on the stack in this way memory is guaranteed to be accessed even if the argument is an actual constant Thus even if the called function simply performs the inline expansion using the function call as a wrapper to sin has the effect of trimming the argument precision down to the expected size Using the Mnobuiltin option on the command line for C accomplishes this task by resolving all math routines in the library libm performing a function call of necessity The other method of generating a function call for math routines but one that may still produce the inline instructions is by using the Kieee switch A second example illustrates the precision control problem using a section of code to determine machine precision program find precision w 1 0 100 W W W y w 1 Z y w if z gt 0 goto 100 C now w is just big enough that w 1 w 1 gt 1 print w e
68. f contains a number Function Inlining 103 of small functions used in the files parser f and alloc f The makefile also maintains the inline library utils il The makefile updates the library whenever you change utils f or one of the include files it uses In turn the makefile compiles parser f and alloc f whenever you update the library Example 4 1 Sample Makefile SRC mydir FC pgf95 FFLAGS 02 main o SRC main f SRC global h S FC S FFLAGS c SRC main f utils o SRC utils f SRC global h SRC utils h FC S FFLAGS c SRC utils f utils il SRC utils f SRC global h SRC utils h FC S FFLAGS Mextract 15 o utils il parser o SRC parser f SRC global h utils il FC S FFLAGS Minline utils il c SRC parser f alloc o SRC alloc f SRC global h utils il FC S FFLAGS Minline utils il c SRC alloc f myprog main o utils o parser o alloc o FC o myprog main o utils o parser o alloc o 4 3 Error Detection during Inlining To request inlining information from the compiler when you invoke the inliner specify the Minfo inline option For example pgf95 Minline mylib il Minfo inline myext f 4 4 Examples Assume the program dhry consists of a single source file dhry f The following command line builds an executable file for dhry in which proc7 is inlined wherever it is called pgf95 dhry f Minline proc7 The following command lines b
69. first argument is an integral type and its second argument is a floating point type the first argument will be passed in the first general purpose register rcx and the second argument will be passed in the second XMM register xmm1 the first XMM register and second general purpose register are ignored Arguments after the first four are passed on the stack 234 Run time Environment Integral and pointer type return values are returned in rax Floating point return values are returned in xmm0 Additional registers with assigned roles in the standard calling sequence rsp rbp The stack pointer holds the limit of the current stack frame which is the address of the stack s bottom most valid word The stack pointer should point to a 16 byte aligned area unless in the prolog or a leaf function The frame pointer if used can provide a way to reference the previous frame on the stack Details are implementation dependent A function must preserve this register value for its caller MXCSR The flags register MXCSR contains the system flags such as the direction x87 flag and the carry flag The six status flags MXCSR 0 5 are volatile the remainder of the register is nonvolatile Floating Point Control Word FPCSR The control word contains the floating point flags such as the rounding mode and exception masking This register is initialized at process initialization time and its value must be preserved Signals can in
70. for S079 Keyword form of argument illegal in this context for S080 Subscript for array is out of bounds 081 Illegal selector Messages 257 082 Illegal substring expression for variable Substring expressions must be of type integer and if constant must be greater than zero S083 Vector expression used where scalar expression required A vector expression was used in an illegal context For example iscalar iarray where a scalar is assigned the value of an array Also character and record references are not vectorizable 084 Illegal use of symbol This message is used for many different errors 085 Incorrect number of arguments to statement function S086 Dummy argument to statement function must be a variable S087 Non constant expression where constant expression required S088 Recursive subroutine or function call of A function may not call itself S089 Illegal use of symbol with character length Symbols of type CHARACTER must be dummy variables and must not be used as statement function dummy parameters and statement function names Also a dummy variable of type CHARACTER cannot be used as a function S090 Hollerith constant more than 4 characters In certain contexts Hollerith constants may not be more than 4 characters long S091 Constant expression of wrong data type 092 Illegal use of variable length character expression A character expression used as an actual arg
71. i ot n 2 k n i n 2 3 continue If the outer loop is parallelized conflicting values will be stored into k by the various processors The variable k cannot be made local to each processor because the value of k must remain coherent among the processors It is possible the loop could be parallelized if all assignments to k are placed in 28 Optimization amp Parallelization critical sections However it is not clear where critical sections should be introduced because in general the value for k could depend on another scalar or on k itself and code to obtain the value of other scalars must reside in the same critical section In the example above the assignment to k within a conditional at label 2 prevents k from being recognized as an induction variable If the conditional statement at label 2 is removed k would be an induction variable whose value varies linearly with j and the loop could be parallelized 2 6 2 4 Scalar Last Values During parallelization scalars within loops often need to be privatized that is each execution thread will have its own independent copy of the scalar Problems can arise if a privatized scalar is accessed outside the loop For example consider the following loop 1 Lech i for f x i gt 5 0 t x i i if we t The value of t may not be computed on the last iteration of the loop Normally if a scalar is assigned within a loop and used following the loop the PGI com
72. int al int el void f1 class A int al void f1 friend void f int il al cfront uses global al f1 cfront uses global f1 ja Only the innermost class scope is incorrectly skipped by cfront as illustrated in the following example int al int bl struct A static int al class B static int bl friend void int il al cfront uses A al 300 C Dialect Supported int jl b1 cfront uses global b1 e operator may be declared as a nonmember function This is flagged as an anachronism by cfront 2 1 e A type qualifier is allowed but ignored on the declaration of a constructor or destructor For example class A A const No error in cfront 2 1 mode C 6 cfront 2 1 3 0 Compatibility Mode The following extensions are accepted in both cfront 2 1 and cfront 3 0 compatibility mode i e these are features or problems that exist in both cfront 2 1 and 3 0 e Type qualifiers on the this parameter may to be dropped in contexts such as this example struct A void const void A fp amp A f This is actually a safe operation A pointer to a const function may be put into a pointer to non const because a call using the pointer is permitted to modify the object and the function pointed to will actually not modify the object The opposite assignment would not be safe Conversion operators specifying conversion to void are allow
73. means code compiled to target a given processor will not necessarily execute correctly on a previous generation processor The most important processor types along with a list of the features utilized by the PGI compilers that distinguish them from a compatibility standpoint are listed in the following table Preface XV Table P 2 Processor Options Processor Prefetch SSE SSE SSE 32 64 Scalar FP 1 2 3 bit bit Default AMD Athlon X x87 AMD Athlon XP MP X X X x87 AMD Athlon64 X X X X X SSE AMD Opteron X X X X X SSE AMD Opteron Rev E X X X X X SSE AMD Turion X X X X X SSE Intel Celeron X x87 Intel Pentium H X x87 Intel Pentium III x x x x87 Intel Pentium 4 x x X X SSE Intel Pentium M X X X X SSE Intel Centrino X X X X SSE Intel Pentium 4 EM64T X X X X SSE Intel Xeon EM64T X X X X SSE Intel Core 2 EM64T X X X X SSE In this manual the convention is to use x86 to specify the group of processors in the previous table that are listed as 32 bit but not 64 bit The convention is to use x64 to specify the group of processors that are listed as both 32 bit and 64 bit x86 processor based systems can run only 32 bit operating systems x64 processor based systems can run either 32 bit or 64 bit operating systems and can execute all 32 bit x86 binaries in either case x64 processors have additional registers and 64 bit addressing capabilit
74. needed subject to the restrictions listed in the description of each clause For the purposes of the OpenMP pragmas a C C structured block is defined to be a statement or compound statement a sequence of statements beginning with and ending with that has a single entry and a single exit No statement or compound statement is a C C structured block if there is a jump into or out of that statement The compiler option mp enables recognition of the parallelization pragmas The use of this option also implies Mreentrant local variables are placed on the stack and optimizations that may result in non reentrant code are disabled e g Mnoframe 125 Also note that calls to I O library functions are system dependent and are not necessarily guaranteed to be thread safe I O library calls within parallel regions should be protected by critical regions see below to ensure they function correctly on all systems In the examples given with each section the functions omp_get_num_threads and omp_get_thread_num are used refer to Section 6 15 Run time Library Routines They return the number of threads currently in the team executing the parallel region and the thread number within the team respectively 6 2 omp parallel The OpenMP omp parallel pragma uses the following syntax Syntax pragma omp parallel clauses lt C C structured block gt Clauses private list shared list default shared none
75. of the currently executing driver If a full pathname is supplied that file is used for the driver configuration file Syntax rc path filename Where path is either a relative pathname relative to the value of DRIVER or a full pathname beginning with Filename is the driver configuration file Default The driver uses the configuration file pgirc Usage In the following example the file pgf95rctest relative to usr pgi linux86 bin the value of DRIVER is the driver configuration file pgf95 rc pgf95rctest myprog f Cross reference show S Stops compilation after the compiling phase and writes the assembly language output to the file filename s where the input file is filename f Default The compiler produces an executable file Usage In this example pgf95 produces the file myprog s in the current directory pgf95 S myprog f Cross reference c P F Mkeepasm o shared Valid only on Linux and is passed to the linker Instructs the linker to produce a shared object dynamically linked library file Cross reference fpic G R show Produce driver help information describing the current driver configuration 90 Command Line Options Usage In the following example the driver displays configuration information to the standard output after processing the driver configuration file pgf95 show myprog f Cross reference V v help rc silent
76. of the basic block containing the use of the variable W280 Syntax error in directive messages 280 300 rsvd for directive handling W281 Directive ignored S300 Too few data constants in initialization of derived type S301 must be TEMPLATE or PROCESSOR S302 Unmatched ENDS statement S303 END statement for required in an interface block 304 EXIT CYCLE statement must appear in a DO DOWHILE loop S305 cannot be named S306 names more than one construct S307 must have the construct name 308 DO may not terminate at an EXIT CYCLE RETURN STOP GOTO or arithmetic IF Messages 275 309 Incorrect name specified in END statement 310 Generic message for MODULE errors W311 Non replicated mapping for array ignored W312 Array should be declared SEQUENCE W313 Subprogram called within INDEPENDENT loop not PURE E314 IPA actual argument is a label but dummy argument is not an asterisk The call passes a label to the subprogram the corresponding dummy argument in the subprogram should be an asterisk to declare this as the alternate return I315 IPA routine constant dummy arguments This many dummy arguments are being replaced by constants due to interprocedural analysis I316 IPA routine INTENT IN dummy arguments This many dummy arguments are being marked as INTENT IN due to interprocedural analysis I317 IPA routine array alignments propagated
77. only Override the normal error severity of the specified diagnostic messages The message s may be specified using a mnemonic error tag or using an error number diag_warning tag pgCC only Override the normal error severity of the specified diagnostic messages The message s may be specified using a mnemonic error tag or using an error number diag_error tag pgCC only Override the normal error severity of the specified diagnostic messages The message s may be specified using a mnemonic error tag or using an error number display_error_number pgCC only Display the error message number in any diagnostic messages that are generated The option may be used to determine the error number to be used when overriding the severity of a diagnostic message no_ exceptions pgCC only Enable disable exception handling support The default is exceptions Command Line Options 97 no llalign pgCC only Do don t align long long integers on long long boundaries The default is Ialign M Generate a list of make dependencies and print them to stdout Compilation stops after the pre processing phase MD Generate a list of make dependencies and print them to the file lt file gt d where lt file gt is the name of the file under compilation dependencies_file lt file gt optk_allow_dollar_in_id_chars pgCC only Accept dollar signs in identifiers P Stops compilation after the preprocessing p
78. onto the stack in the same manner as integral arguments described above This provides call by value semantics letting the called function modify its arguments without affecting the calling function s object In the example below the argument s is a structure consisting of more than 2 words Run time Environment 215 Table A 6 Structure and Union Arguments Call Argument Stack Address i 1 s 1 8 ebp word 0 s 12 ebp word 1 s 16 ebp Implementing a Stack In general compilers and programmers must maintain a software stack Register esp is the stack pointer Register esp is set by the operating system for the application when the program is started The stack must be a grow down stack A separate frame pointer enables calls to routines that change the stack pointer to allocate space on the stack at run time e g alloca Some languages can also return values from a routine allocated on stack space below the original top of stack pointer Such a routine prevents the calling function from using esp relative addressing to get at values on the stack If the compiler does not call routines that leave esp in an altered state when they return a frame pointer is not needed and is not used if the compiler option Mnoframe is specified Although not required the stack should be kept aligned on 8 byte boundaries so that 8 byte locals are favorably aligned with respect to performance PG
79. passed on the stack which results in a memory store and load Finally Kieee also disables reciprocal division for constant divisors That is for a b with unknown a and constant b the expression is usually converted at compile time to a 1 b thus turning an expensive divide into a relatively fast scalar multiplication However numerical discrepancies can occur when this optimization is used Understanding and correctly using the pc Mnobuiltin and Kieee switches should enable you to produce the desired and expected precision for calculations which utilize floating point operations Usage pgf95 pc 64 myprog c ps Linux only Instructs the compiler to instrument the generated executable for gprof style sample based profiling Must be used at both the compile and link steps A gmon out style trace is generated when the resulting program is executed and and can be analyzed using gprof or pgprof Syntax P8 Default The compiler does not instrument the generated executable for gprof style profiling Selects variations for compilation There are four uses for the Q option Syntax Qdirdirectory The first variety using the dir keyword lets you supply a directory parameter that indicates the directory where the compiler driver is located Qoptionprog opt 88 Command Line Options The second variety using the option keyword lets you supply the option opt to the program prog The prog parameter can be one
80. register usage and parameter passing Register Usage Conventions The following table defines the standard for register allocation The x64 Architecture provides a variety of registers All the general purpose registers XMM registers and x87 registers are global to all procedures in a running program 218 Run time Environment Table A 7 Register Allocation Type Name Purpose General rax st return register Jorbx callee saved optional base pointer orcx pass 4th argument to functions rdx pass 3rd argument to functions 2nd return register rsp stack pointer rbp callee saved optional stack frame pointer Prsi pass 2nd argument to functions rdi pass 1st argument to functions r8 pass 5th argument to functions r9 pass 6th argument to functions r10 temporary register pass a function s static chain pointer rll temporary register r12 r15 callee saved registers XMM xmm0 xmm1 pass and return floating point arguments xmm2 xmm7 pass floating point arguments xmmsg xmm15 temporary registers x87 st 0 temporary register return long double arguments st 1 temporary register return long double arguments st 2 P st 7 temporary registers In addition to the registers each function has a frame on the run time stack This stack grows downward from high addresses The next table shows the stack frame organization Run time Environment 219
81. routines and environment variables that can be used to specify shared memory parallelism in Fortran C and C programs The directives include a parallel region construct for writing coarse grain SPMD programs work sharing constructs which specify that DO loop iterations should be split among the available threads of execution and synchronization constructs The data environment is controlled using clauses on the directives or with additional directives Run time library routines are provided to query the parallel runtime environment for example to determine how many threads are participating in execution of a parallel region Finally environment variables are provided to control the execution behavior of parallel programs For more information on OpenMP see http www openmp org For an introduction to how to execute programs that use multiple processors along with some pointers to example code see Section 1 4 Parallel Programming Using the PGI Compilers The file ftp ftp pgroup com pub SMP fftpde tar gz contains a more advanced self guided tutorial on how to parallelize the NAS FT fast Fourier transform benchmark using OpenMP directives You can retrieve it using a web browser and unpack it using the following commands within a shell command window gunzip fftpde tar gz tar xvf fftpde tar d i Follow the instructions in the README file to work through the tutorial 5 1 Parallelization Directives Parallelization directive
82. selected For compatibility with other vendors directives the prefix cpgi may be substituted with cdir or cvd 141 7 2 Fortran Directive Summary The next table summarizes the supported Fortran directives The scope entry indicates the allowed scope indicators for each directive the default scope is surrounded by parentheses The system field indicates the target system type for which the pragma applies Many of the directives can be preceded by NO The default entry in the table indicates the default of the directive n a appears if a default does not apply The name of a directive may also be prefixed with M for example the directive Mbounds is equivalent to bounds and Mopt is equivalent to opt 142 Optimization Directives and Pragmas Table 7 1 Fortran Optimization Directive Summary Directive Function Default Scope altcode noaltcode Do don t generate alternate code for vec altcode Drg torized and parallelized loops assoc noassoc Do don t perform associative transforma assoc Drg tions bounds nobounds Do don t perform array bounds checking nobounds r g cncall nocncall Loops are considered for parallelization nocncall Drg even if they contain calls to user defined subroutines or functions or if their loop counts do not exceed usual thresholds concur noconcur Do don t enable auto concurrentization of concur Drg loops depchk nodepchk Do don
83. sonsonen 75 EM DO DE 75 Mprefeteh ves ccc esse periaos ieina 75 MpreprocesSs ccsescceeeseeeeeeeeeesseneeescees 82 sMprOf EE 62 EMI Be gue eege E ee 75 MIrBintrinSiCs e ccccceeeeeesteeeeeeeeneeeeees 75 Mrecursive sees dek tdeee adage denen eens 63 Mreentrannt ccceccccceeeeeeeeteeeeeeeeeneeeeeees 63 Mref_externals ccceeececceeeeeeseeeeeees 63 Masate Jastbva eeren 63 SMS ate pii aae ie rerin sapada iaa iatera aste 76 SMSAVG ties eege a A e 67 Macalarese A 76 E Een e 69 Msecond underecore ssnsnnnsnnnneneeeeeen 63 Msignextend cceecccceeseeeeesseeeeeereeeseners 63 eMsingl ie see Sid tie eat eee 69 SMSMart zua egene eet de dE e 76 Gd Ee Ee DEEN 68 SMStrid Os 1 4 scsi eh eaaa a 63 Muchas cel Ee See dee 69 Gd GE 64 Muntslogieal AA 68 MUDCASC dees tial cantante 68 SC Te EE 64 Ne Sik at teh ape genee Ue ota had 77 e DEEN 85 88 optk_allow_dollar_in_id_chars 9 304 E 98 e RE 86 LEE 98 SOC EE 99 PFOINCIUAG s ccieaieeeegereesteceeeieecdeseeeeestraeesuees 99 SE EE 88 SP WEE 89 gll EE 89 SET 90 BO EE E 90 Shared ee eeo e R e it Ede 90 le UE 90 SE ieee tee event Seta ier 91 RE 99 SUING EE 91 PUD EE 91 SE 93 USO e RE 99 NM EE 93 e dE EE EE 93 PV ech ee E 94 Mee Eed Ed vice AE 94 Commandline Options SVMAX te Geers eed teased A laensts 2 Compilation driver ceeeeeceeeseeeeeeeseeeeneeeeeeee 1 Compilers Invo
84. standard search path for include files Command Line Options 51 Usage The Fortran INCLUDE statement directs the compiler to begin reading from another file The compiler uses two rules to locate the file 1 If the file name specified in the INCLUDE statement includes a path name the compiler begins reading from the file it specifies 2 If no path name is provided in the INCLUDE statement the compiler searches in order e any directories specified using the I option in the order specified e the directory containing the source file e the current directory For example the compiler applies rule 1 to the following statements INCLUDE bob include file1 absolute path name INCLUDE filel relative path name and rule 2 to this statement INCLUDE filel Cross reference Mnostdinc i2 i4 and i8 Treat INTEGER and LOGICAL variables as either two four or eight bytes INTEGER 8 values not only occupy 8 bytes of storage but operations use 64 bits instead of 32 bits K lt flag gt Requests that the compiler provide special compilation semantics Syntax K lt flag gt Where flag is one of the following ieee Perform floating point operations in strict conformance with the IEEE 754 standard Some optimizations are disabled and on some systems a more accurate math library is linked if Kieee is used during the link step noieee Use the fastest available means to perform floating poin
85. subsection of code within a parallel region referred to as a critical section which will be executed one thread at a time The first thread to arrive at a critical section will be the first to execute the code within the section The second thread to arrive will not begin execution of statements in the critical section until the first thread has exited the critical section Likewise each of the remaining threads will wait its turn to execute the statements in the critical section An optional name may be used to identify the critical region Names used to identify critical regions have external linkage and are in a name space separate from the name spaces used by labels tags members and ordinary identifiers Critical sections cannot be nested and any such specifications are ignored Branching into or out of a critical section is illegal include lt stdlib h gt main int a 100 100 mx 1 Imx 1 i j for j 0 j lt 100 j for i 0 i lt 100 i a i j 1 int 10 0 rand RAND _MAX 1 0 pragma omp parallel private i firstprivate 1mx pragma omp for for j 0 j lt 100 j for i 0 i lt 100 i lmx lmx gt ali j lm a i j pragma omp critical mx mx gt lmx mx lmx printf max value of a is d n mx 6 4 omp master The OpenMP omp master pragma uses the following syntax pragma omp master lt C C structured block gt In a parallel region of code there
86. supports SSE or SSE2 instructions SSE operations can be used to operate on pairs of single precision floating point numbers and do not apply to double precision floating point numbers SSE2 instructions can be used to operate on quads of single precision floating point numbers or on pairs of double precision floating point numbers Loops vectorized using SSE or SSE2 instructions operate much more efficiently when processing vectors that are aligned to a cache line boundary You can cause unconstrained data objects of size 16 bytes or greater to be cache aligned by compiling with the Mcache_align switch An unconstrained data object is a data object that is not a common block member and not a member of an aggregate data structure Note In order for stack based local variables to be properly aligned the main program or function must be compiled with Mcache_align Optimization amp Parallelization 23 The Mcache_align switch has no effect on the alignment of Fortran allocatable or automatic arrays If you have arrays that are constrained for example vectors that are members of Fortran common blocks you must specifically pad your data structures to ensure proper cache alignment Mcache_align causes only the beginning address of each common block to be cache aligned The following examples show results of compiling the example code with and without Mvect sse Example 2 3 Vector operation using SSE instructions program vector
87. the directive prefix may contain upper or lower case letters case is not significant Case is significant for any variable names that appear in the body of the directive if the command line option Mupcase is selected An example using prefetch directives to prefetch data in a matrix multiplication inner loop where a row of one source matrix has been gathered into a contiguous vector might look as follows real 8 a m n b n p c m p arow n do j 1 p cSmem prefetch arow 1 b 1 j cSmem prefetch arow 5 b 5 j cSmem prefetch arow 9 b 9 j do k 1 n 4 cSmem prefetch arow k 12 b k 12 j c i j c i j arow k b k j e i j c i j arow k 1 b k 1 j e i j c i j arow k 2 b k 2 j c i j c i j arow k 3 b k 3 j enddo This pattern of prefetch directives will cause the compiler to emit prefetch instructions whereby elements of arow and b are fetched into the data cache starting 4 iterations prior to first use By varying the prefetch distance in this way it is possible in some cases to reduce the effects of main memory latency and improve performance 158 Optimization Directives and Pragmas Chapter 8 Libraries and Environment Variables This chapter discusses issues related to PGI supplied compiler libraries It also addresses the use of C C builtin functions in place of the corresponding libc routines creation of dynamically linked libraries also known as shared objects or shared libraries
88. the vectorizer to assume that the maximum vector length is less than or equal to n The vectorizer uses this information to eliminate generation of the stripmine loop for vectorized loops wherever possible If the size n is omitted the default is 100 Note No space is allowed on either side of the colon sse Instructs the vectorizer to search for vectorizable loops and where possible make use of SSE SSE2 and prefetch instructions Mnovect instructs the compiler not to perform vectorization can be used to override a previous instance of Mvect on the command line in particular for cases where Mvect is included in an aggregate option such as fastsse Mnovintr instructs the compiler not to perform idiom recognition or introduce calls to hand optimized vector functions Default For arguments that you do not specify the default optimization control options are as follows depchk noprefetch i4 nounroll nofprelaxed novect noipa nor8 nolre nor8intrinsics If you do not supply an option to Mvect the compiler uses defaults that are dependent upon the target system Usage In this example the compiler invokes the vectorizer with use of packed SSE instructions enabled pgf95 Mvect sse Mcache align myprog f Cross reference g O 78 Command Line Options M lt pgflag gt Syntax Manno Mbounds Mnobounds Mbyteswapio Mchkfpstk Mchkptr Mchkstk Miscellaneous Controls
89. whether IPA information has changed optimize references to global variables default is noglobals perform automatic function inlining If the optional n is provided limit inlining to at most n levels IPA based function inlining is performed from leaf routines upward save IPA information in a ipo file rather than incorporating it into the object file keep the optimized object files using file name mangling to reduce re compile time in subsequent builds default is keepobj optimize calls to certain standard C library routines default is nolibc allow inlining of routines from libraries implies Mipa inline default is nolibinline allow recompiling and optimization of routines from libraries using IPA information default is nolibopt equivalent to arg plus externalization of local pointer targets default is nolocalarg specify a function to appear as a global entry point may appear multiple times disables linking enable pointer disambiguation across procedure calls default is noptr pure function detection default is nopure 73 required return an error condition if IPA is inhibited for any reason rather than the default behavior of linking without IPA optimization safe lt function gt l lt library gt declares that the named function or all functions no safeall no shape summary no vestigial in the named library are safe a safe procedure does not call back into the know
90. 2003 x64 Editions running on an x64 processor based system collectively all Win32 and Win64 platforms supported by the PGI compilers The following table lists the PGI compilers and tools and their corresponding commands xiv Preface Table P 1 PGI Compilers and Commands Compiler or Tool Language or Function Command PGF77 FORTRAN 77 pgf77 PGF95 Fortran 90 95 pgf95 PGHPF High Performance Fortran pghpf PGCC C ANSI C99 and K amp R C pgcec PGC ANSI C with cfront features pgCC PGDBG Source code debugger pgdbg PGPROF Performance profiler pgprof In general the designation PGF95 is used to refer to The Portland Group s Fortran 90 95 compiler and pgf95 is used to refer to the command that invokes the compiler A similar convention is used for each of the PGI compilers and tools For simplicity examples of command line invocation of the compilers generally reference the pgf95 command and most source code examples are written in Fortran Usage of the PGF77 compiler whose features are a subset of PGF95 is similar Usage of PGHPF PGC and PGCC ANSI C99 is consistent with PGF95 and PGF77 but there are command line options and features of these compilers that do not apply to PGF95 and PGF77 and vice versa There are a wide variety of x86 compatible processors in use All are supported by the PGI compilers and tools Most of these processors are forward compatible but not backward compatible That
91. 6 A 3 4 Win64 Fortran Supplement ccccceeeeceeeeeeeeeeeeeeeeceaeeeeeeeeesseeeseneeeeaas 239 A 3 4 1 Fortran Fundamental Types ccecccccceeeesceeeeeeeetceeeeeeenaeeeeeeenaeeeeeneaa 240 A 3 4 2 Fortran Naming Conventions c cccccecceeseeeeeeeeeeeeeeeeececaeeseeeeeeseaeeneneees 241 A 3 4 3 Fortran Argument Passing and Return Conventions s es 241 A 3 4 4 Interlanguage Calling cccccccceeeeceeceeeeeeceeeeeeeaaeseeneeeseaaeeseeeeeessaeeneneees 241 Appendix B Messages ssssssssssnnunnnnennnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn nannan 247 B 1 Diagnostic MESSAGES eseina burii iv air ie i 247 B 2 Phase Invocation Messages eeeeceeceeeeeeneeeeeeeeeaeeeeeeeaaaeeeeetenaaeeeeenenaes 248 B 3 Fortran Compiler Error Messages cccecceeeseeeeeeeeeeeeeeeeeeeeeeseaaeeteneeees 248 B 3 1 Message Fonal ek 248 B 3 2 Message BCEE 248 B 4 Fortran Runtime Error Messages eeeeecceeeeeeneeeeeeeenaeeeeeeeeaaeeeeeeeaaas 290 B 4 1 Message Format ccceccscceeeeescneeeeeeecneeeeeeeeaeeeeeeeaaeeeeeeeeaeeeeeneaeeeeeenes 290 B 4 2 Message BCEE 290 Appendix C C Dialect Supported s ssssunsenunsnnnnennnnunnnnnnnnnnnnnnnnnnnnnnnnnennn nanna 295 Contents vii viii Anachronisms Accepted oreroraa a AS 295 New Language Features Accepted ccceeeeeeeeeeeeeeeecaeeeeeeeeeeeeeeeeeeeees 296 The following language features are not accepted ccceeeeeeeees 298 Extensions Accep
92. 9 PARALLEL WORKSHARE 1 1 11 118 userdirected ee ee 84 SECTIONS EE 118 Parallelization Directives c cccceereen 107 SING A e A ie tat oes sued Ee 113 Parallelization Pragmas NEE stile cere Pate Bee ea te 125 THREADPRIVATE oseese 121 Pragmas EE 116 EE 3 OpenMP Fortran Support Routines GEES 152 omp desto Jack EEN 123 SCOPE te Geshe Ae Needed he el ieee 155 OMP_Get_AYNAMIC EE 123 Prefetch directives seet 157 omp_get_max_threads EE 122 Preprocessor omp oel meter eee 123 EE eege ele ege 4 omp Oel mum Droen 122 e E 4 omp_get_num_threads AN 122 omp_get_thread_nNum eeen 122 R ie une e Run time Environment 0 209 omp In Dorallel AN 122 omp init EE het tas ates 123 S omp_set_dynamic ccccceeeseeeeeeeeeeeeeee 122 Shared object files ec eeeeeeeeeeeeseeeeeeees 159 omp Set lock aein miana 123 omp Set nesied 123 T omp Set num Threacdei 122 Timing OMP_teSt_lOCK erer 123 CPU CLOCK eara eana 38 omp_unset_lock ENEE 123 OXCCULION EE 38 E Le 141 SYSTEM CLOCK aranan annaa 38 C C pragmas eeeeeceeeeeeeteeeeeeeeees 37 151 Tools C C PragMAS sCope te 155 POD G e ee theaere se ak ida XV CACHE tIlING sseesesesseeeeseeeseetsseeesee tenets 77 PGPROF geen XV Fortran directives ccccccceessessseeeeeees 37 141 Fortran directives SCOPE En 148 v function inlining 0 02 2 ee eeeeeeeeeeeeteeee eee 14 101 raion global optimization h 14 17 Ne EE
93. 99 Note that you must explicitly link in the PGF95 runtime support libraries when linking pgf95 compiled program units into C or C main programs When linking pgf77 compiled program units into C or C main programs you need only link in Ipgftnrtl 10 14 Win32 Calling Conventions Aside from name mangling considerations in C the calling convention i e the symbol name to which the subroutine or function name is mapped and the means by which arguments are passed for C C is identical between most compilers on Win32 and Linux Win64 However Fortran calling conventions vary widely between legacy Win32 Fortran compilers and Linux Win64 Fortran compilers 10 14 1 Win32 Fortran Calling Conventions Four styles of calling conventions are supported using the PGI Fortran compilers for Win32 Default C STDCALL and UNIX e Default Default is the method used in the absence of compilation flags or directives to alter the default e C or STDCALL The C or STDCALL conventions are used if an appropriate compiler directive is placed in a program unit containing the call The C and STDCALL conventions are typically used to call routines coded in C or assembly language that depend on these conventions UNIX The UNIX convention is used in any Fortran program unit compiled using the Munix compilation flag The following table outlines each of these calling conventions 200 Inter language Calling Table 10 3 Calling Conventions Supp
94. A memory reference occurred whose address does not meet its data alignment requirement 200 Missing UNIT FILE specifier 201 Illegal I O specifier 202 Repeated I O specifier S203 FORMAT statement has no label s204 Miscellaneous I O error 205 Illegal specification of scale factor The integer following or has been omitted or P does not follow the integer value S206 Repeat count is zero 268 Messages 207 Integer constant expected in edit descriptor S208 Period expected in edit descriptor 209 Illegal edit descriptor 210 Exponent width not used in the Ew dEe or Gw dEe edit descriptors 211 Internal I O not allowed in this I O statement 212 Illegal NAMELIST I O Namelist I O cannot be performed with internal unformatted formatted and list directed I O Also I O lists must not be present 213 is not a NAMELIST group name 214 Input item is not a variable reference 215 Assumed sized array name cannot be used as an I O item or specifier An assumed sized array was used as an item to be read or written or as an I O specifier Ge FMT array name In these contexts the size of the array must be known 216 STRUCTURE UNION cannot be used as an I O item 217 ENCODE DECODE buffer must be a variable array or array element Messages 269 S218 Statement labeled S219 lt reserved message number gt S220 Redefining predefined macro S221 elif after else A preprocesso
95. DLL Libraries and Environment Variables 165 Mmakeimplib Generate an import library without generating a DLL Use this flag when you want to generate an import library for a DLL but are not yet ready to build the DLL itself This situation might arise for example when building DLLs with mutual imports see Example 4 below o lt file gt Passed to the linker Name the DLL or import library lt file gt def lt file gt When used with Mmakedll this flag is passed to the linker and a def file named lt file gt is generated for the DLL The def file contains the symbols exported by the DLL Generating a def file is not required when building a DLL but can be a useful debugging tool if the DLL does not contain the symbols that you expect it to contain When used with Mmakeimplib this flag is passed to lib which requires a def file to create an import library The def file can be empty if the list of symbols to export are passed to lib on the command line or explicitly marked as dllexport in the source code implib lt file gt Passed to linker Generate an import library named lt file gt for the DLL A DLL s import library is the interface used when linking an executable that depends on routines in a DLL To use the PGI compilers to create an executable that links to the DLL form of the runtime use the compiler flag Mdll The executable built will be smaller than one built without Mdll the PGI runtime DLLs howe
96. E PRECISION variables and constants respectively DOUBLE PRECISION elements are 8 bytes in length pgf77 pgf95 and pghpf only the compiler does not promote REAL variables and constants to DOUBLE PRECISION REAL variables will be single precision 4 bytes in length pgf77 and pgf95 only the compiler treats the intrinsics CMPLX and REAL as DCMPLX and DBLE respectively 75 Mnor8gintrinsics pgf77 and pgf95 only the compiler does not promote the intrinsics CMPLX and REAL to DCMPLX and DBLE respectively Msafeptr option option pgcc and pgCC only instructs the C C compiler to override data dependencies between pointers of a given storage class Possible values of option include all assume all pointers and arrays are independent and safe for aggressive optimizations and in particular that no pointers or arrays overlap or conflict with each other arg instructs the compiler that arrays and pointers are treated with the same copyin and copyout semantics as Fortran dummy arguments global instructs the compiler that global or external pointers and arrays do not overlap or conflict with each other and are independent local auto instructs the compiler that local pointers and arrays do not overlap or conflict with each other and are independent static instructs the compiler that static pointers and arrays do not overlap or conflict with each other and are independent Mscalarsse Use SSE SSE2 instructi
97. E variables but in addition are initialized from the original object existing prior to entering the parallel region Variables that appear in the list of a REDUCTION clause must be SHARED A private copy of each variable in list is created for each thread as if the PRIVATE clause had been specified Each private copy is initialized according to the operator as specified in the following table Table 5 1 Initialization of REDUCTION Variables Operator Intrinsic Initialization 0 1 0 AND TRUE OR FALSE EQV TRUE NEQV FALSE MAX Smallest Representable Number MIN Largest Representable Number IAND All bits on IOR 0 IEOR 0 At the end of the parallel region a reduction is performed on the instances of variables appearing in list using operator or intrinsic as specified in the REDUCTION clause The initial value of each REDUCTION variable is included in the reduction operation If the operator intrinsic portion of the REDUCTION clause is omitted the default reduction operator is addition 110 OpenMP Directives for Fortran The COPYIN clause applies only to THREADPRIVATE common blocks In the presence of the COPYIN clause data from the master thread s copy of the common block is copied to the threadprivate copies upon entry to the parallel region In the presence of an IF clause the parallel region will be executed in parallel only if the c
98. EAL 4 A 100 B 100 Z INTEGER I DO I 1 100 2 Z Z A i B i Z Z A i 1 B i 1 END DO END Using the Minfo option the compiler informs you when a loop is being unrolled For example a message indicating the line number and the number of times the code is unrolled similar to the following will display when a loop is unrolled dot 5 Loop unrolled 5 times Using the c lt m gt and n lt m gt sub options to Munroll or using Mnounroll you can control whether and how loops are unrolled on a file by file basis Using directives or pragmas as specified in Chapter 7 Optimization Directives and Pragmas you can precisely control whether and how a given loop is unrolled See Chapter 3 Command Line Options for a detailed description of the Munroll option 2 5 Vectorization using Mvect The Mvect option is included as part of fastsse on all x86 and x64 targets If your program contains computationally intensive loops the Mvect option may be helpful If in addition you specify Minfo and your code contains loops that can be vectorized the compiler reports relevant information on the optimizations applied When a PGI compiler command is invoked with the Mvect option the vectorizer scans code searching for loops that are candidates for high level transformations such as loop distribution loop interchange cache tiling and idiom recognition replacement of a recognizable code sequence suc
99. HPF User s Guide and The High Performance Fortran Handbook for more details on constructing and executing data parallel programs on shared memory or distributed memory cluster systems using PGHPF 8 Getting Started 1 5 Using the PGI Compilers on Linux 1 5 1 Linux Header Files The Linux system header files contain many GNU gcc extensions Many of these extensions are supported This should allow the PGCC C and C compilers to compile most programs compilable with the GNU compilers A few header files not interoperable with previous revisions of the PGI compilers have been rewritten and are included in PGI linux86 include These files are sigset h asm byteorder h stddef h asm posix_types h and others Also PGI s version of stdarg h should support changes in newer versions of Linux If you are using the PGCC C or C compilers please make sure that the supplied versions of these include files are found before the system versions This will happen by default unless you explicitly add a I option that references one of the system include directories 1 5 2 Running Parallel Programs on Linux You may encounter difficulties running auto parallel or OpenMP programs on Linux systems when the per thread stack size is set to the default 2MB If you have unexplained failures please try setting the environment variable MPSTKZ to a larger value such as 8MB This can be accomplished with the command setenv MPSTKZ 8M in csh or with
100. However rather than 5 4 byte quantities as in the Default convention there are 3 4 byte quantities and 1 8 byte quantity the double precision value of a C The symbol name for the subroutine is constructed by pre pending an underscore and converting to all lower case Character strings are truncated to the first character in the string which is passed by value as the first byte in a 4 byte word The following is an example of the pseudo code for the above call using C conventions call work val E Sval a addr b val n UNIX The symbol name for the subroutine is constructed by pre pending an underscore converting to all lower case and appending an underscore Byte counts for character strings appear in sequence following the last argument in the argument list The following is an example of the pseudo code for the above call using UNIX conventions call work addr ERR addr a addr b addr n 3 10 14 3 Using the Default Calling Convention Using the Default calling convention is straightforward Use the default convention if no directives are inserted to modify calling conventions and if the Munix compilation flag is not used See the previous section for a complete description of the Default convention 10 14 4 Using the STDCALL Calling Convention Using the STDCALL calling convention requires the insertion of a compiler directive into the declarations section of any Fortran program unit whi
101. I s compilers allocate stack space for each routine in multiples of 8 bytes Variable Length Parameter Lists Parameter passing in registers can handle a variable number of parameters The C language uses a special method to access variable count parameters The stdarg h and varargs h files define several functions to access these parameters A C routine with variable parameters must use the va_start macro to set up a data structure before the parameters can be used The va_arg macro must be used to access the successive parameters C Parameter Conversion In C for a called prototyped function the parameter type in the called function must match the argument type in the calling function If the called function is not prototyped the calling convention uses the types of the arguments but promotes char or short to int and unsigned char or unsigned short to unsigned int and promotes float to double unless you use the Msingle option For more 216 Run time Environment information on the Msingle option refer to Chapter 3 If the called function is prototyped the unused bits of a register containing a char or short parameter are undefined and the called function must extend the sign of the unused bits when needed Calling Assembly Language Programs Example A 1 C Program Calling an Assembly language Routine File testmain c main long 1 paral 0x3 800000 float f para2 1 0 double d_para3 0 5 float retur
102. If you want to get started quickly with optimization a good set of options to use with any of the PGI compilers is fastsse Mipa fast For example pgf95 fastsse Mipa fast prog f For all of the PGI Fortran C and C compilers this option will generally produce code that is well optimized without the possibility of significant slowdowns due to pathological cases The fastsse option is an aggregate option that includes a number of individual PGI compiler options which PGI compiler options are included depends on the target for which compilation is performed The Mipa fast option invokes interprocedural analysis including several IPA suboptions For C programs add Minline levels 10 no_exceptions pgCC fastsse Mipa fast Minline levels 10 no exceptions prog cc Note a C program compiled with no_execptions will fail if the program uses exception handling By experimenting with individual compiler options on a file by file basis further significant performance gains can sometimes be realized However individual optimizations can sometimes cause slowdowns depending on coding style and must be used carefully to ensure performance improvements result In addition to fastsse the optimization flags most likely to further improve performance are O3 Mpfi Mpfo Minline and on targets with multiple processors Mconcur In addition the Msafeptr option can significantly improve performance of C C pr
103. In this example the compiler instruments the executable produced from myprog f to perform array bounds checking at runtime pgf95 C myprog f Cross reference Mbounds C Stops after the assembling phase Use the c option to halt the compilation process after the assembling phase and write the object code to the file filename o where the input file is filename f Default The compiler produces an executable file does not use the c option 46 Command Line Options Usage In this example the compiler produces the object file myprog o in the current directory pgf95 c myprog f Cross reference P Mkeepasm o and S D Defines a preprocessor macro Use the D option to create a macro with a given value The value must be either an integer or a character string You can use the D option more than once on a compiler command line The number of active macro definitions is limited only by available memory You can use macros with conditional compilation to select source text during preprocessing A macro defined in the compiler invocation remains in effect for each module on the command line unless you remove the macro with an undef preprocessor directive or with the U option The compiler processes all of the U options in a command line after processing the D options Syntax Dname value Where name is the symbolic name and value is either an integer value or a character string Default If you de
104. PGI User s Guide Parallel Fortran C and C for Scientists and Engineers The Portland Group STMicroelectronics Two Centerpointe Drive Lake Oswego OR 97035 While every precaution has been taken in the preparation of this document The Portland Group a wholly owned subsidiary of STMicroelectronics makes no warranty for the use of its products and assumes no responsibility for any errors that may appear or for damages resulting from the use of the information contained herein The Portland Group retains the right to make changes to this information at any time without notice The software described in this document is distributed under license from STMicroelectronics and may be used or copied only in accordance with the terms of the license agreement No part of this document may be reproduced or transmitted in any form or by any means for any purpose other than the purchaser s personal use without the express written permission of The Portland Group Many of the designations used by manufacturers and sellers to distinguish their products are claimed as trademarks Where those designations appear in this manual The Portland Group was aware of a trademark claim The designations have been printed in caps or initial caps Thanks is given to the Parallel Tools Consortium and in particular to the High Performance Debugging Forum for their efforts PGF95 PGF90 PGC Cluster Development Kit CDK PGI Unified Binary PGI Visual For
105. RY_PATH is already initialized it is important not to overwrite its contents Assuming you have placed tobeshared so in a directory home myusername bin you can initialize LD_LIBRARY_PATH to include that directory and preserve its existing contents as follows setenv LD_LIBRARY_PATH LD_LIBRARY_PATH home myusername bin If you know that tobeshared so will always reside in a specific directory you can create the executable myprog in a form that assumes this using the R link time option For example you can link as follows o pgf95 o myprog myprof f tobeshared so R home myusername bin Note that there is no space between R and the directory name As with the L option no space can be present If the R option is used it is not necessary to initialize LD_LIBRARY_PATH In the example above the dynamic linker will always look in home myusername bin to resolve references to tobeshared so By default if the LD_LIBRARY_PATH environment variable is not set the linker will only search usr lib for shared objects The command ldd is a useful tool when working with shared object files and executables that reference them When applied to an executable as follows o ldd myprog ldd lists all shared object files referenced in the executable along with the pathname of the directory from which they will be extracted If the pathname is not hard coded using the R option and if LD_LIBRARY_PATH is not initialized the pathname is listed
106. The PGI environment variable specifies the root directory where the PGI compilers and tools are installed The default value of this variable is usr pgi In most cases the name of this root directory is derived dynamically by the PGI compilers and tools through determination of the path to the instance of the compiler or tool that has been invoked However there are still some dependencies on the PGI environment variable and it can be used as a convenience when initializing your environment for use of the PGI compilers and tools For example assuming you use csh and want the 64 bit linux86 64 versions of the PGI compilers and tools to be default ole setenv PGI usr pgi setenv MANPATH SMANPATH PGI linux86 6 0 man setenv LM LICENSE FILE PGI license dat set path PGI linux86 64 6 0 bin path dp de E PGI_CONTINUE If the PGI CONTINUE environment variable is set upon execution of a program compiled with Mchkfpstk the stack will be automatically cleaned up and execution will continue There is a performance penalty associated with the stack cleanup If PGI_CONTINUE is set to verbose the stack will be automatically cleaned up and execution will continue after printing of a warning message 172 Libraries and Environment Variables STATIC_RANDOM_ SEED The first call to the Fortran 90 95 RANDOM_SEED intrinsic without arguments will reset the random seed to a default value then advance the seed by a variable amount based on tim
107. Y ccccccececceeeeeeeeeeeeeeeeeeeeeeeaeeseeeeetaeeseeeeeeeeaaeeee 153 Representation of Fortran Data Types ccccecceeceeeeeeeenneeeeeeeennaeeeeeeeeaees 176 Real Data Type Ranges crcrr asenaan e a R E A 178 Scalar Type Alonment nern nesnnnennrenntnnnnstnessresssennsennnenn 178 C C Scalar Data Tvpes 181 e ET ln Cu 182 Fortran and C C Data Type Compatbiliiv esenee eenen ene 190 Fortran and C C Representation of the COMPLEX Type 190 Calling Conventions Supported by the PGI Fortran Compilers 1522 201 Register AllOCaton EE 210 Standard Stack Frame cccccccccceeeeeeeeeeeeeeeeeeeseeeeeeaaeesecaeeeseeeeseeeeeeaeeee 211 Stack Contents for Functions Returning struct Union 0 ccceeceeeeeees 214 Integral and Pointer Arguments eee eeeeeee eee eeeeeeeeeeeeeeeaeeeeeeeeaeeeeenenaaes 215 Floating point ArQuiment ecccceeeseeeeeeeeeeeeeeeeeneeeeeeeenseneeeeeenseeeeeeeenenees 215 Structure and Union Arguments ssseeesseeseeeseresirresrenestnnrrnnrsrnntnnessnssne 216 Redister Aller auser tina ite es 219 Standard Stack Frame c cccecccceeeeeeeceeeeeeeeeeeceeeeeeaaeeseaeeeseeeeseeeeeeneeeees 220 Register Allocation for Example A2 224 Linux86 64 Fortran Fundamental Types ccccccceeeesceeeeeeeseaeeeeeeeeaees 227 Fortran and C C Data Type Compatbiltv eeren 230 Fortran and C C Representation of the COMPLEX Type 230 Register AllOCation EE 233 Standard Stack Frame 234
108. YPE ENDTYPE STRUCTURE ENDSTRUCTURE UNION ENDUNION MAP ENDMAP declaration contains no members 152 All dimension specifiers must ber 153 Array objects are not conformable 154 DISTRIBUTE target must be a processor S155 S156 Number of colons and triplets must be equal in ALIGN with 157 Illegal subscript use of ALIGN dummy S158 Alternate return not specified in SUBROUTINE or ENTRY An alternate return can only be used if alternate return specifiers appeared in the SUBROUTINE or ENTRY statements S159 Alternate return illegal in FUNCTION subprogram An alternate return cannot be used in a FUNCTION 264 Messages 160 ENDSTRUCTURE ENDUNION or ENDMAP does not match top 161 Vector subscript must be rank one array W162 Not equal test of loop control variable replaced with lt or gt test 163 lt reserved message number gt 164 Overlapping data initializations of An attempt was made to data initialize a variable or array element already initialized S165 appeared more than once as a subprogram A subprogram name appeared more than once in the source file The message is applicable only when an assembly file is the output of the compiler 166 cannot be a common block and a subprogram A name appeared as a common block name and a subprogram name The message is applicable only when an assembly file is the output of the compiler I167 Inconsistent size of common block
109. _op parameter N 9999 real 4 x n y n z n w n i i 2 i i 4 i do j 1 200000 call loop x y z w 1 0e0 n enddo print x 1 x 771 x 3618 x 6498 x 9999 end subroutine loop a b c d s n integer i n real 4 a n b n c n d n s do i 1 n a i b i c i s d i enddo end Assume the above program is compiled as follows pgf95 fast Minfo vadd f vector op 4 Loop unrolled 4 times loop 18 Loop unrolled 4 times Following is the result if the generated executable is run and timed on a standalone AMD Opteron 2 2 Ghz system bin time a Out 1 000000 771 000 3618 000 6498 00 9999 00 5 15user 0 00system 0 05 16 elapsed 99 CPU Now recompile with SSE vectorization enabled 24 Optimization amp Parallelization pgf95 fast Mvect sse Minfo vadd f vector op 4 Unrolling inner loop 8 times Loop unrolled 7 times completely unrolled loop 18 Generating vector sse code for inner loop Generated 3 prefetch instructions for this loop Note the informational message indicating that the loop has been vectorized and SSE instructions have been generated The second part of the informational message notes that prefetch instructions have been generated for 3 loads to minimize latency of transfers of data from main memory Executing again you should see results similar to the following bin time a Out 1 000000 771 000 3618 00 6498 00 9999 0 3 55user 0 00system 0 03 56elaps
110. a sseseeeeeeeee 30 Building a Program Without IPA Single Step ceeeeeeeeeeeteeeeeeeeee 30 Building a Program Without IPA Several Steps ccceccesssteeeeeenes 31 Building a Program Without IPA Using Make sssrin 31 Building a Program With IPA aseen iEn Ea a Ea era N AENA 32 Building a Program with IPA Single Step ccccceeseeeeeeeeeeesteeeeeeeees 32 Building a Program with IPA Several Steps cccccesesteeeeeeeeetteeeees 33 Building a Program with IPA Using Make eeccseceeeeeeeneeeeeeeeeteeeeeeeeee 34 Questions about IPA eebe TEEN NEEN eege 34 Profile Feedback Optimization using Mpfi Mpfo seesseeeeeeeeeeeeeeeene 35 Default Optimization Levels c ccceeccceesseeceeneeeeeseeeeeeeeeeeaaeeseeeeeeeaeeeeaes 36 Local Optimization Using Directives and Pragmas cccccsseeeeeeeses 37 Execution Timing and Instruction Counting 0 c cceeeeeeeeeseeeeetteeeeeneeeees 38 Command Line Options ceeeccceseeeeeeesseeeeeeeseeeeeeeesseeeeeeeneeeees 39 Generic PGI Compiler Options 45 C and C specific Compiler Options ccceccceeeeeeeeeeeeeteeeeeteeeeeeeeees 94 Function Inlining EE 101 INVOKING FUNCTION Inlmtmg verienenn ieties 101 Usngam Inine LHaren a a NRE EAEE 102 Creating an Inline Library ssseseeesseeeinesirsssereserrssnrsrnrsrrnnsrnsssresssrnnsns 103 Working with Inline Libraries 103 Updating Inline Libraries Makefiles AAA 103 Error
111. a F extension or C and C source files When linking a program with a library the linker extracts only those library components that the program needs The compiler drivers link in several libraries by default For more information about libraries refer to Chapter 8 Libraries and Environment Variables 1 3 2 Output Files By default an executable output file produced by one of the PGI compilers is placed in the file a out or on Windows a filename based on the name of the first source or object file on the command line As shown in the preceding section you can use the o option to specify the output file name If you use one of the options F Fortran only P C C only S or c the compiler produces a file containing the output of the last phase that completes for each input file as specified by the option supplied The output file will be a preprocessed source file an assembly language file or an unlinked object file respectively Similarly the E option does not produce a file but displays the preprocessed source file on the standard output Using any of these options the o option is valid only if you specify a single input file If no errors occur during processing you can use the files created by these options as input to a future invocation of any of the PGI compiler drivers The following table lists the stop after options and the output files that the compilers create when you use these options Get
112. a the letters L R and G indicate loop routine and global scope respectively The default scope is surrounded by parentheses The in the scope field indicates that the scope is the code following the pragma until the end of the routine for R scoped pragmas as opposed to the entire routine or until the end of the file for G scoped pragmas as opposed to the entire file Many of the pragmas can be preceded by no The default entry in the table indicates the default of the pragma N A appears if a default does not apply The name of any pragma may be prefixed with M for example Mnoassoc is equivalent to noassoc and Mvintr is equivalent to vintr The section following the table provides brief descriptions of the pragmas that are unique to C C Pragmas that have a corresponding directive in Fortran are described in Section 7 2 Fortran Directive Summary 152 Optimization Directives and Pragmas Table 7 2 C C Pragma Summary Pragma Function Default Scope altcode noaltcode Do don t generate alternate code for vector altcode L RG ized and parallelized loops assoc noassoc Do don t perform associative transforma assoc L RG tions bounds nobounds Do don t perform array bounds checking nobounds R G concur noconcur Do don t enable auto concurrentization of concur L RG loops depchk nodepchk Do don t ignore potential data dependencies depch
113. a period lets the compiler know that the file on the command line is an inline library pgf95 Minline proc lib il myprog f 102 Function Inlining 4 2 Creating an Inline Library You can create or update an inline library using the Mextract command line option If you do not specify a selection criteria along with the Mextract option the compiler attempts to extract all subprograms When you use the Mextract option only the extract phase is performed the compile and link phases are not performed The output of an extract pass is a library of functions available for inlining It is placed in the inline library file specified on the command line with the o filename specification If the library file exists new information is appended to it If the file does not exist it is created You can use the Minline option with the Mextract option In this case the extracted library of functions can have other functions inlined into the library Using both options enables you to obtain more than one level of inlining In this situation if you do not specify a library with the Minline option the inline process consists of two extract passes The first pass is a hidden pass implied by the Minline option during which the compiler extracts functions and places them into a temporary library The second pass uses the results of the first pass but puts its results into the library that you specify with the o option 4 2 1 Working
114. ables for each thread in the team When an assignment to a private variable occurs each thread assigns to its local copy of the variable When operations involving a private variable occur each thread performs the operations using its local copy of the variable Important points about private variables are e Variables declared private in a parallel region are undefined upon entry to the parallel region If the first use of a private variable within the parallel region is in a right hand side expression the results of the expression will be undefined i e this is probably a coding error e Likewise variables declared private in a parallel region are undefined when serial execution resumes at the end of the parallel region The variables specified in a SHARED list are shared between all threads in a team meaning that all threads access the same storage area for SHARED data OpenMP Directives for Fortran 109 The DEFAULT clause lets you specify the default attribute for variables in the lexical extent of the parallel region Individual clauses specifying PRIVATE SHARED etc status override the declared DEFAULT Specifying DEFAULT NONE declares that there is no implicit default and in this case each variable in the parallel region must be explicitly listed with an attribute of PRIVATE SHARED FIRSTPRIVATE LASTPRIVATE or REDUCTION Variables that appear in the list of a FIRSTPRIVATE clause are subject to the same semantics as PRIVAT
115. ace Use these naming conventions If you call a C C function from Fortran you should rename the C C function by appending an underscore or use CSPRAGMA C in the Fortran program refer to Chapter 7 Optimization Directives and Pragmas for details on CSPRAGMA C If you call a Fortran function from C C you should append an underscore to the Fortran function name in the calling program 10 5 Compatible Data Types The next table shows compatible data types between Fortran and C C Table 10 2 Fortran and C C Representation of the COMPLEX Type shows how the Fortran COMPLEX type may be represented in C C If you can make your function subroutine parameters and return values match types you should be able to use inter language calling Inter language Calling 189 Table 10 1 Fortran and C C Data Type Compatibility Fortran Type lower C C Type Size bytes case character x char x 1 character n x char x n n real x float x 4 real 4 x float x 4 real 8 x double x 8 double precision double x 8 integer x int x 4 integer 1 x signed charx 1 integer 2 x short x 2 integer 4 x int x 4 integer 8 x long long x 8 logical x int x 4 logical 1 x char x 1 logical 2 x short x 2 logical 4 int x 4 logical 8 long long x 8 Table 10 2 Fortran and C C Representation of the COMPLEX Type Fortran Type lower
116. agmas with loop scope That is they apply to the code following the pragma e opt and safe When the opt and safe pragmas are placed within a routine they apply to the entire routine as if they had been placed at the beginning of the routine 78 Prefetch Directives When vectorization is enabled using the Mvect or Mprefetch compiler options or an aggregate option such as fastsse that incorporates Mvect the PGI compilers selectively emit instructions to explicitly prefetch data into the data cache prior to first use It is possible to control how these prefetch instructions are emitted using prefetch directives These directives only have an effect when vectorization or prefetching are enabled on the command line See Table 2 Processor Options in the Preface for a list of processors that support prefetch instructions The syntax of a prefetch directive is as follows Optimization Directives and Pragmas 157 c mem prefetch lt varli gt lt var2 gt where lt varn gt is any valid variable or array element reference NOTE The sentinel for prefetch directives is c mem which is distinct from the cpgi sentinel used for optimization directives Any prefetch directives that use the cpgi sentinel will be ignored by the PGI compilers The c must be in column 1 Either or is allowed in place of c The scope indicators g r and 1 used with the cpgi sentinel are not supported The directive name including
117. al function omp_get_dynamic 122 OpenMP Directives for Fortran is designed to allow the user to query whether automatic dynamic adjustment of the number of threads used for execution of parallel regions is enabled This function is recognized but currently always returns FALSE subroutine omp_set_nested scalar_logical_exp is designed to allow enabling disabling of nested parallel regions This function is recognized but currently has no effect logical function omp_get_nested is designed to allow the user to query whether dynamic adjustment of the number of threads available for execution of parallel regions is enabled This function is recognized but currently always returns FALSE double precision function omp_get_wtime returns the elapsed wall clock time in seconds as a DOUBLE PRECISION value Times returned are per thread times and are not necessarily globally consistent across all threads double precision function omp_get_wtick returns the resolution of omp_get_wtime in seconds as a DOUBLE PRECISION value subroutine omp_init lock integer var initializes a lock associated with the variable integer_var for use in subsequent calls to lock routines This initial state of integer_var is unlocked It is illegal to make a call to this routine if integer_var is already associated with a lock subroutine omp_destroy lock integer var disassociates a lock associated with the variable integer_var subroutine omp_s
118. all threads in a team meaning that all threads access the same storage area for shared data The default clause allows the user to specify the default attribute for variables in the lexical extent of the parallel region Individual clauses specifying private shared etc status override the declared default Specifying default none declares that there is no implicit default and in this case each variable in the parallel region must be explicitly listed with an attribute of private shared firstprivate or reduction Variables that appear in the list of a firstprivate clause are subject to the same semantics as private variables but in addition are initialized from the original object existing prior to entering the parallel region Variables that appear in the list of a reduction clause must be shared A private copy of each variable in list is created for each thread as if the private clause had been specified Each private copy is initialized according to the operator as specified in the following table OpenMP Pragmas for C and C 127 Table 6 1 Initialization of Reduction Variables Operator Initialization 0 1 0 amp 0 0 A 0 amp amp 1 I 0 e At the end of the parallel region a reduction is performed on the instances of variables appearing in list using operator as specified in the reduction clause The initial value of each reduction variable is included in the reduction operation
119. alled with the default release and must have a release number greater than or equal to 4 1 the first release for which this functionality is supported Usage The following command line shows the output using the V option pgf95 V myprog f The following command line causes PGF95 to compile using the 5 2 release instead of the default pgcce V5 2 myprog c Cross reference Minfo v V Use the v option to display the invocations of the compiler assembler and linker These invocations are command lines created by the compiler driver from the files and the W options you specify on the compiler command line Default The compiler does not display individual phase invocations Command Line Options 93 Cross reference Minfo V W Passes arguments to a specific phase Use the W option to specify options for the assembler compiler or linker Note A given PGI compiler command invokes the compiler driver which parses the command line and generates the appropriate commands for the compiler assembler and linker Syntax W 0lall option option Where 0 the number zero specifies the compiler a specifies the assembler 1 lowercase letter 1 specifies the linker option is a string that is passed to and interpreted by the compiler assembler or linker Options separated by commas are passed as separate command line arguments Note You cannot have a space between the W and the single
120. and omp parallel for loops that include the schedule runtime clause The default value for this variable is static If the optional chunk size is not set a chunk size of is assumed except in the case of a static schedule For a static schedule the default is as defined in Section 6 6 omp for Examples of the use of OMP_SCHEDULE are as follows setenv OMP_ SCHEDULE Static 5 setenv OMP SCHEDULE guided 8 x g setenv OMP_SCHEDULE dynamic OMP_DYNAMIC currently has no effect OMP_NESTED currently has no effect OpenMP Pragmas for C and C 139 MPSTKZ increase the size of the stacks used by threads executing in parallel regions For use with programs that utilize large amounts of thread local storage in the form of private variables or local variables in functions or subroutines called within parallel regions The value should be an integer lt n gt concatenated with M or m to specify stack sizes of n megabytes For example setenv MPSTKZ 8M 140 OpenMP Pragmas for C and C Chapter 7 Optimization Directives and Pragmas Directives are Fortran comments that the user may supply in a Fortran source file to provide information to the compiler Directives alter the effects of certain command line options or default behavior of the compiler While a command line option affects the entire source file that is being compiled directives apply or disable the effects of a command line option to selected subprog
121. any x86 processor based system x64 generate 64 bit unified binary code including full optimizations and support for both AMD and Intel x64 processors See Table 2 Processor Options for a concise list of the features of these processors that distinguish them as separate targets when using the PGI compilers and tools Syntax for 64 bit targets tp k8 64 k8 64e p7 64 core2 64 x64 Syntax for 32 bit targets tp k8 32 p6 p7 core2 piii px Usage In the following example pgf95 sets the target architecture to EM64T pgf95 tp p7 64 myprog f 92 Command Line Options Default The default style of code generation is auto selected depending on the type of processor on which compilation is performed The tp x64 style of unified binary code generation is only enabled by an explicit tp x64 option U Undefines a preprocessor macro Use the U option or the undef preprocessor directive to undefine macros Syntax Usymbol Where symbol is a symbolic name Usage The following examples undefine the macro test pgf95 Utest myprog F pgf95 Dtest Utest myprog F Cross reference D _Mnostdde V release_number Displays additional information including version messages If a release_number is appended the compiler driver will attempt to compile using the specified release instead of the default release There can be no space between V and release number The specified release must be co inst
122. applications or DLL s REFERENCE name Specifies that the argument name is being passed by reference Often this attribute is used in conjuction with STDCALL where STDCALL refers to an entire routine then individual arguments are modified with REFERENCE STDCALL routine_name Specifies that routine routine_name will have its arguments passed by value When a routine marked STDCALL is called arguments except arrays and characters will be sent by value The standard F90 F95 calling convention is by reference VALUE name Specifies that the argument name is being passed by value Often used to specify that a particular argument is being passed by value Loop Distribution Directive DECS DISTRIBUTE POINT This directive is front end based and tells the compiler at what point within a loop to split into two loops subroutine dist a b n integer i integer n integer a integer b do i 1 n 150 Optimization Directives and Pragmas a i a i 2 DECS DISTRIBUTE POINT b i b i 4 enddo end subroutine DECS DISTRIBUTEPOINT is same as DEC DISTRIBUTE POINT ALIAS Attribute DECS ALIAS same as DEC ATTRIBUTES ALIAS 7 5 Adding Pragmas to C and C Pragmas may be supplied in a C C source file to provide information to the compiler Like directives in Fortran pragmas alter the effects of certain command line options or default behavior of the compiler many pragmas have a corresponding command l
123. applies e It is referenced in a statement function e A common block mismatch exists i e the caller must contain all common blocks specified in the callee and elements of the common blocks must agree in name order and type except that the caller s common block can have additional members appended to the end of the common block e An argument mismatch exists i e the number and type size of actual and formal parameters must be equal e A name clash exists e g a call to subroutine xyz in the extracted subprogram and a variable named xyz in the caller The following types of C and C functions cannot be inlined e Functions whose return type is a struct data type or functions which have a struct argument This limitation applies only to x86 targets e Functions containing switch statements Functions which reference a static variable whose definition is nested within the function Function which accept a variable number of arguments Certain C C functions can only be inlined into the file that contains their definition Function Inlining 105 106 Static functions Functions which call a static function Functions which reference a static variable Function Inlining Chapter5 OpenMP Directives for Fortran The PGF77 and PGF95 Fortran compilers support the OpenMP Fortran Application Program Interface The OpenMP shared memory parallel programming model is defined by a collection of compiler directives library
124. ation of array An asterisk may be used only as the upper bound of the last dimension 254 Messages S049 Illegal use of sr in non subroutine subprogram The alternate return specifier is legal only in the subroutine statement Programs functions and block data are not allowed to have alternate return specifiers S050 Assumed size array is not a dummy argument 051 Unrecognized built in function The allowable built in functions are VAL REF LOC and FILL One was encountered that did not match one of these allowed forms 052 Illegal argument to VAL or LOC 053 SREF or VAL not legal in this context The built in functions REF and VAL can only be used as actual parameters in procedure calls W054 Implicit character used in a previous implicit statement An implicit character has been given an implied data type more than once The implied data type for the implicit character is changed anyway w055 Multiple implicit none statements The IMPLICIT NONE statement can occur only once in a subprogram w056 Implicit type declaration The dclchk switch and an implicit declaration following an IMPLICIT NONE statement will produce a warning message for IMPLICIT statements S057 Illegal equivalence of dummy variable Dummy arguments may not appear in EQUIVALENCE statements S058 Equivalenced variables and not in same common block A common block variable must not be equivalenced with a variable i
125. b 200 200 float c 200 200 d 200 200 int i j for i 0 i lt 200 i for j 0 j lt 200 j a i j a i j b i j c i j c i j c i j b i j d i j When this source is compiled using the Mvect command line option func2 is vectorized but func1 is not vectorized In the following example the global novector pragma turns off vectorization for the entire file 156 Optimization Directives and Pragmas include math h funcl1 pragma global novector float a 100 100 b 100 100 float c 100 100 d 100 100 int 1 54 for i 0 1 lt 100 i for j 0 j lt 100 j a i j a i j b i j c i j c i j c i j b i j d i j func2 float a 200 200 b 200 200 float c 200 200 d 200 200 tnt aire for i 0 i lt 200 i for j 0 j lt 200 j a i j a i j b i j c i j c i j c i j b i j d i j Special Scope Rules Special rules apply for a pragma with loop routine and global scope When the pragma is placed within a routine it applies to the routine from its point in the routine to the end of the routine The same rule applies for one of these pragmas with global scope However there are several pragmas for which only routine and global scope applies and which affect code immediately following the pragma bounds and fcon The bounds and fcon pragmas behave in a similar manner to pr
126. bles to be made private to each thread but global within the thread This pragma must appear in the declarations section of a program unit after the declaration of any variables listed On entry to a parallel region data in a threadprivate variable is undefined unless copyin is specified on the omp parallel pragma When a variable appears in an omp threadprivate pragma each thread s copy is initialized once at an unspecified point prior to its first use as the master copy would be initialized in a serial execution of the program The following restrictions apply to the omp threadprivate pragma e The omp threadprivate pragma must appear after the declaration of every threadprivate variable included in list e It is illegal for an omp threadprivate variable to appear in any clause other than a copyin schedule or if clause e If a variable is specified in an omp threadprivate pragma in one translation unit it must be specified in an omp threadprivate pragma in every translation unit in which it appears e The address of an omp threadprivate variable is not an address constant e An omp threadprivate variable must not have an incomplete type or a reference type 6 15 Run time Library Routines User callable functions are available to the OpenMP C C programmer to query and alter the parallel execution environment Any program unit that invokes these functions should include the statement include lt omp h gt The omp h include file contains
127. c filel c pgce o a out filel o file2 o file3 o d i d i 2 7 3 Building a Program Without IPA Using Make The program compilation and linking process can be simplified greatly using the make utility on systems where it is supported Using a file makefile containing the following lines a out filel o file2 o file3 o pgcc OPT o a out filel o file2 o file3 o filel o filel c pgce OPT c filel c file2 o file2 c pgcc OPT c file2 c file3 o file3 c pgce OPT c file3 c Optimization amp Parallelization 31 It is possible to type a single make command make The make utility determines which object files are out of date with respect to their corresponding source files and invokes the compiler to recompile only those source files and to relink the executable If you subsequently edit one or more source files the executable can be rebuilt with the minimum number of recompilations using the same single make command 2 7 4 Building a Program with IPA Interprocedural analysis and optimization IPA by the PGI compilers is designed to alter the standard and make utility command level interfaces outlined above as little as possible IPA occurs in three phases e Collection Create a summary of each function or procedure collecting the useful information for interprocedural optimizations This is done during the compile step if the Mipa switch is present on the command line summary information is collected and stored in t
128. c to each occurrence of the message B 3 2 Message List Error message severities I informative WwW warning S severe error F fatal error V variable v000 Internal compiler error This message indicates an error in the compiler rather than a user error although it may be possi ble for a user error to cause an internal error The severity may vary if it is informative or warning correct object code was probably generated but it is not safe to rely on this Regardless of the severity or cause internal errors should be reported to trs pgroup com F001 Source input file name not specified On the command line source file name should be specified either before all the switches or after them F002 Unable to open source input file 248 Messages Source file name misspelled file not in current working directory or file is read protected F003 Unable to open listing file Probably user does not have write permission for the current working directory F004 Generic message for file errors F005 Unable to open temporary file Compiler uses directory usr tmp or tmp in which to create temporary files If neither of these directories is available on the node on which the compiler is being used this error will occur s006 Input file empty Source input file does not contain any Fortran statements other than comments or compiler directives F007 Subprogram too large to compile at this optimization le
129. case C C Type Size bytes complex x struct float r i x 8 complex 8 x struct float r i x 8 double complex x struct double dr di x 16 10 5 1 Fortran Named Common Blocks 190 Inter language Calling A named Fortran common block can be represented in C C by a structure whose members correspond to the members of the common block The name of the structure in C C must have the added underscore For example the Fortran common block INTEGER I COMPLEX C DOUBLE COMPLEX CD DOUBLE PRECISION D COMMON COM i c cd d is represented in C with the following equivalent extern struct int i struct float real imag c struct double real imag cd double d com and in C with the following equivalent extern C struct int i struct float real imag c struct double real imag cd double d com_ 10 6 Argument Passing and Return Values In Fortran arguments are passed by reference i e the address of the argument is passed rather than the argument itself In C C arguments are passed by value except for strings and arrays which are passed by reference Due to the flexibility provided in C C you can work around these differences Solving the parameter passing differences generally involves intelligent use of the amp and operators in argument passing when C C calls Fortran and in argument declarations when Fortran calls C C For strings declared in Fortran as typ
130. cceceseeceeeeeeeneeeeeeeeeeseeeeseeueeeseneeeeed 197 Example C Calling Fortran ccccccceceeeceeeeeeeeeeeeeeseaeeseeeeeeeenaeeneeeees 199 Win32 Calling CONVENTIONS c cceceeeeeeeeeee cece eeeeeeeeeeaeeeeeaaeesecaeeeeeaeeeees 200 Win32 Fortran Calling Conventions c ccccceeeeeeceeeeeeeeeeeteeeeeteeeseeeeeaes 200 Symbol Name Construction and Calling Example 201 Using the Default Calling CONVENTION 00 0 eee eeeeeceeeteeeeeeeeetetteeeteneeeeeaees 202 Using the STDCALL Calling Convention ceccccceeceeeseeseeeeeeseeeeenees 202 Using the C Calling Convention cceccececeeeeeeeeeeeeeeeeseeeeseeeeeeeeeeeeeaees 203 Using the UNIX Calling Convention cccceeeeeceeeeeeeeeeeeeeseeeeeeeeeeeeeaees 203 C Name Mangling cseccccseseeeeeeeeeneeeenseeeeeeeeseeeeeeeseneeeeeneeeeees 205 Contents 11 1 Types of Mangling serrit ete aa Ae Mei ete lan ies Gel ery 205 11 2 Mangling SUMMALry cccccccceeeceeeeeeeeeeeeeeeeeeceeaeeeeaaeeeeeaeeeeeaaeseeneeeesnaaeesenes 206 11 2 1 Type Name Mang hing e sar oertein aan a TR EEA ENEE 206 11 2 2 Nested Class Name Mangling ccececceceeeeeeeeeeeeeeeeeeeeeeseceeeseeeeesaees 207 11 2 3 Local Class Name Mangling ccccccceesceeceeeeeeeeeeeseeeeeeeeaeseeeeeeessaaeeenes 207 11 2 4 Template Class Name Mangling ccccsceeeeeeeeeeeeeeeeeeeeeeeaeeseeeeeeeeaeeeees 207 Appendix A Run time Environment c sseeeeeeeneieeeeeeeeeeeeeeee
131. ccess performance and make use of packed SSE instructions which perform the same operation on multiple data items concurrently Unrolling replicates the body of loops to reduce loop branching overhead and provide better opportunities for local optimization vectorization and scheduling of instructions Performance for loops on systems with multiple processors may also improve using the parallelization features of the PGI compilers Inter Procedural Analysis and Optimization IPA Interprocedural analysis allows use of information across function call boundaries to perform optimizations that would otherwise be unavailable For example if the actual argument to a function is in fact a constant in the caller it may be possible to propagate that constant into the callee and perform optimizations that are not valid if the dummy argument is treated as a variable A wide range of optimizations are enabled or improved by using IPA including but not limited to data alignment optimizations argument removal constant propagation pointer disambiguation pure function detection F90 F95 array shape propagation data placement vestigial function removal automatic function inlining inlining of functions from pre compiled libraries and interprocedural optimization of functions from pre compiled libraries Function Inlining This optimization allows a call to a function to be replaced by a copy of the body of that function This optimization will sometimes sp
132. ce you need to develop a special procedure Follow these guidelines Note that if a C function contains objects with constructors and destructors calling such a function from either C or Fortran will not be possible unless the initialization in the main program is performed from a C program where constructors and destructors are properly initialized 187 In general you can call a C function from C without problems as long as you use the extern C keyword to declare the C function in the C program This prevents name mangling for the C function name If you want to call a C function from C likewise you have to use the extern C keyword to declare the C function This keeps the C compiler from mangling the name of the function e You can use the __cplusplus macro to allow a program or header file to work for both C and C For example the following defines in the header file stdio h allow this file to work for both C and C ifndef _STDIO H define _STDIO_H ifdef cplusplus extern C endif __cplusplus Functions and data types defined ifdef cplusplus endif cplusplus endif C member functions cannot be declared extern as their names will always be mangled Therefore C member functions cannot be called from C or Fortran 10 3 Functions and Subroutines Fortran C and C define functions and subroutines differently For a Fortran program calling a C or C
133. ce xiii AMD64 EM64T x64 linux86 linux86 64 Win32 Win64 Windows a 64 bit processor from AMD designed to be binary compatible with IA32 processors and incorporating new features such as additional registers and 64 bit addressing support for improved performance and greatly increased memory range a 64 bit IA32 processor with Extended Memory 64 bit Technology extensions that are binary compatible with AMD64 processors collectively all AMD64 and EM64T processors supported by the PGI compilers 32 bit Linux operating system running on an x86 or x64 processor based system with 32 bit GNU tools utilities and libraries used by the PGI compilers to assemble and link for 32 bit execution 64 bit Linux operating system running on an x64 processor based system with 64 bit and 32 bit GNU tools utilities and libraries used by the PGI compilers to assemble and link for execution in either linuxS6 or linux86 64 environments The 32 bit development tools and execution environment under linux86 64 are considered a cross development environment for x86 processor based applications any of the 32 bit Microsoft Windows Operating Systems XP 2000 Server 2003 running on an x86 or x64 processor based system On these targets the PGI compiler products include additional tools and libraries needed to build executables for 32 bit Windows systems any of the 64 bit Microsoft Windows Operating Systems XP Professional Windows Server
134. cedural analysis has verified that the prescriptive or descriptive distribution formats of this many array dummy arguments match the formats of the actual argument I390 IPA array distribution targets propagated Interprocedural analysis has found this many array dummy arguments that could have the transcriptive distribution target replaced by a descriptive target I391 IPA array distribution targets verified Interprocedural analysis has verified that the prescriptive or descriptive distribution targets of this many array dummy arguments match the targets of the actual argument I392 IPA common blocks optimized Interprocedural analysis has found this many common blocks that could be optimized I393 IPA common blocks not optimized Interprocedural analysis has found this many common blocks that could not be optimized either because the common block was not declared in the main program or because it was declared differently in different subprograms I394 IPA replaced by constant value The dummy argument was replaced by a constant as per interprocedural analysis I395 IPA changed to INTENT IN The dummy argument was changed to INTENTCN as per interprocedural analysis I396 IPA array alignment propagated to The template alignment for the dummy argument was changed as per interprocedural analysis I397 IPA distribution format propagated to The distribution format for the dummy argument was changed as per
135. ch calls the STDCALL program unit This directive has no effect when the Munix compilation flag is used meaning you cannot mix UNIX style argument passing and STDCALL calling conventions within the same file Syntax for the directive is as follows 202 Inter language Calling MSSATTRIBUTES STDCALL work Where work is the name of the subroutine to be called using STDCALL conventions More than one subroutine may be listed separated by commas See Section 10 14 2 Symbol Name Construction and Calling Example for a complete description of the implementation of STDCALL NOTE The directive prefix DEC is also supported but requires a space between the prefix and the directive keyword ATTRIBUTES The must begin the prefix when compiling using Fortran 90 freeform format The characters C or can be used in place of in either form of the prefix when compiling used fixed form F 77 style format The directives are completely case insensitive 10 14 5 Using the C Calling Convention Using the C calling convention requires the insertion of a compiler directive into the declarations section of any Fortran program unit which calls the C program unit This directive has no effect when the Munix compilation flag is used meaning you cannot mix UNIX style argument passing and C calling conventions within the same file Syntax for the directive is as follows IMSSATTRIBUTES C work Where work is the name of the subroutine to be cal
136. ck while allowing parallel execution of statements outside the code block The following additional restrictions apply to the ordered pragma e The ordered code block must be a structured block It is illegal to branch into or out of the block A given iteration of a loop with a DO directive cannot execute the same ORDERED directive more than once and cannot execute more than one ORDERED directive 6 12 omp atomic The omp atomic pragma uses the following syntax pragma omp atomic lt C C expression statement gt The omp atomic pragma is semantically equivalent to subjecting the following single C C expression statement to an omp critical pragma The expression statement must be of one of the following forms x lt binary_operator gt expr X x x X where x is a scalar variable of intrinsic type expr is a scalar expression that does not reference x lt binary_operator gt is not overloaded and is one of amp l lt lt or gt gt OpenMP Pragmas for C and C 135 6 13 omp flush The omp flush pragma uses the following syntax pragma omp flush list The omp flush pragma ensures that all processor visible data items or only those specified in list when it s present are written back to memory at the point at which the directive appears 6 14 omp threadprivate The omp threadprivate pragma uses the following syntax pragma omp threadprivate list Where list is a list of varia
137. compilers allow many variations on these general program development steps These variations include the following od Stop the compilation after preprocessing compiling or assembling to save and examine intermediate results e Provide options to the driver that control compiler optimization or that specify various features or limitations e Include as input intermediate files such as preprocessor output compiler output or assembler output 1 2 Invoking the Command level PGI Compilers To translate and link a Fortran C or C program the pgf77 pgf95 pghpf pgcc and pgCC commands do the following e Preprocess the source text file e Check the syntax of the source text Generate an assembly language file e Pass control to the subsequent assembly and linking steps For example if you enter the following simple Fortran program in the file hello f print hello end You can compile it from a shell prompt using the default pgf95 driver options PGI pgf95 hello f Linking PGIS By default the executable output is placed in the file a out or on Windows platforms a filename based on the name of the first source or object file on the command line Use the o option to specify an output file name To place the executable output in the file hello PGI pgf95 o hello f Linking PGIS To execute the resulting program simply type the filename at the command prompt and press the Return or Enter key on your keyboard
138. copying the array to sequential storage The most common cause is passing an ALLOCATABLE array or array expression to a dummy argument that is declared with explicit bounds Declaring the dummy argument as assumed shape with bounds will remove this warning 280 Messages W368 Array valued expression passed to scalar dummy argument The actual argument is an array valued expression but the dummy argument is a scalar variable W369 Dummy argument has different rank than actual argument The actual argument is an array or array valued expression with a different rank than the dummy argument W370 Dummy argument has different shape than actual argument The actual argument is an array or array valued expression with a different shape than the dummy argument this may require copying the actual argument into sequential storage W371 Dummy argument is INTENT IN but may be modified The dummy argument was declared as INTENT IN but analysis has found that the argument may be modified the INTENT IN declaration should be changed W372 Cannot propagate alignment from to The most common cause is when passing an array with an inherited alignment to a dummy argument with non inherited alignment I373 Cannot propagate distribution format from to The most common cause is when passing an array with a transcriptive distribution format to a dummy argument with prescriptive or descriptive distribution format I374 Cann
139. ct files The driver passes files with s extensions to the assembler and files with o so a and lib extensions to the linker Input files with unrecognized extensions or no extension are also passed to the linker Files with a F Capital F or FOR suffix are first preprocessed by the Fortran compilers and the output is passed to the compilation phase The Fortran preprocessor functions similar to cpp for C C programs but is built in to the Fortran compilers rather than implemented through an invocation of cpp This ensures consistency in the pre processing step regardless of the type or revision of operating system under which you re compiling Getting Started Any input files not needed for a particular phase of processing are not processed For example if on the command line you use an assembly language file filename s and the S option to stop before the assembly phase the compiler takes no action on the assembly language file Processing stops after compilation and the assembler does not run in this case compilation must have been completed in a previous pass which created the s file Refer to the following section Output Files for a description of the S option In addition to specifying primary input files on the command line code within other files can be compiled as part of include files using the INCLUDE statement in a Fortran source file or the preprocessor include directive in Fortran source files that use
140. ction from C C illustrates these caller provided extra parameters CHARACTER FUNCTION CHF C1 I CHARACTER C1 INTEGER I END xtern void chf har tmp 10 har c1 9 nt i hf tmp 10 cl Si 9 Qeaa oO The extra parameters tmp and 10 are supplied for the return value while 9 is supplied as the length of cl Refer to Section 10 6 Argument Passing and Return Values for additional information Run time Environment 245 246 Run time Environment Appendix B Messages This appendix describes the various messages that the compiler produces These messages include the sign on message and diagnostic messages for remarks warnings and errors The compiler always displays any error messages along with the erroneous source line on the screen If you specify the Mlist option the compiler places any error messages in the listing file You can also use the v option to display more information about the compiler assembler and linker invocations and about the host system For more information on the Mlist and v options refer to Chapter 3 Command Line Options BI Diagnostic Messages Diagnostic messages provide syntactic and semantic information about your source text Syntactic information includes information such as syntax errors Semantic includes information includes such as unreachable code You can specify that the compiler displays error messages at a certain level with the M
141. ctions pragma must be a structured block and the code in each omp section must be a structured block The available clauses are as defined in Section 6 2 omp parallel and Section 6 6 omp for 6 10 omp parallel sections The omp parallel sections pragma uses the following syntax pragma omp parallel sections clauses pragma omp section lt C C structured block executed by processor i gt pragma omp section lt C C structured block executed by processor j gt ES Clauses private list shared list default shared none firstprivate list lastprivate list reduction operator list copyin list if scalar_expression num_threads scalar_integer_ expression nowait 134 OpenMP Pragmas for C and C The omp parallel sections pragma defines a non iterative work sharing construct without the need to define an enclosing parallel region Semantics are identical to a parallel region containing only an omp sections pragma and the associated structured block 6 11 omp ordered The OpenMP ordered pragma uses the following syntax pragma omp ordered lt C C structured block gt The ordered pragma can appear only in the dynamic extent of a for or parallel for pragma that includes the ordered clause The structured code block appearing after the ordered pragma is executed by only one thread at a time and in the order of the loop iterations This sequentializes the ordered code blo
142. ctive and C C optimization pragma and shows examples of their use Chapter 8 Libraries and Environment Variables discusses PGI support libraries shared object files and environment variables that affect the behavior of the PGI compilers Chapter 9 Fortran C and C Data Types describes the data types that are supported by the PGI Fortran C and C compilers Chapter 10 Inter language Calling provides examples showing how to place C Language calls in a Fortran program and Fortran Language calls in a C program Chapter 11 C Name Mangling describes the name mangling facility and explains the transformations of names of entities to names that include information on aspects of the entity s type and a fully qualified name Appendix A Run time Environment describes the assembly language calling conventions and examples of assembly language calls xii Preface Appendix B Messages provides a list of compiler error messages Appendix C C Dialect Supported lists more details of the version of the C language that PGC supports Hardware and Software Constraints This guide describes versions of the PGI compilers that produce assembly code for x86 and x64 processor based systems Details concerning environment specific values and defaults and system specific features or limitations are presented in the release notes delivered with the PGI compilers Conventions The PGI User s Guid
143. ctives or pragmas is invoked using the mp option described in detail in Chapter 5 OpenMP Directives for Fortran and Chapter 6 OpenMP Pragmas for C and C 2 3 Local and Global Optimization using O Using the PGI compiler commands with the Olevel option you can specify any of the following optimization levels the capital O is for Optimize 00 level zero specifies no optimization A basic block is generated for each language statement 16 Optimization amp Parallelization Sal level one specifies local optimization Scheduling of basic blocks is performed Register allocation is performed 02 level two specifies global optimization This level performs all level one local optimization as well as level two global optimization 03 level three specifies aggressive global optimization This level performs all level one and level two optimizations and enables more aggressive hoisting and scalar replacement optimizations that may or may not be profitable Level zero optimization specifies no optimization O0 At this level the compiler generates a basic block for each statement This level is useful for the initial execution of a program Performance will almost always be slowest using this optimization level Level zero is useful for debugging since there is a direct correlation between the program text and the code generated Level one optimization specifies local optimization O1 The compiler performs schedulin
144. ctorizer transformations can be controlled by arguments to the Mvect command line option The following sections describe the arguments that affect the operation of the vectorizer In addition some of these vectorizer operations can be controlled from within code using directives and pragmas For details on the use of directives and pragmas refer to Chapter 7 Optimization Directives and Pragmas The vectorizer performs the following operations e Loop interchange e Loop splitting e Loop fusion e Memory hierarchy cache tiling optimizations Generation of SSE instructions on processors where these are supported Generation of prefetch instructions on processors where these are supported e Loop iteration peeling to maximize vector alignment e Alternate code generation By default Mvect without any sub options is equivalent to Mvect assoc cachesize 262144 This enables the options for nested loop transformation and various other vectorizer options These defaults may vary depending on the target system 2 5 1 1 Assoc Option The option Mvect assoc instructs the vectorizer to perform associativity conversions that can change the results of a computation due to roundoff error Mvect noassoc disables this option For example a typical optimization is to change one arithmetic operation to another arithmetic operation Optimization amp Parallelization 21 that is mathematically correct but can be computationally
145. ctured block enclosed by the WORKSHARE END WORKSHARE directive pair can consist only of the following types of statements and constructs e Array assignments e Scalar assignments FORALL statements or constructs e WHERE statements or constructs e OpenMP ATOMIC CRITICAL or PARALLEL constructs The work implied by the above statements and constructs is split up between the threads executing the WORKSHARE construct in a way that is guaranteed to maintain standard Fortran semantics The goal of the WORKSHARE construct is to effect parallel execution of non iterative but implicitly data parallel array assignments FORALL and WHERE statements and constructs intrinsic to the Fortran language beginning with Fortran 90 The Fortran structured block contained within a WORKSHARE construct must not contain any user defined function calls unless the function is ELEMENTAL 5 8 BARRIER The OpenMP BARRIER directive uses the following syntax SOMP BARRIER There may be occasions in a parallel region when it is necessary that all threads complete work to that point before any thread is allowed to continue The BARRIER directive synchronizes all threads at such a point in a program Multiple barrier points are allowed within a parallel region The BARRIER directive must either be executed by all threads executing the parallel region or by none of them 5 9 DOA CROSS The C DOACROSS directive is not part of the OpenMP standard but is supported for c
146. d TYPE ATTENDEE ATTLIST 100 ATTLIST 1 SNAME JOHN DOE 9 2 C and C Data Types 9 2 1 C and C Scalars The next table lists C and C scalar data types their size and format The alignment of a scalar data type is equal to its size Table 9 5 Scalar Alignment shows scalar alignments that apply to individual scalars and to scalars that are elements of an array or members of a structure or union Wide characters are supported character constants prefixed with an L The size of each wide character is 4 bytes 180 Fortran C and C Data Types Table 9 4 C C Scalar Data Types Size Data T F t R ata Type bytes orma ange unsigned char 1 ordinal 0 to 255 signed char 1 two s complement 128 to 127 integer unsigned short 2 ordinal 0 to 65535 signed short 2 two s complement 32768 to 32767 integer unsigned int 4 ordinal 0 to 232 1 signed int signed long int 4 two s complement 231 to 231 1 integer unsigned long int 4 ordinal 0 to 232 1 signed long long int 8 two s complement 263 to 263 1 integer unsigned long long int 8 ordinal 0 to 264 1 float 4 IEEE single preci 107 to 10 8 1 sion floating point double 8 IEEE double preci 1030 to 103 8 1 sion floating point long double 8 IEEE double preci 10307 to 1938 1 sion floating point bit field 2 unsigned value 1 to 32 ordinal 0 to 25 7 1 where size is bits the number
147. d Pragmas By default Mconcur without any sub options is equivalent to Mconcur dist block This enables parallelization of loops with blocked iteration allocation across the available threads of execution These defaults may vary depending on the target system 2 6 1 1 Altcode Option The option Mconcur altcode instructs the parallelizer to generate alternate serial code for parallelized loops If altcode is specified without arguments the parallelizer determines an appropriate cutoff length and generates serial code to be executed whenever the loop count is less than or equal to that length If altcode n is specified the serial altcode is executed whenever the loop count is less than or equal to n If noaltcode is specified no alternate serial code is generated 2 6 1 2 Dist Option The option Mconcur dist blocklcyclic option specifies whether to assign loop iterations to the available threads in blocks or in a cyclic round robin fashion Block distribution is the default If cyclic is specified iterations are allocated to processors cyclically That is processor 0 performs iterations 0 3 6 etc processor performs iterations 1 4 7 etc and processor 2 performs iterations 2 5 8 etc 26 Optimization amp Parallelization 2 6 1 3 Cncall Option The option Mconcur cncall specifies that it is safe to parallelize loops that contain subroutine or function calls By default such loops are excluded from consid
148. d by all of the threads This is equivalent to the default status of a variable that is not listed as PRIVATE in a parallel region or DO this same default status is used in C DOACROSS loops as well 5 10 PARALLEL DO The OpenMP PARALLEL DO directive uses the following syntax Syntax SOMP PARALLEL DO CLAUSES lt Fortran DO loop to be executed in parallel gt SOMP END PARALLEL DO Clauses PRIVATE list SHARED list DEFAULT PRIVATE SHARED NONE FIRSTPRIVATE list LASTPRIVATE list REDUCTION operator intrinsic list COPYIN list OpenMP Directives for Fortran 117 IF scalar_logical_ expression NUM_THREADS scalar_integer_expression SCHEDULE type chunk ORDERED The semantics of the PARALLEL DO directive are identical to those of a parallel region containing only a single parallel DO loop and directive Note that the END PARALLEL DO directive is optional The available clauses are as defined in Section 5 2 PARALLEL END PARALLEL and Section 5 6 DO END DO 5 11 PARALLEL WORKSHARE The OpenMP PARALLEL WORKSHARE directive uses the following syntax Syntax SOMP PARALLEL WORKSHARE CLAUSES lt Fortran structured block to be executed in parallel gt SOMP END PARALLEL WORKSHARE Clauses PRIVATE list SHARED list DEFAULT PRIVATE SHARED NONE FIRSTPRIVATE list LASTPRIVATE list REDUCTION operator intrinsic list COPYIN list IF scalar_
149. d chf char tmp 10 char c1 9 int i chf tmp 10 cl amp i 9 If the Fortran function is declared to return a character value of constant length for example CHARACTER 4 FUNCTION CHFO the second extra parameter representing the length must still be supplied but is not used NOTE The value of the character function is not automatically NULL terminated 192 Inter language Calling 10 6 3 Complex Return Values When a Fortran function returns a complex value an argument needs to be added at the beginning of the C C calling function s argument list this argument is the address of the complex return value Example 10 2 COMPLEX Return Values illustrates the extra parameter cplx supplied by the caller Example 10 2 COMPLEX Return Values COMPLEX FUNCTION CF C I INTEGER I END extern void cf typedef struct float real imag cplx cplx c1 int i cf amp c1 amp i 10 7 Array Indices C C arrays and Fortran arrays use different default initial array index values By default C C arrays start at 0 and Fortran arrays start at 1 If you adjust your array comparisons so that a Fortran second element is compared to a C C first element and adjust similarly for other elements you should not have problems working with this difference If this is not satisfactory you can declare your Fortran arrays to start at zero Another difference between Fortran and C C arrays is the storage method
150. d the first application When you build the second application pgcc o app2 main2 c sub c the IPA linker will create two more IPA optimized object files main2 ipa4_ app2 o0 sub_ipa4 appi oo Note there are now three object files for sub c the original sub o and two IPA optimized objects one for each application in which it appears How is the mangled name for the IPA optimized object files generated The mangled name has _ipa appended followed by the decimal number of the length of the executable file name followed by an underscore and the executable file name itself The suffix is changed to oo on Linux or oobj on Windows so linking o or obj does not pull in the IPA optimized objects If the IPA linker determines that the file would not benefit from any interprocedural optimizations it does not have to recompile the file at link time and will use the original object 2 8 Profile Feedback Optimization using Mpfi Mpfo The PGI compilers support many common profile feedback optimizations including semi invariant value optimizations and block placement These are performed under control of the Mpfi Mpfo command line options Optimization amp Parallelization 35 When invoked with the Mpfi option the PGI compilers instrument the generated executable for collection of profile and data feedback information This information can be used in subsequent compilations that include the Mpfo optimization option Mpf
151. d to specify that obj1 lib the DLL s import library should be used to resolve imports Step 3 Ensure that obt dl is in your path then run the executable prog to determine if the DLL was successfully created and linked progl subl adata 11 subl i 12 main adata 12 Should you wish to change obj1 dll without changing the subroutine or function interfaces no rebuilding of prog is necessary Just recreate obj1 dll and the new obj1 dll will be loaded at runtime Example 2 Build a DLL out of a single source file object2 c which exports data and a subroutine using __declspec dllexport Build the main source file prog2 c which uses __declspec dllimport to import the data and subroutine from the DLL object2 c int declspec dllexport data void _ declspec dllexport func2 printf in func2 data d n data prog2 c int declspec dllimport data void _ declspec dllimport func2 int main Libraries and Environment Variables 167 data 11 func2 return 0 Step 1 Create the DLL obt dl and its import library obj2 lib using the following series of commands d i pgcec c object2 c pgcc Mmakedll object2 obj o obj2 dll Step 2 Compile the main program o pgcc Mdll o prog2 prog2 c defaultlib obj2 The Mdll switch causes the compiler to link against the PGI runtime DLLs instead of the PGI runtime static libraries The Mdll switch is required when linki
152. de within the SECTIONS END SECTIONS directives must be a structured block and the code in each SECTION must be a structured block The available clauses are as defined in Section 5 2 PARALLEL END PARALLEL and Section 5 6 DO END DO 5 13 PARALLEL SECTIONS The OpenMP PARALLEL SECTIONS END SECTIONS directive pair uses the following syntax Syntax SOMP PARALLEL SECTIONS CLAUSES SOMP SECTION lt Fortran code block executed by processor i gt SOMP SECTION lt Fortran code block executed by processor j gt SOMP END SECTIONS NOWAIT Clauses PRIVATE list SHARED list DEFAULT PRIVATE SHARED NONE FIRSTPRIVATE list LASTPRIVATE list OpenMP Directives for Fortran 119 REDUCTION operator intrinsic list COPYIN list IF scalar_logical_ expression NUM_THREADS scalar_integer_expression The PARALLEL SECTIONS END SECTIONS directives define a non iterative work sharing construct without the need to define an enclosing parallel region Each section is executed by a single processor If there are more processors than sections some processors will have no work and will jump to the implied barrier at the end of the construct If there are more sections than processors one or more processors will execute more than one section A SECTION directive may only appear within the lexical extent of the enclosing PARALLEL SECTIONS END SECTIONS directives In addition the code with
153. definitions for each of the C C library routines and two required type definitions include lt omp h gt int omp_get_num_threads void 136 OpenMP Pragmas for C and C returns the number of threads in the team executing the parallel region from which it is called When called from a serial region this function returns 1 A nested parallel region is the same as a single parallel region By default the value returned by this function is equal to the value of the environment variable OMP_NUM_THREADS or to the value set by the last previous call to the omp_set_num_threads function defined below include lt omp h gt void omp_set_num_ threads int num_threads sets the number of threads to use for the next parallel region This function can only be called from a serial region of code If it is called from within a parallel region or within a function that is called from within a parallel region the results are undefined This function has precedence over the OMP_NUM_THREADS environment variable include lt omp h gt int omp_get_thread_num void returns the thread number within the team The thread number lies between 0 and omp_get_num_threads 1 When called from a serial region this function returns 0 A nested parallel region is the same as a single parallel region include lt omp h gt int omp_get_max_threads void returns the maximum value that can be returned by calls to omp_get_num_threads If omp_set_nu
154. dely used library Here s an example struct A struct B public A DE operator A amp J 298 C Dialect Supported C 5 By default as well as in cfront compatibility mode there will be no implicit declaration of B operator const B amp whereas in strict ANSI mode B operator A amp is not a copy assignment operator and B operator const B amp is implicitly declared Implicit type conversion between a pointer to an extern C function and a pointer to an extern C function is permitted Here s an example extern C void s type has extern C linkage void pf pf points to an extern C function amp error unless implicit conv is allowed This extension is allowed in environments where C and C functions share the same calling conventions though it is pointless unless DEFAULT_C_AND_CPP_FUNTION_TYPES_ARE_DISTINCT is TRUE When DEFAULT_IMPL_CONV_BETWEEN_C_AND_CPP_FUNCTION_PTRS_ALLOWED is set it is enabled by default it can also be enabled in cfront compatibility mode or with command line option implicit_extern_c_type_conversion It is disabled in strict ANSI mode cfront 2 1 Compatibility Mode The following extensions are accepted in cfront 2 1 compatibility mode in addition to the extensions listed in the following 2 1 3 0 section i e these are things that were corrected in the 3 0 release of cfront The dependent statement of an if while do while or for
155. di rsi Yordx Yorcx r8 r9 even if non integral types are also present in the eight bytes If the eight bytes only contain floating point types these eight bytes will be passed in the first available XMM register of the sequence from xmm0 to xmm7 If the structure or union is larger than eight bytes but smaller than 17 bytes examine the type or types of the fields making up the second eight bytes of the structure or union If the eight bytes contain at least one integral type the eight bytes will be passed in the next available general purpose register of the sequence rdi Grat Yordx Yrcx r8 r9 even if non integral types are also present in the eight bytes If these eight bytes only contain floating point types these eight bytes will be passed in the next available XMM register of the sequence from xmm0 to xmm7 If the first or second eight bytes of the structure or union cannot be passed in a register for some reason the entire structure or union must be passed in memory Passing Arguments on the Stack If there are arguments left after every argument register has been allocated the remaining arguments are passed to the function on the stack The unassigned arguments are pushed on the stack in reverse order with the last argument pushed first Table A 9 Register Allocation for Example A 2 shows the register allocation and stack frame offsets for the function declaration and call shown in the following example Both tab
156. different and generate faster code This option is provided to enable or disable this transformation since roundoff error for such associativity conversions may produce unacceptable results 2 5 1 2 Cachesize Option The option Mvect cachesize n instructs the vectorizer to tile nested loop operations assuming a data cache size of n bytes By default the vectorizer attempts to tile nested loop operations such as matrix multiply using multi dimensional strip mining techniques to maximize re use of items in the data cache 2 5 1 3 SSE Option The option Mvect sse instructs the vectorizer to automatically generate packed SSE SSE2 streaming SIMD extensions and prefetch instructions when vectorizable loops are encountered SSE instructions first introduced on Pentium III and AthlonXP processors operate on single precision floating point data and hence apply only to vectorizable loops that operate on single precision floating point data SSE2 instructions first introduced on Pentium 4 Xeon and Opteron processors operate on double precision floating point data Prefetch instructions first introduced on Pentium HI and AthlonXP processors can be used to improve the performance of vectorizable loops that operate on either 32 bit or 64 bit floating point data See table P 2 for a concise list of processors that support SSE SSE2 and prefetch instructions Note Programs units compiled with Mvect sse will not execute on Pentium Pent
157. e Subsequent calls to RANDOM_SEED without arguments will reset the random seed to the same initial value as the first call Unless the time is exactly the same each time a program is run a different random number sequence will be generated You can force the seed returned by RANDOM_SEED to be constant thereby generating the same sequence of random numbers at each execution of the program by setting the environment variable STATIC_RANDOM_SEED to yes PGI_TERM The stack traceback and just in time debugging functionality is controlled by the PGI_TERM environment variable The run time libraries use the value of PGI_TERM to determine what action to take when a progam abnormally terminates PGI_TERM_DEBUG The PGI_TERM_DEBUG variable may be set to override the default behavior when PGI_TERM is set to debug TMPDIR Can be used to specify the directory that should be used for placement of any temporary files created during execution of the PGI compilers and tools TZ Can be used to explicitly set the time zone and is used in some contexts by the PGC compiler For more information on the possible settings for TZ use the tzselect utility on Linux for a detailed description of possible settings and step by step instructions for setting the value of TZ for a given time zone Libraries and Environment Variables 173 174 Libraries and Environment Variables Chapter 9 Fortran C and C Data Types This chapter describes the scalar and aggre
158. e and produces correct results is not always organized for efficient execution Normally the first step in the program development process involves producing code that executes and produces the correct results This first step usually involves compiling without much worry about optimization After code is compiled and debugged code optimization and parallelization become an issue Invoking one of the PGI compiler commands with certain options instructs the compiler to generate optimized code Optimization is not always performed since it increases compilation time and may make debugging difficult However optimization produces more efficient code that usually runs significantly faster than code that is not optimized The compilers optimize code according to the specified optimization level Using the O Mvect Mipa and Mconcur options you can specify the optimization levels In addition several M lt pgflag gt switches can be used to control specific types of optimization and parallelization This chapter describes the optimization options and describes how to choose optimization options to use with the PGI compilers Chapter 4 Function Inlining describes how to use the function inlining options 2 1 Overview of Optimization In general optimization involves using transformations and replacements that generate more efficient code This is done by the compiler and involves replacements that are independent of the particu
159. e CHARACTER an argument representing the length of the string is passed to a calling function On Linux systems or when using the UNIX calling convention on Windows Munix the compiler places the length argument s at the end of the parameter list following the other formal arguments The length argument is passed by value not by reference Inter language Calling 191 10 6 1 Passing by Value VAL When passing parameters from a Fortran subprogram to a C C function it is possible to pass by value using the VAL function If you enclose a Fortran parameter with VAL the parameter is passed by value For example the following call passes the integer i and the logical bvar by value integer 1li logical 1lbvar call cvalue VAL i VAL bvar 10 6 2 Character Return Values Section 10 3 Functions and Subroutines describes the general rules for return values for C C and Fortran inter language calling There is a special return value to consider When a Fortran function returns a character two arguments need to be added at the beginning of the C C calling function s argument list e the address of the return character or characters e the length of the return character Example 10 1 Character Return Parameters illustrates the extra parameters tmp and 10 supplied by the caller Example 10 1 Character Return Parameters CHARACTER FUNCTION CHF C1 I CHARACTER CL INTEGER I END extern voi
160. e amounts of thread local storage in the form of private variables or local variables in functions or subroutines called within parallel regions The value should be an integer lt n gt concatenated with M or m to specify stack sizes of n megabytes For example setenv MPSTKZ 8M MP_BIND the MP_BIND environment variable can be set to yes or y to bind processes or threads executing in a parallel region to physical processors or to no or n to disable such binding The default is to not bind processes to processors This is an execution time environment variable interpreted by the PGI runtime support libraries It does not affect the behavior of the PGI compilers in any way Note the MP_BIND environment variable is not supported on all platforms MP_BLIST In addition to the MP_BIND variable it is possible to define the thread CPU relationship For example setting MP_BLIST 3 2 1 0 maps CPUs 3 2 1 and 0 to threads 0 1 2 and 3 respectively MP_SPIN When a thread executing in a parallel region enters a barrier it spins on a semaphore MP_SPIN can be used to specify the number of times it checks the semaphore before calling sched_yield on linux or _sleep on Windows These calls cause the thread to be re scheduled allowing other processes to run The default values are 100 Linux and 10000 Windows Libraries and Environment Variables 171 MP WARN By default a warning will be printed to stderr if you execute an OpenMP or auto
161. e and use the information to enable or guide optimizations Optimization no prefetch disable enable generation of prefetch instruc tions Optimization preprocess perform cpp like preprocessing on assembly lan guage and Fortran input source files Miscellaneous prof set profile options function level and line level profiling are supported Code Generation nors determines whether the compiler promotes REAL variables and constants to DOUBLE PRECISION pgf77 pgf95 and pghpf only Optimization nor8intrinsics determines how the compiler treats the intrinsics CMPLX and REAL pgf77 pgf95 and pghpf only Optimization no recursive allocate do not allocate local variables on the stack this allows recursion SAVEd data initial ized or namelist members are always allocated statically regardless of the setting of this switch pgf77 pgf95 and pghpf only Code Generation no reentrant specifies whether the compiler avoids optimiza tions that can prevent code from being reentrant Code Generation no ref_externals do don t force references to names appearing in EXTERNAL statements pgf77 pgf95 and pghpf only Code Generation safeptr instructs the compiler to override data dependen cies between pointers and arrays pgcc and pgCC only Optimization Command Line Options 59 pgflag Description Category
162. e been passed 290 Messages 206 attempt to open a named file as SCRATCH 207 file is already connected to another unit 208 NEW specified for file that already exists 209 OLD specified for file that does not exist 210 dynamic memory allocation failed Memory allocation operations occur only in conjunction with namelist I O The most probable cause of fixed buffer overflow is exceeding the maximum number of simultaneously open file units 211 invalid file name 212 invalid unit number A file unit number less than or equal to zero has been specified 215 formatted unformatted file conflict Formatted unformatted file operation conflict 217 attempt to read past end of file 219 attempt to read write past end of record For direct access the record to be read written exceeds the specified record length 220 write after last internal record 221 syntax error in format string A runtime encoded format contains a lexical or syntax error 222 unbalanced parentheses in format string Messages 291 223 illegal P or T edit descriptor value missing 224 illegal Hollerith or character string in format An unknown token type has been found in a format encoded at run time 225 lexical error unknown token type 226 unrecognized edit descriptor letter in format An unexpected Fortran edit descriptor FED was found in a runtime format item 228 end of file reached without finding group 229 end of file reached whi
163. e computed on the last iteration of the loop If a scalar assigned within a loop is used outside the loop we normally save the last value of the scalar Essentially the value of the scalar on the last iteration is saved in this case wheni N 146 Optimization Directives and Pragmas If the loop is parallelized and the scalar is not assigned on every iteration it may be difficult to determine on what iteration t is last assigned without resorting to costly critical sections Analysis allows the compiler to determine if a scalar is assigned on every iteration thus the loop is safe to parallelize if the scalar is used later An example loop is d t ser RN if x i gt 0 0 E 2 0 else E 2S EN endif EE ee burg enddo V E where t is assigned on every iteration of the loop However there are cases where a scalar may be privatizable If it is used after the loop it is unsafe to parallelize Examine this loop do i LN if x i gt 0 0 t x i Wicd yy Stes es endif enddo We E where each use of t within the loop is reached by a definition from the same iteration Here t is privatizable but the use of t outside the loop may yield incorrect results since the compiler may not be able to detect on which iteration of the parallelized loop t is assigned last The compiler detects the above cases Where a scalar is used after the loop but is not defined on every iteration of the loop parallelization will not occur If y
164. e default is llalign M Generate make dependence lists MD Generate make dependence lists MD filename pgCC only Generate make dependence lists and print them to file filename optk_allow_dollar_in_id_chars pgCC only Accept dollar signs in identifiers pch pgCC only Automatically use and or create a precompiled header file pch_dir directoryname no_ pch_messages pgCC only The directory dirname in which to search for and or create a precompiled header file pgCC only Enable disable the display of a message indi cating that a precompiled header file was created or used p pgCC only Disallow all anachronistic constructs cfront option P Stops after the preprocessing phase and saves the prepro cessed file in filename i preinclude lt filename gt pgCC only Specify file to be included at the beginning of compilation to set system dependent macros types etc 44 Command Line Options Option Description t Control instantiation of template functions EDG option use_pch filename pgCC only Use a precompiled header file of the specified name as part of the current compilation no_Jusing_std pgCC only Enable disable implicit use of the std namespace when standard header files are included X pgCC only Generate cross reference information and place output in specified file EDG option Xm pgCC only A
165. e defines the standard for register allocation The 64 bit AMD64 Architecture AMD64 provides a number of registers All the general purpose registers XMM registers and x87 registers are global to all procedures in a running program 232 Run time Environment Table A 13 Register Allocation Type Name Purpose General Prax return value register Yorbx callee saved DCH pass Ist argument to functions Yordx pass 2nd argument to functions Jorsp stack pointer Yorbp callee saved optional stack frame pointer Prsi callee saved rdi callee saved or8 pass 3rd argument to functions or9 pass 4th argument to functions r10 r11 temporary registers used in syscall sysret instructions or 12 115 callee saved registers XMM Gamm pass Ist floating point argument return value register xmm1 pass 2nd floating point argument Gamm pass 3rd floating point argument xmm3 pass 4th floating point argument xmm4 xmm5 temporary registers xmm6 xmm15 callee saved registers In addition to the registers each function has a frame on the run time stack This stack grows downward from high addresses The next table shows the stack frame organization Run time Environment 233 Table A 14 Standard Stack Frame Position Contents Frame 8n 120 rbp argument eightbyte n previous 80 rbp argument eightbyte 5 88 rbp r9 home 96 rbp
166. e description of name mangling 11 1 Types of Mangling The following entity names are mangled e Function names including non member function names are mangled to deal with overloading Names of functions with extern C linkage are not mangled 205 e Mangled function names have the function name followed by __ followed by F followed by the mangled description of the types of the parameters of the function If the function is a member function the mangled form of the class name precedes the F If the member function is static an S also precedes the F int float Ff class A int f float _1AFf static int g float 9g__1ASFf Special and operator function names like constructors and operator The encoding is similar to that for normal functions but a coded name is used instead of the routine name class A int operator float _pl_1Aff A float ot _1Aff int operator A float pl F1Af Static data member names The mangled form is the member name followed by __ followed by the mangled form of the class name class A static int i i_1A e Names of variables generated for virtual function tables These have names like vtblmangled class name or vtblmangled base class namemangled class name e Names of variables generated to contain runtime type information These have names like Ttype encoding and TIDtype encoding 11 2 Mangling Summary This section lists some of the C entitie
167. e does not contain alternate returns call it from C C as a void function The following table provides the C C data type corresponding to each Fortran data type Run time Environment 229 A 2 4 5 Table A 11 Fortran and C C Data Type Compatibility Fortran Type C C Type Size bytes CHARACTER n x char x n n REAL x float x 4 REAL 4 x float x 4 REAL 8 x double x 8 DOUBLE PRECISION x double x 8 INTEGER x int x 4 INTEGER 1 x signed char x 1 INTEGER 2 x short x 2 INTEGER 4 x int x 4 INTEGER 8 x long x orlong long x 88 LOGICAL x int x 4 LOGICAL 1 x char x 1 LOGICAL 2 x short x 2 LOGICAL 4 x int x 4 LOGICAL 8 x long x orlong long x 88 Table A 12 Fortran and C C Representation of the COMPLEX Type Fortran Type C C Type Size bytes COMPLEX x struct float r I x 8 COMPLEX 8 x struct float r I x 8 COMPLEX 16 x struct double dr di x 16 DOUBLE COMPLEX x struct double dr di x 16 Arrays C C arrays and Fortran arrays use different default initial array index values By default C C 230 Run time Environment arrays start at 0 and Fortran arrays start at 1 A Fortran array can be declared to start at zero Another difference between Fortran and C C arrays is the storage method used Fortran uses column major order and C C use row major order For one dimensional arrays this poses no problems For two dim
168. e of high level languages e g Fortran C C and assembly language in the software development process and should have some level of understanding of programming The PGI compilers are available on a variety of x86 or x64 hardware platforms and operating systems You need to be familiar with the basic commands available on your system Finally your system needs to be running a properly installed and configured version of the compilers For information on installing PGI compilers and tools refer to the Release and Installation notes included with your software Compatibility and Conformance to Standards For further information refer to the following e American National Standard Programming Language FORTRAN ANSI X3 1978 1978 e ISO IEC 1539 1991 Information technology Programming Languages Fortran Geneva 1991 Fortran 90 e ISO IEC 1539 1997 Information technology Programming Languages Fortran Geneva 1997 Fortran 95 e Fortran 95 Handbook Complete ISO ANSI Reference Adams et al The MIT Press Cambridge Mass 1997 e High Performance Fortran Language Specification Revision 1 0 Rice University Houston Texas 1993 http www crpc rice edu HPFF e High Performance Fortran Language Specification Revision 2 0 Rice University Houston Texas 1997 http www crpc rice edu HPFF xi OpenMP Application Program Interface Version 2 5 May 2005 http www openmp org e Programming in VAX Fort
169. e option This directive enables the checking of array bounds when subscripted array references are performed By default array bounds checking is not performed cncall nocncall Loops within the specified scope are considered for parallelization even if they contain calls to user defined subroutines or functions or if their loop counts do not exceed the usual thresholds A nocncall directive cancels the effect of a previous cncall concur noconcur This directive alters the effects of the Mconcur command line option The directive instructs the auto parallelizer to enable auto concurrentization of loops If concur is specified multiple processors will be used to execute loops which the auto parallelizer determines to be parallelizable The noconcur directive disables these transformations This directive affects the compiler only when Mconcur is enabled on the command line depchk nodepchk This directive alters the effects of the Mdepchk command line option When potential data dependencies exist the compiler by default assumes that there is a data dependence that in turn may inhibit certain optimizations or vectorizations nodepchk directs the compiler to ignore unknown data dependencies eqvchk noeqvchk When examining data dependencies noeqvchk directs the compiler to ignore any dependencies between variables appearing in EQUIVALENCE statements Optimization Directives and Pragmas 145 invarif noinvarif The
170. e programs refer to Appendix A 10 1 Overview of Calling Conventions This chapter includes information on the following topics e Functions and subroutines in Fortran C and C e Naming and case conversion conventions Compatible data types e Argument passing and special return values e Arrays and Indexes Win32 calling conventions Default Fortran calling conventions under Win32 differ from those used under Linux and Win64 operating systems Win32 programs compiled using the Munix Fortran command line option use the Linux Win64 convention rather than the default Win32 convention Sections 6 1 through 6 13 describe how to perform inter language calling using the Linux Win64 convention All information in those sections pertaining to compatibility of arguments applies to Win32 as well See Section 10 14 Win32 Calling Conventions for details on the symbol name and argument passing conventions used on Win32 platforms 10 2 Inter language Calling Considerations In general when argument data types and function return values agree you can call a C or C function from Fortran and likewise you can call a Fortran function from C or C You may need to develop special procedures in cases where data types for arguments do not agree For example the Fortran COMPLEX type does not have a matching type in C it is still possible to provide inter language calls but there are no general calling conventions for such cases In this instan
171. e uses the following conventions italic is used for commands filenames directories arguments options and for emphasis Constant Width is used in examples and for language statements in the text including assembly language statements item square brackets indicate optional items In this case item1 is optional item2 item 3 braces indicate that a selection is required In this case you must select either item2 or item3 filename ellipsis indicate a repetition Zero or more of the preceding item may occur In this example multiple filenames are allowed FORTRAN Fortran language statements are shown in the text of this guide using upper case characters and a reduced point size The PGI compilers and tools are supported on both 32 bit and 64 bit variants of the Linux and Windows operating systems on a variety of x86 compatible processors There are a wide variety of releases and distributions of each of these types of operating systems The PGI User s Guide defines the following terms with respect to these platforms x86 a processor designed to be binary compatible with 1386 1486 and previous generation processors from Intel Corporation Used to refer collectively to such processors up to and including 32 bit variants TA32 an Intel Architecture 32 bit processor designed to be binary compatible with x86 processors but incorporating new features such as streaming SIMD extensions SSE for improved performance Prefa
172. ect4 o0 Mmakedll o obj4 dll L lobj3 d i d i Step 2 Compile the main program and link against the import libraries for obj3 dll and obj4 dll pgcc Mdll o prog2 prog2 c L lobj3 lobj4 Step 3 Execute prog to ensure that the DLLs were created properly prog2 func _3a calling a routine in obj4 dll func Ab func_4a calling a routine in obj3 dll func _3b 164 Libraries and Environment Variables 8 4 Creating and Using Dynamic Link Libraries on Windows Some of the PGI compiler runtime libraries are available in both static library and dynamic link library DLL form for Windows The static libraries are always used by default To use the PGI Workstation C and Fortran compilers to create an executable that links to the runtime DLLs use the compiler flag Mdll at the link step There are several differences between static and dynamic link libraries Both libraries are used when resolving external references when linking an executable but the process differs for each type of library When linking with a static library the code needed from the library is incorporated into the executable When linking with a DLL external references are resolved using the DLL s import library not the DLL itself The code in the DLL associated with the external references does not become a part of the executable The DLL is loaded when the executable that needs it is run For the DLL to be loaded in this manner the DLL must be in your path
173. ed A nonstandard friend declaration may introduce a new type A friend declaration that omits the elaborated type specifier is allowed in default mode but in cfront mode the declaration is also allowed to introduce a new type name struct A friend B e The third operator of the operator is a conditional expression instead of an assignment expression as it is in the current X3J16 WG21 Working Paper e A reference to a pointer type may be initialized from a pointer value without use of a temporary even when the reference pointer type has additional type qualifiers above those present in the pointer value For example int p const int amp r p No temporary used A reference may be initialized with a null C Dialect Supported 301 302 C Dialect Supported Index A Auto paralleltzaton 25 B Basic Bl Ck ansni endive eege tei tates 13 Bounds checking A 79 Cc C Name Mangling ceceeceeeeeeeeeeeeeeees 205 C Standard Template Library 170 C C Builtin Functons eee eeeeeeeeeeee 159 C C Math Header File 159 Cache tiling failed cache tiling s e 81 With MV Ct setii ci peiiini ea anaitasun 77 Command line Optons 3 39 54 SHH i AEE AETA A S EEEE EE 45 RE 94 Stee Eege 95 Seen 96 Dvteswapio tee 46 R E E ERTE 46 AE EE death T A 46 fr p2 enee ieie Ees 96 SLO D EE S D EE E E E ETES 96 Create e EE 97 BEE 47 CAG We EEN 97 diag_remark cceeecceeseeeeesseeeeseeeeeeeeees
174. ed 99 CPU The result is a speed up of 45 over the equivalent scalar i e non SSE version of the program Speed up realized by a given loop or program can vary widely based on a number of factors Performance improvement using vector SSE or SSE2 instructions is most effective when the vectors of data are resident in the data cache e If data is aligned properly performance will be better in general than when using vector SSE operations on unaligned data e If the compiler can guarantee that data is aligned properly even more efficient sequences of SSE instructions can be generated SSE2 vector instructions can operate on 4 single precision elements concurrently but only 2 double precision elements As a result the efficiency of loops that operate on single precision data can be higher Note Compiling with Mvect sse can result in numerical differences from the generated executable Certain vectorizable operations for example dot products are sensitive to order of operations and the associative transformations necessary to enable vectorization or parallelization 2 6 Auto Parallelization using Mconcur With the Mconcur option the compiler scans code searching for loops that are candidates for auto parallelization Mconcur must be used at both compile time and link time When the parallelizer finds opportunities for auto parallelization it parallelizes loops and you are informed of the line or loop being parallelized if
175. ed with Mipa If you compile with Mipa and link without Mipa the linker is invoked on the original object files A legal executable will be generated while this will not have the benefit of interprocedural optimizations any other optimizations will apply 34 Optimization amp Parallelization What if I compile without Mipa and link with Mipa At link time the IPA linker must have summary information about all the functions or routines used in the program This information is created only when a file is compiled with Mipa If you compile a file without Mipa and then try to get interprocedural optimizations by linking with Mipa the IPA linker will issue a message that some routines have no IPA summary information and will proceed to run the system linker using the original object files If some files were compiled with Mipa and others were not it will determine the safest approximation of the IPA summary information for those files not compiled with Mipa and use that to recompile the other files using interprocedural optimizations Can I build multiple applications in the same directory with Mipa Yes Suppose you have three source files mam ec main2 c sub c where sub c is shared between the two applications When you build the first application with Mipa pgcc o appl mainl c sub c the IPA linker will create two IPA optimized object files mainl ipa4_ appl oo sub_ipa4 appl oo and use them to buil
176. eed up execution by eliminating the function call and return overhead Function inlining may also create opportunities for other types of optimization Function inlining is not always beneficial When used improperly it may increase code size and generate less efficient code Profile Feedback Optimization PFO Profile feedback optimization makes use of information from a trace file produced by specially instrumented executables which capture and save information on branch frequency function and subroutine call frequency semi invariant values loop index ranges and other input data dependent 14 Optimization amp Parallelization information that can only be collected dynamically during execution of a program By definition use of profile feedback optimization is a two phase process compilation and execution of a specially instrumented executable followed by a subsequent compilation which reads a trace file generated during the first phase and uses the information in the trace file to guide compiler optimizations 2 2 Getting Started with Optimizations Your first concern should be getting your program to execute and produce correct results To get your program running start by compiling and linking without optimization Use the optimization level OU or select g to perform minimal optimization At this level you will be able to debug your program easily and isolate any coding errors exposed during porting to x86 or x64 platforms
177. eeeeeceaaeeeeeeeeeseaeeeseeeeeteaeeneneees 11 Optimization A ParalleliZation ccceccseseeeeeseseeeeeeeeeeeeeeeeeeeeees 13 Overview Of Optimization cccceccseeeeceeeeeeceeeeeeeeeeeeeeseeeeeeaeeeseneeeeeeeeees 13 Getting Started with Optimizations ccceeeceeeceeeeeeeeeeeeeeeeeeeeeeetaeeeeeneeees 15 Local and Global Optimization USING 03 16 Scalar SSE Code Generation ccccccceeceeeeeeeeeneeeeeeeeeeeaeeeeeeeeseeaeeneeeees 18 Loop Unrolling using Muno 0 2 ee eeceeeeeeeneeeeeeeeneeeeeeeeaaeeeeeeeeaeeeeeneaas 19 Vectorization USING MVECt serierna aai anaa RAE RA E 20 Vectorization SUD OptiONS 21 ASSOC le Le NEE 21 Gachesize BIO iieeeber iergendee Miesch Ee ueet A e eieiei Marleen 22 E e EE 22 Prefetch Option ugeet sac a ine Geer 22 Vectorization Example Using SSE SSE2 Instructions 23 Auto Parallelization using Meonceur 0 eeeeeeeeee eect eee teeeeeeeeeteeeeeeeeeeeneaees 25 Auto parallelization SUD OptionS 26 iii 2 6 1 1 2 6 1 2 2 6 1 3 2 6 2 2 6 2 1 2 6 2 2 2 6 2 3 2 6 2 4 2 7 2 7 1 2 7 2 2 7 3 2 7 4 2 7 5 2 7 6 2 7 7 2 7 8 2 8 2 9 2 10 2 11 Chapter 3 3 1 3 2 Chapter 4 4 1 4 1 1 5 1 Altcode elt Le E 26 SEI eege Ee eege 26 Creal e le 27 Loops That Fail to Paralelizmi a E A 27 dall E e EE 27 TIMING Neel CN 27 SCALANS wesc bedside lect ege EE A A tel ees diesen EE leede teats 28 Scalar Last EE 29 Inter Procedural Analysis and Optimization using Mip
178. eeeeeeeseeneeeeeeeeenes 209 A 1 Linux86 and Win32 Programming Model ceeceeeeeeeteeeeeeeeeeteeeeeeenee 209 A 1 1 Function Calling Sequence c cceceteeeeeeeeeeeeeeeeaeeeeeaeeeseneeeteaeeeeeeeeeed 209 A 1 1 1 Register Usage Conventions ccccccceeeeeeeeeeceeeeeeeeeeseaeeeeneeeeseaeeeneneees 209 A 1 2 Function Return Values 212 A 1 3 ArQUMENt PASSING BEE 214 A 2 Linux86 64 Programming Model A 218 A 2 1 Function Calling Sequence c ccececeeeeeeeeeeteeeeeeeaeeeseaeeeseneeeseaeeeeeaeeeeed 218 A 2 2 Function Return Values 221 A 2 3 Argument PASSING EE 222 A 2 4 Linux86 64 Fortran Supplement cceeeeeeceeeeeeeeeeeeeeaeeeeeeeeeeseaeeeseneees 226 A 2 4 1 Fortran Fundamental Types 0 cceccsccceeeesceeeeeeeeecneeeeeeeaeeeeeeenaaeeeeeneaa 227 A 2 4 2 Naming Conventions cccccccecececceeeeeneeececaeeeseaeeeseaeeesseaeeeeeeeeesiaaeeesaes 228 A 2 4 3 Argument Passing and Return Conventions cccceceseeeeeteeeeeseeteees 228 A 2 4 4 Inter language Calling c ccceeeccceceeceeeeeeeceeeeeeeaaeeeceeeeseaaeeeeeeeeestaaeeneneees 228 A 2 4 5 AAV S EE ee EREM et 230 A 2 4 6 Gommon BIOCKS iesi iiaeia iin cain hae at 231 A 3 Win64 Programming Model cececeeeeeeeeneeeeeeeeeeeeeeseeenaeeeeeeeaaeeeeneeaas 232 A 3 1 Function Calling Sequence cctccceeeeeeeseeeeeeeteeeteeeeesetaeeeseneeeteaeeeeeeeeend 232 A 3 2 Function Return Values 235 A 3 3 Argument Passing EE 23
179. eeneeceeeeeenseeeeeeenneaeeeeeenenaaes 183 Bit field Alignment eects eeeeeeeeeeeeeeeeeeesaaeeeeeeteaeeeeeeseeeeeeeeeneeaaeess 184 Other Type Keywords in C and CA 185 Inter lanQuage Calling ccccsseeceseeeeeeeeseeeeeeeeseeeeeeneeeneeeeneeeeneeens 187 Overview of Calling Conventions 0 eceeeceeeeeeeeeeeeeeeeeeaeeeeeeeeeeteaeeeeaes 187 Inter language Calling Considerations ccccceeeeeeeceeeeeeteeeetteeeeeneeees 187 Functions ANd Subroutines 0 2 eeeeeece cee eeeeeeee cease eeeeeeeseaeeseeneeeseaeesseneees 188 Upper and Lower Case Conventions Underscores ccceeeeerees 189 Compatible Data Ty POS szesnaste eaae aa an aa AR iain 189 Fortran Named Common BIOCKS cccccceeeeeeeeeeeeeeeeeeeeeseneeseceeeeeeeeees 190 Argument Passing and Return Values eeecseeeeeeeenneeeeeeeenaeeeeeeeeaaaes 191 Passing by Value VAL eeeececceeeeeneeeee eee eeceee eee eaaeeeeeeeaaeeeeeetaaeeeeeeeaa 192 Character Return Values AA 192 Complex Return Values ecccecesceeeeeeeeeeneeececeeeeeneeeeeaeeeesaaeesecaeeeenaeeesaas 193 e lu EE 193 Example Fortran Calling C A 193 Example C Calling Fortran cccccccscceeeeeeeeeeeeeeeeeeeeeeaeeeeeneeeseaaeesennees 194 Example C Calling Co o c ececcceecccceceeeeceeeeeeeeeaeeeceeeeesaeeseeneeessaeenennees 196 Example C Calling C ececcccccceeceeeecenneeceeeeeeeeaaeeeceeeeesaeeseeeeesnaeeseanees 197 Example Fortran Calling C cccccecc
180. el region before this subregion and then starting it up again after this subregion the MASTER END MASTER directive pair let you conveniently designate code that executes on the master thread and is skipped by the other threads There is no implied barrier on entry to or exit from a MASTER END MASTER section of code Nested master sections are ignored Branching into or out of a master section is not supported PROGRAM MASTER _USE INTEGER A 0 1 INTEGER omp_get_thread_num A 1 SOMP PARALLEL A omp_get_thread_num omp_get_thread_num SOMP MASTER PRINT YOU SHOULD ONLY SEE THIS ONCE SOMP END MASTER SOMP END PARALLEL PRINT A O A O A 1 A 1 END 112 OpenMP Directives for Fortran 3 5 SINGLE END SINGLE The OpenMP SINGLE END SINGLE directive uses the following syntax SOMP SINGLE Clauses lt Fortran code in body of SINGLE processor section gt SOMP END SINGLE NOWAIT Clauses PRIVATE list FIRSTPRIVATE list COPYPRIVATE list In a parallel region of code there may be a sub region of code that will only execute correctly on a single thread Instead of ending the parallel region before this subregion and then starting it up again after this subregion the SINGLE END SINGLE directive pair lets you conveniently designate code that executes on a single thread and is skipped by the other threads There is an implied barrier on exit from a SINGLE END SINGLE section of code unless the op
181. ensional arrays where there are an equal number of rows and columns row and column indexes can simply be reversed Inter language function mixing is not recommended for arrays other than single dimensional arrays and square two dimensional arrays Structures Unions Maps and Derived Types Fields within Fortran structures and derived types and multiple map declarations within a Fortran union conform to the same alignment requirements used by C structures A 2 4 6 Common Blocks A named Fortran common block can be represented in C C by a structure whose members correspond to the members of the common block The name of the structure in C C must have the added underscore For example the Fortran common block INTEGER I J COMPLEX C DOUBLE COMPLEX CD DOUBLE PRECISION D COMMON COM i j c cd d is represented in C with the following equivalent extern struct int i int j struct float real imag c struct double real imag cd double d com and in C with the following equivalent extern C struct int i int j struct float real imag c struct double real imag cd double d com Note that the compiler provided name of the BLANK COMMON block is implementation specific Run time Environment 231 Calling Fortran COMPLEX and CHARACTER functions from C C is not as straightforward as calling other types of Fortran functions Additional arguments must be passed to the Fortran functi
182. ent which in turn must precede an executable statement S071 Executable statements not allowed in block data 256 Messages 072 Assignment operation illegal to The destination of an assignment operation must be a variable array reference or vector reference The assignment operation may be by way of an assignment statement a data statement or the index variable of an implied DO loop The compiler has determined that the identifier used as the destination is not a storage location The error message attempts to indicate the type of entity used entry point An assignment to an entry point that was not a function procedure was attempted external procedure An assignment to an external procedure or a Fortran intrinsic name was attempted if the identifier is the name of an entry point that is not a function an external procedure 073 Intrinsic or predeclared cannot be passed as an argument 074 Illegal number or type of arguments to The indicated symbol is an intrinsic or generic function or a predeclared subroutine or function requiring a certain number of arguments of a fixed data type 075 Subscript substring or argument illegal in this context for This can happen if you try to doubly index an array such as ra 2 3 This also applies to substring and function references S076 Subscripts specified for non array variable S077 Subscripts omitted from array S078 Wrong number of subscripts specified
183. ents Many of the directives are presented in pairs and must be used in pairs In the examples given with each section the routines omp_get_num_threads and omp_get_thread_num are used refer to Section 5 18 Run time Library Routines for more information They return the number of threads currently in the team executing the parallel region and the thread number within the team respectively 5 2 PARALLEL END PARALLEL The OpenMP PARALLEL END PARALLEL directive is supported using the following syntax Syntax SOMP PARALLEL Clauses lt Fortran code executed in body of parallel region gt SOMP END PARALLEL Clauses PRIVATE list SHARED list DEFAULT PRIVATE SHARED NONE FIRSTPRIVATE list REDUCTION operator intrinsic list COPYIN list IF scalar_logical_ expression NUM_THREADS scalar_integer_expression This directive pair declares a region of parallel execution It directs the compiler to create an executable in which the statements between PARALLEL and END PARALLEL are executed by multiple lightweight threads The code that lies between PARALLEL and END PARALLEL is called a parallel region The OpenMP parallelization directives support a fork join execution model in which a single thread executes all statements until a parallel region is encountered At the entrance to the parallel region a system dependent number of symmetric parallel threads begin executing all statements in the parallel
184. eous and before the start of a parallel region Useful when many private variables are declared concur enable auto concurrentization of loops Multiple Optimization processors or cores will be used to execute paral lelizable loops cpp run the PGI cpp like pre processor without per Miscellaneous forming subsequent compilation steps Command Line Options 55 pgflag Description Category cray Force Cray Fortran CF77 compatibility pgf77 Optimization pgf95 and pghpf only no daz Do don t treat denormalized numbers as zero Code Generation no dclchk determines whether all program variables must be Fortran Language declared pgf77 pgf95 and pghpf only no defaultunit determines how the asterisk character is Fortran Language treated in relation to standard input and standard output regardless of the status of I O units 5 and 6 pgf77 pgf95 and pghpf only no depchk checks for potential data dependencies Optimization no dlines determines whether the compiler treats lines con Fortran Language taining the letter D in column one as executable statements pgf77 pgf95 and pghpf only dll Link with the DLL version of the runtime libraries Miscellaneous Windows only dollar specifies the character to which the compiler Fortran Language maps the dollar sign code pgf77 pgf95 and pghpf only dwarf1 when used with g genera
185. er Run time Environment 221 Functions Returning Structures or Unions A function can use either registers or memory to return a structure or union The size and type of the structure or union determine how it is returned If a structure or union is larger than 16 bytes it is returned in memory allocated by the caller To determine whether a 16 byte or smaller structure or union can be returned in one or more return registers examine the first eight bytes of the structure or union The type or types of the structure or union s fields making up these eight bytes determine how these eight bytes will be returned If the eight bytes contain at least one integral type the eight bytes will be returned in rax even if non integral types are also present in the eight bytes If the eight bytes only contain floating point types these eight bytes will be returned in xmm0 If the structure or union is larger than eight bytes but smaller than 17 bytes examine the type or types of the fields making up the second eight bytes of the structure or union If these eight bytes contain at least one integral type these eight bytes will be returned in rdx even if non integral types are also present in the eight bytes If the eight bytes only contain floating point types these eight bytes will be returned in xmm1 If a structure or union is returned in memory the caller provides the space for the return value and passes its address to the function as a
186. er eax and can be used to access the object Both the calling and the called functions must cooperate to pass the return value successfully e The calling function must supply space for the return value and pass its address in the stack frame e The called function must use the address from the frame and copy the return value to the object so supplied e The called function must remove this address from the stack before returning Failure of either side to meet its obligation leads to undefined program behavior The standard function calling sequence does not include any method to detect such failures nor to detect structure and union type mismatches Therefore you must declare the function properly The following table illustrates the stack contents when the function receives control after the call instruction and when the calling function again receives control after the ret instruction Run time Environment 213 Table A 3 Stack Contents for Functions Returning struct union Position After Call After Return Position 4n 8 esp argument word n argument word n 4n 4 esp 8 esp argument word 1 argument word 1 0 esp 4 esp value address undefined 0 esp return address The following sections of this appendix describe where arguments appear on the stack The examples are written as if the function prologue described above had been used ALA Argument Passing
187. er case This selection affects the linking process If you compile and link the same source code using Mupcase on one occasion and Mnoupcase on another you may get two different executables depending on whether the source contains uppercase letters The standard libraries are compiled using Mnoupcase Default For arguments that you do not specify the defaults are as follows nobackslash noiomutex nodclchk noonetrip nodefaultunit nosave nodlines nounixlogical dollar noupcase M lt pgflag gt C C Language Controls This section describes the M lt pgflag gt options that affect C C language interpretations by the PGI C and C compilers These options are only valid to the pgcc and pgCC compiler drivers Masmkeyword 68 Syntax instructs the compiler to allow the asm keyword in C source files The syntax of the asm statement is as follows Command Line Options asm statement Where statement is a legal assembly language statement The quote marks are required Mnoasmkeyword instructs the compiler not to allow the asm keyword in C source files If you use this option and your program includes the asm keyword unresolved references will be generated Mdollar char char specifies the character to which the compiler maps the dollar sign The PGCC compiler allows the dollar sign in names ANSI C does not allow the dollar sign in names Mfcon instructs the compiler to treat floating point c
188. eration for auto parallelization Also no minimum loop count threshold must be satisfied before parallelization will occur and last values of scalars are assumed to be safe The environment variable NCPUS is checked at runtime for a parallel program If NCPUS is set to 1 a parallel program runs serially but will use the parallel routines generated during compilation If NCPUS is set to a value greater than 1 the specified number of processors will be used to execute the program Setting NCPUS to a value exceeding the number of physical processors can produce inefficient execution Executing a program on multiple processors in an environment where some of the processors are being time shared with another executing job can also result in inefficient execution As with the vectorizer the Mconcur option can speed up code if it contains well behaved countable loops and or computationally intensive nested loops that operate on arrays However it is possible that some codes will show a decrease in performance on multi processor systems when compiled with Mconcur due to parallelization overheads memory bandwidth limitations in the target system false sharing of cache lines or other architectural or code generation factors For this reason it is recommended that you check carefully whether particular program units or loops show improved performance when compiled using this option If the compiler is not able to successfully auto parallelize your app
189. erent distribution format in another subprogram W350 Common has different alignment The array was declared in a common block with a different alignment in another subprogram W351 Wrong number of arguments passed to The subroutine or function statement for the given subprogram has a different number of dummy arguments than appear in the call W352 Wrong number of arguments passed to when bound to The subroutine or function statement for the given subprogram has a different number of dummy arguments than appear in the call to the EXTERNAL name given W353 Subprogram is missing A call to a subroutine or function with this name appears but it could not be found or analyzed I354 Subprogram is not called No calls to the given subroutine or function appear anywhere in the program W355 Missing argument in call to A nonoptional argument is missing in a call to the given subprogram I356 Array section analysis incomplete Interprocedural analysis for array section arguments is incomplete some information may not be available for optimization I357 Expression analysis incomplete Messages 279 Interprocedural analysis for expression arguments is incomplete some information may not be available for optimization W358 Dummy argument is EXTERNAL but actual is not subprogram The call statement passes a scalar or array to a dummy argument that is declared EXTERNAL W359 SUBROUTINE passed to FUNCTION dum
190. essors while the tp p7 64 option results in generation of code that is supported on and optimized for Intel x64 processors Performance of k8 64 or k8 64e code executed on Intel x64 processors or of p7 64 code executed on AMD x64 processors can often be significantly less than that obtained with a Command Line Options 91 native binary The tp x64 option results in generation of unified binary object and executable files which are supported on and include optimized code sequences for both AMD and Intel x64 processors Following is a list of possible sub options to tp and the processors they are intended to target k8 32 generate 32 bit code for AMD Athlon64 AMD Opteron and compatible processors k8 64 generate 64 bit code for AMD Athlon64 AMD Opteron and compatible processors k8 64e generate 64 bit code for AMD Opteron Revision E AMD Turion and compatible processors p6 generate 32 bit code for Pentium Prof and AthlonXP compatible processors p7 generate 32 bit code for Pentium 4 and compatible processors p7 64 generate 64 bit code for Intel P4 Xeon EM64T and compatible processors core2 generate 32 bit code for Intel Core 2 Duo and compatible processors core2 64 generate 64 bit code for Intel Core 2 Duo EM64T and compatible processors piii generate 32 bit code for Pentium II and compatible processors including support for single precision vector code using SSE instructions px generate 32 bit code that is useable on
191. esult in faster code if the program uses only float parameters 154 Optimization Directives and Pragmas Note Since ANSI C specifies that routines must convert float parameters to double in non prototyped functions this pragma results in non ANSI conforming code 7 7 Scope of C C Pragmas and Command Line Options This section presents several examples showing the effect of pragmas and the use of the pragma scope indicators Note during compilation a pragma either turns an option on or turns an option off Pragmas apply to the section of code corresponding to the specified scope that is the entire file the following loop or the following or current routine For pragmas that have only routine and global scope there are two rules for determining the scope of a pragma We cover these special scope rules at the end of this section In all cases pragmas override a corresponding command line option Consider the program main float a 100 100 b 100 100 c 100 100 int time maxtime n i j maxtime 10 n 100 for time 0 time lt maxtime time for j 0 j lt n j for i 0 i lt n i c i j a i j b i j When this is compiled using the Mvect command line option both interior loops are interchanged with the outer loop Pragmas alter this behavior either globally or on a routine or loop by loop basis To ensure that vectorization is not applied use the novector pragma with global scope main
192. et_lock integer var causes the calling thread to wait until the specified lock is available The thread gains ownership of the lock when it is available It is illegal to make a call to this routine if integer_var has not been associated with a lock subroutine omp_unset_lock integer_var causes the calling thread to release ownership of the lock associated with integer_var It is illegal to make a call to this routine if integer_var has not been associated with a lock logical function omp_test_lock integer var causes the calling thread to try to gain ownership of the lock associated with integer_var The function returns TRUE if the thread gains ownership of the lock and FALSE otherwise It is illegal to make a call to this routine if integer_var has not been associated with a lock OpenMP Directives for Fortran 123 5 19 Environment Variables OMP_NUM_THREADS specifies the number of threads to use during execution of parallel regions The default value for this variable is 1 For historical reasons the environment variable NCPUS is supported with the same functionality In the event that both OMP_NUM_THREADS and NCPUS are defined the value of OMP_NUM_THREADS takes precedence NOTE OMP_NUM_THREADS threads will be used to execute the program regardless of the number of physical processors available in the system As a result you can run programs using more threads than physical processors and they will execute correctly Ho
193. fine a macro name without specifying a value the preprocessor assigns the string 1 to the macro name Usage In the following example the macro PATHLENGTH has the value 256 until a subsequent compilation If the D option is not used PATHLENGTH s value is set to 128 pgf95 DPATHLENGTH 256 myprog F Where the source text is ifndef PATHLENGTH define PATHLENGTH 128 endif SUBROUTINE SUB CHARACTER PATHLENGTH path END Cross reference U Command Line Options 47 dryrun Use the dryrun option to display the invocations of the compiler assembler and linker but do not execute them These invocations are command lines created by the compiler driver from the PGIRC file and the command line supplied with dryrun Default The compiler does not display individual phase invocations Usage The following command line requests verbose invocation information pgf95 dryrun myprog f Cross reference Minfo V H E Stops after the preprocessing phase Use the E option to halt the compilation process after the preprocessing phase and display the preprocessed output on the standard output Default The compiler produces an executable file Usage In the following example the compiler displays the preprocessed myprog f on the standard output pgf95 E myprog f Cross reference See the options C c Mkeepasm o F S F Stops compilation after the preprocessing phase Use t
194. fined in object4 c and vice versa This situation of mutual imports requires two steps to build each DLL object3 c extern void func _4b void void func _3a void printf func 3a calling a routine in obj4 dl11l n func _4b void func _3b void printf func _3b n object4 c Libraries and Environment Variables 163 extern void func_3b void void func _4a void printf func_4a calling a routine in obj3 d11 n func _3b void func _4b void printf func _4b n prog2 c extern void func _3a void extern void func _4a void int main func _3a func _4a Step 1 To make obj3 dll and obj4 dll first compile the source and create an import library for each DLL that will be built pgce c object3 c pgcc object3 o0 Mmakedll o obj3 dll out implib obj3 1ib Creating library file obj3 1lib object3 o text 0x24 object3 c undefined reference to func_4b pgcc c object4 c pgcc object4 o0 Mmakedll o obj4 dll out implib obj4 1lib Creating library file obj4 1lib object4 o0 text 0x24 object4 c undefined reference to func_3b d i d i The undefined reference errors are to be expected in the first step of building DLLs with mutual imports These errors will be resolved in the next step where each DLL is built and linked against the import library previously created for the other DLL pgcc object3 o Mmakedll o obj3 dll L lobj4 pgcc obj
195. for Fortran x expr operator x X intrinsic x expr X intrinsic expr x where x is a scalar variable of intrinsic type expr is a scalar expression that does not reference x intrinsic is one of MAX MIN LAND IOR or IEOR and operator is one of AND OR EQV or NEQV 5 16 FLUSH The OpenMP FLUSH directive uses the following syntax SOMP FLUSH list The FLUSH directive ensures that all processor visible data items or only those specified in list when it s present are written back to memory at the point at which the directive appears 5 17 THREADPRIVATE The OpenMP THREADPRIVATE directive uses the following syntax SOMP THREADPRIVATE list Where list is a comma separated list of named variables to be made private to each thread or named common blocks to be made private to each thread but global within the thread Common block names must appear between slashes i e common_blockn This directive must appear in the declarations section of a program unit after the declaration of any common blocks or variables listed On entry to a parallel region data in a THREADPRIVATE common block or variable is undefined unless COPYIN is specified on the PARALLEL directive When a common block or variable that is initialized using DATA statements appears in a THREADPRIVATE directive each thread s copy is initialized once prior to its first use The following restrictions apply to the THREADPRIVATE directive
196. func a function in the source text Multiple functions can be listed comma separated name func A function name which is a string matching func a function in the source text size n A size which instructs the compiler to select functions with a statement count less than or equal to n the specified size Note the size n may not exactly equal the number of statements in a selected function the size parameter is used as a rough gauge 101 levels n A level number which represents the number of function calling levels to be inlined The default number is one 1 Using a level greater than one indicates that function calls within inlined functions may be replaced with inlined code This allows the function inliner to automatically perform a sequence of inline and extract processes lib file ext A library file name This instructs the inliner to inline the functions within the library file file ext Create the library file using the Mextract option If no inline library is specified functions are extracted from a temporary library created during an extract prepass If you specify both a function name and a size n the compiler inlines functions that match the function name or have n or fewer statements If a keyword name lib or size is omitted then a name with a period is assumed to be an inline library a number is assumed to be a size and a name without a period is assumed to be a function name In the following
197. g of basic blocks as well as register allocation This optimization level is a good choice when the code is very irregular that is it contains many short statements containing IF statements and the program does not contain loops DO or DO WHILE statements For certain types of code this optimization level may perform better than level two O2 although this case rarely occurs The PGI compilers perform many different types of local optimizations including but not limited to e Algebraic identity removal Constant folding Common subexpression elimination e Local register optimization e Peephole optimizations Redundant load and store elimination Strength reductions Level two optimization O2 or O specifies global optimization The fast option generally will specify global optimization however the fast switch will vary from release to release depending on a reasonable selection of switches for any one particular release The O or O2 level performs all level one local optimizations as well as global optimizations Control flow analysis is applied and global registers are allocated for all functions and subroutines Loop regions are given special consideration This optimization level is a good choice when the program contains loops the loops are short and the structure of the code is regular The PGI compilers perform many different types of global optimizations including but not limited to e Branch to branc
198. g templates exceptions and support for the anachronisms described in section 18 of the ARM This is the same language defined by the language reference for ATT s cfront version 3 0 1 with the addition of exceptions PGC optionally accepts a number of features erroneously accepted by cfront version 2 1 Using the b option PGC accepts these features which may never have been legal C but have found their way into some user s code Command line options provide full support of many C variants including strict standard conformance PGC provides command line options that enable the user to specify whether anachronisms and or cfront 2 1 compatibility features should be accepted Refer to Section C 4 for details on features that are not part of the ARM but are part of the ANSI C working draft X3J16 WG21 C 1 Anachronisms Accepted The following anachronisms are accepted when anachronisms are enabled when the p option is not used e overload is allowed in function declarations It is accepted and ignored e Definitions are not required for static data members that can be initialized using default initialization This anachronism does not apply to static data members of template classes they must always be defined e The number of elements in an array may be specified in an array delete operation The value is ignored e A single operator and operator function can be used to overload both prefix and postfix operations
199. gate data types recognized by the PGI Fortran C and C compilers the format and alignment of each type in memory and the range of values each type can take on x86 or x64 processor based systems running a 32 bit operating system For more information on x86 specific data representation refer to the System V Application Binary Interface Processor Supplement listed in the This chapter specifically does not address x64 processor based systems running a 64 bit operating system because the application binary interface ABI for those systems is still evolving See http www x86 64 org abi pdf for the latest version of this ABI 9 1 Fortran Data Types 9 1 1 Fortran Scalars A scalar data type holds a single value such as the integer value 42 or the real value 112 6 The next table lists scalar data types their size format and range Table 9 2 Real Data Type Ranges shows the range and approximate precision for Fortran real data types Table 9 3 Scalar Type Alignment shows the alignment for different scalar data types The alignments apply to all scalars whether they are independent or contained in an array a structure or a union 175 Table 9 1 Representation of Fortran Data Types Fortran Data Type Format Range INTEGER 2 s complement integer 231 to 231 1 INTEGER 2 2 s complement integer 32768 to 32767 INTEGER 4 same as INTEGER INTEGER 8 same as
200. h as a reduction loop with optimized code sequences or function calls When the vectorizer finds vectorization opportunities it internally rearranges or replaces sections of loops the vectorizer changes the code generated your source code s loops are not altered In addition to performing these 20 Optimization amp Parallelization loop transformations the vectorizer produces extensive data dependence information for use by other phases of compilation and detects opportunities to use vector or packed Streaming SIMD Extensions SSE instructions on processors where these are supported The Mvect option can speed up code which contains well behaved countable loops which operate on large REAL REAL 4 REAL 8 INTEGER 4 COMPLEX or COMPLEX DOUBLE arrays in Fortran and their C C counterparts However it is possible that some codes will show a decrease in performance when compiled with Mvect due to the generation of conditionally executed code segments inability to determine data alignment and other code generation factors For this reason it is recommended that you check carefully whether particular program units or loops show improved performance when compiled with this option enabled 2 5 1 Vectorization Sub options The vectorizer performs high level loop transformations on countable loops A loop is countable if the number of iterations is set only before loop execution and cannot be modified during loop execution Some of the ve
201. h elimination Constant propagation Copy propagation e Dead store elimination Global register allocation Optimization amp Parallelization 17 e Invariant code motion e Induction variable elimination You select the optimization level on the command line For example level two optimization results in global optimization as shown below pgf95 O02 prog f Specifying O on the command line without a level designation is equivalent to O2 The default optimization level changes depending on which options you select on the command line For example when you select the g debugging option the default optimization level is set to level zero O0 However you can override this default by placing Olevel option after g on the command line if you need to debug optimized code Refer to Section 2 8 Default Optimization Levels for a description of the default levels As noted above the fast option includes O2 on all x86 and x64 targets If you wish to override this with O3 while maintaining all other elements of fast simply compile as follows pgf95 fast 03 prog f Note Most modern x86 processors support SSE instructions so using the fastsse option instead of fast is recommended for those processors Only use fast for older x86 processors as described in the following section 2 3 1 Scalar SSE Code Generation For all processors prior to Intel Pentium 4 and AMD Opteron Athlon64 for exam
202. h the number of characters present 124 Relational expression mixes character with numeric data A relational expression is used to compare two arithmetic expressions or two character expressions A character expression cannot be compared to an arithmetic expression I125 Dummy procedure not declared EXTERNAL Messages 261 A dummy argument which is not declared in an EXTERNAL statement is used as the subprogram name in a CALL statement or is called as a function and is therefore assumed to be a dummy procedure This message can result from a failure to declare a dummy array I126 Name is not an intrinsic function I127 Optimization level for changed to opt 1 W128 Integer constant truncated to fit data type An integer constant will be truncated when assigned to data types smaller than 32 bits such as a BYTE I129 Floating point overflow Check constants and constant expressions I130 Floating point underflow Check constants and constant expressions I131 Integer overflow Check floating point expressions cast to integer I132 Floating pt invalid oprnd Check constants and constant expressions I133 Divide by 0 0 Check constants and constant expressions 134 Illegal attribute W135 Missing STRUCTURE name field A STRUCTURE name field is required on the outermost structure 262 Messages W136 Field namelist not allowed The field namelist field of the STRUCTURE statement is disallowed on the outermost st
203. hase Use the P option to halt the compilation process after preprocessing and write the preprocessed output to the file filename i where the input file is filename c or filename cc Use the suffix option with this option to save the intermediate file in a file with the specified suffix Default The compiler produces an executable file Usage In the following example the compiler produces the preprocessed file myprog i in the current directory pgCC P myprog cc Cross references C c E Mkeepasm o S pch pgCC only Automatically use and or create a precompiled header file If use_pch or create_pch manual PCH mode appears on the command line following this option its effect is erased 98 Command Line Options pch_dir directoryname pgCC only The directory in which to search for and or create a precompiled header file This option may be used with automatic PCH mode pch or manual PCH mode create_pch or use_pch no_ pch_messages pgCC only Enable or disable the display of a message indicating that a precompiled header file was created or used in the current compilation preinclude lt filename gt pgCC only Specifies the name of a file to be included at the beginning of the compilation This option can be used to set system dependent macros and types for example use_pch filename pgCC only Use a precompiled header file of the specified name as part of the current c
204. hat can contain different data types The members are allocated in the order they appear in the definition but may not occupy contiguous locations When a struct is defined with member functions its alignment issues are the same as those for a class union is a single location that can contain any of a specified set of scalar or aggregate data types A union can have only one value at a time The data type of the union member to which data is assigned determines the data type of the union after that assignment 182 Fortran C and C Data Types 9 2 3 Class and Object Data Layout Class and structure objects with no virtual entities and with no base classes that is just direct data field members are laid out in the same manner as C structures The following section describes the alignment and size of these C like structures C classes and structures as a special case of a class are more difficult to describe Their alignment and size is determined by compiler generated fields in addition to user specified fields The following paragraphs describe how storage is laid out for more general classes The user is warned that the alignment and size of a class or structure is dependent on the existence and placement of direct and virtual base classes and of virtual function information The information that follows is for informational purposes only reflects the current implementation and is subject to change Do not make assumptions about the layout of
205. havior The standard calling sequence does not include any method to detect such failures nor to detect return value type mismatches Therefore the user must declare all functions properly There is no difference in the representation of single double or extended precision values in floating point registers 212 Run time Environment e Functions that return no value also called procedures or void functions put no particular value in any register e A call instruction pushes the address of the next instruction the return address onto the stack The return instruction pops the address off the stack and effectively continues execution at the next instruction after the call instruction A function that returns a scalar or no value must preserve the caller s registers as described above Additionally the called function must remove the return address from the stack leaving the stack pointer esp with the value it had before the call instruction was executed Functions Returning Structures or Unions If a function returns a structure or union then the caller provides space for the return value and places its address on the stack as argument word zero In effect this address becomes a hidden first argument A function that returns a structure or union also sets eax to the value of the original address of the caller s area before it returns Thus when the caller receives control again the address of the returned object resides in regist
206. he F option to halt the compilation process after preprocessing and write the preprocessed output to the file filename f where the input file is filename P Default The compiler produces an executable file Usage In the following example the compiler produces the preprocessed file myprog f in the current directory pgf95 F myprog F Cross reference c E Mkeepasm o S fast A generally optimal set of options is chosen depending on the target system In addition the appropriate tp option is automatically included to enable generation of code optimized for the type of system on which compilation is performed 48 Command Line Options Note Auto selection of the appropriate tp option means that programs built using the fast option on a given system are not necessarily backward compatible with older systems Cross reference O Munroll Mnoframe Mvect mp Mscalarsse fastsse A generally optimal set of options is chosen for targets that support SSE capability In addition the appropriate tp option is automatically included to enable generation of code optimized for the type of system on which compilation is performed Note Auto selection of the appropriate tp option means that programs built using the fastsse option on a given system are not necessarily backward compatible with older systems Cross reference O Munroll Mnoframe Mscalarsse Mvect Mcache_align
207. he executables which they generate Many of these environment variables are documented in context in other sections of the PGI User s Guide They are gathered here for easy reference Specifically excluded are environment variables specific to OpenMP which are used to control the behavior of OpenMP programs See section 5 17 Environment Variables for a list and description of environment variables that affect the execution of Fortran OpenMP programs See section 6 16 Environment Variables for a list and description of environment variables that affect the execution of C and C OpenMP programs Also excluded are environment variables that control the behavior of the PGDBG debugger or PGPROF profiler See the PGI Tools Guide for a description of environment variables that affect these tools FORTRAN_OPT If this variable exists and contains the value vaxio the record length in the open statement is in units of 4 byte words and the edit descriptor only has an effect for lines beginning with a space or If this variable exists and contains the value format_relaxed an I O item corresponding to a numerical edit descriptor F E I etc is not required to be a type implied by the descriptor For example setenv FORTRAN OPT vaxio will cause the PGI Fortran compilers to use VAX I O conventions as defined above MPSTKZ increase the size of the stacks used by threads executing in parallel regions It is for use with programs that utilize larg
208. he compiler option Mnoframe is specified The stack must be kept aligned on 16 byte boundaries Variable Length Parameter Lists Parameter passing in registers can handle a variable number of parameters The C language uses a special method to access variable count parameters The stdarg h and varargs h files define several functions to access these parameters A C routine with variable parameters must use the va_start macro to set up a data structure before the parameters can be used The va_arg macro must be used to access the successive parameters For calls that use varargs or stdargs the register rax acts as a hidden argument whose value is the number of XMM registers used in the call 224 Run time Environment C Parameter Conversion In C for a called prototyped function the parameter type in the called function must match the argument type in the calling function If the called function is not prototyped the calling convention uses the types of the arguments but promotes char or short to int and unsigned char or unsigned short to unsigned int and promotes float to double unless you use the Msingle option For more information on the Msingle option refer to Chapter 3 Calling Assembly Language Programs Example A 3 C Program Calling an Assembly language Routine File testmain c main long JL paral 0x3 800000 float Cf para2 1 0 double d_para3 0 5 float return extern float sum 3
209. he extract directory required in order to save and re use inline libraries If you specify both name and size the compiler extracts functions that match func or that have number or fewer statements For examples of extracting functions see Function Inlining Minline option option This passes options to the function inliner where option can be any of Command Line Options except func instructs the inliner to inline all eligible functions except func a function in the source text Multiple functions can be listed comma separated name func instructs the inliner to inline the function func The func name should be a non numeric string that does not contain a period You can also use a name prefix followed by the function name If name is specified what follows is always the name of a function 65 lib filename ext instructs the inliner to inline the functions within the library file filename ext The compiler assumes that a filename ext option containing a period is a library file Create the library file using the Mextract option You can also use a lib prefix followed by the library name If lib is specified no period is necessary in the library name Functions from the specified library are inlined If no library is specified functions are extracted from a temporary library created during an extract prepass size number instructs the inliner to inline functions with number or fewer statements Yo
210. he following command setenv NCPUS lt number gt in a Windows command prompt window setenv NCPUS lt number gt in a shell command window under csh or with Getting Started 7 NCPUS lt number gt export NCPUS in sh ksh or BASH command window Note If you set NCPUS to a number larger than the number of physical processors your program may execute very slowly 1 4 2 Running Data Parallel HPF Programs When you execute an HPF program by default it will use only one processor If you wish to run on more than one processor use the pghpf np runtime option For example to compile and run the hello f example defined above on one processor you would issue the following commands pghpf o hello hello f Linking hello hello a 5 To execute it on two processors you would issue the following commands hello pghpf np 2 hello a 5 Note If you specify a number larger than the number of physical processors your program will execute very slowly Note that you still only see a single hello printed to your screen This is because HPF is a single threaded model meaning that all statements execute with the same semantics as if they were running in serial However parallel statements or constructs operating on explicitly distributed data are in fact executed in parallel The programmer must manually insert compiler directives to cause data to be distributed to the available processors See the PG
211. he object file e Propagation Processing all the object files to propagate the interprocedural summary information across function and file boundaries This is done during the link step when all the object files are combined if the Mipa switch is present on the link command line Recompile Optimization Each of the object files is recompiled with the propagated interprocedural information producing a specialized object file This is also done during the link step when the Mipa switch is present on the link command line When linking with Mipa the PGI compilers automatically regenerate IPA optimized versions of each object file essentially recompiling each file If there are IPA optimized objects from a previous build the compilers will minimize the recompile time by reusing those objects if they are still valid They will still be valid if the IPA optimized object is newer than the original object file and the propagated IPA information for that file has not changed since it was optimized After each object file has been recompiled the regular linker is invoked to build the application with the IPA optimized object files The IPA optimized object files are saved in the same directory as the original object files for use in subsequent program builds 2 7 5 Building a Program with IPA Single Step By adding the Mipa command line switch several source files can be compiled and linked with interprocedural optimizations with
212. hunk size is chosen equal to number_of iterations omp_num threads 1 omp_num_threads e When SCHEDULE DYNAMIC chunk is specified iterations are allocated in contiguous blocks of size chunk As each thread finishes a piece of the iteration space it dynamically obtains the next set of iterations The chunk must be a scalar integer expression If no chunk is specified a default chunk size is chosen equal to 1 When SCHEDULE GUIDED chunk is specified the chunk size is reduced in an exponentially decreasing manner with each dispatched piece of the iteration space Chunk specifies the minimum number of iterations to dispatch each time except when there are less than chunk iterations remaining to be processed at which point all remaining iterations are assigned If no chunk is specified a default chunk size is chosen equal to 1 When SCHEDULE RUNTIME is specified the decision regarding iteration scheduling is deferred until runtime The schedule type and chunk size can be chosen at runtime by setting the OMP_SCHEDULE environment variable If this environment variable is not set the resulting schedule is equivalent to SCHEDULE STATIC OpenMP Directives for Fortran 115 5 7 WORKSHARE END WORKSHARE The OpenMP WORKSHARE END WORKSHARE directive pair uses the following syntax Syntax SOMP WORKSHARE lt Fortran structured block to be executed in parallel gt SOMP END WORKSHARE NOWAIT The Fortran stru
213. i must be used at both compile time and link time Programs compiled with Mpfi include extra code to collect run time statistics and write them out to a trace file When the resulting program is executed a profile feedback trace file pgfi out is generated in the current working directory Note Programs compiled and linked with Mpfi will execute more slowly due to the instrumentation and data collection overhead You should use executables compiled with Mpfi only for execution of training runs When invoked with the Mpfo option the PGI compilers use data from a pgfi out profile feedback tracefile to enable or enhance certain performance optimizations Use of this option requires the presence of a pgfi out trace file in the current working directory 2 9 Default Optimization Levels The following table shows the interaction between the O g and M lt opt gt options In the table level can be 0 1 2 or 3 and lt opt gt can be vect unroll or ipa The default optimization level is dependent upon these command line options 36 Optimization amp Parallelization Table 2 1 Optimization and O g and M lt opt gt Options Optimize Debug M lt opt gt Optimization Option Option Option Level none none none 1 none none M lt opt gt 2 none g none 0 O none or g none 2 Olevel none or g none level Olevel lt 2 none or g M lt opt gt 2 03 none or g none 3 U
214. ication binary interface ABI as defined in the System V Application Binary Interface Intel Processor Supplement and the System V Application Binary Interface listed in the Related Publications section in the Preface A 1 1 Function Calling Sequence This section describes the standard function calling sequence including the stack frame register usage and parameter passing A 1 1 1 Register Usage Conventions The following table defines the standard for register allocation The 32 bit x86 Architecture provides a number of registers All the integer registers and all the floating point registers are global to all procedures in a running program 209 Table A 1 Register Allocation Type Name Purpose General Joeax integer return value Joedx dividend register for divide operations pecx count register shift and string operations Joebx local register variable ebp optional stack frame pointer oesi local register variable Jedi local register variable hesp stack pointer Floating point st 0 floating point stack top return value st 1 floating point next to stack top st st 7 floating point stack bottom In addition to the registers each function has a frame on the run time stack This stack grows downward from high addresses The next table shows the stack frame organization 210 Run time Environment Table A 2 Standard Stack Frame P
215. id _ declspec dllimport func_4a void int main func _3a func dai 0 return 1 D Step 1 To make obj3 dll and obj4 dll first compile the source and create an import library for each DLL that will be built The PGI drivers call the Microsoft lib tool to create import libraries The lib tool will only create an import library if a module definition def file is provided A def file contains symbols to export In this example the symbols to be exported are already marked as such by __declspec dllexport so the det file should be empty pgcce c object3 c pgcc Mmakeimplib o obj3 lib object3 obj def obj3 def d i d i pgce c object4 c pgcc Mmakeimplib o obj4 lib object4 obj def obj4 def d i d i Step 2 Create the DLLs using the import libraries obj3 lib and obj4 lib o pgcc Mmakedll o obj3 dll object3 obj defaultlib obj4 o pgcc Mmakedll o obj4 dll object4 obj defaultlib obj3 Step 3 Compile the main program and link against the import libraries for obj3 dll and obj4 dll o pgcc Mdll prog3 c o prog3 defaultlib obj3 defaultlib obj4 Libraries and Environment Variables 169 Step 4 Execute prog3 to ensure that the DLLs were created properly prog3 func_3a calling a routine in obj4 dll func Ab func Aa calling a routine in obj3 dll func 3b 8 5 Using LIB3F The PGI Fortran compilers include complete support for the de facto standard LIB3F library routines on bo
216. ies that are utilized by the PGI compilers and tools when running on a 64 bit operating system The prefetch SSE1 SSE2 and SSE3 processor features further distinguish the various processors Where such distinctions are important with respect to a given compiler option or feature it is explicitly noted in this manual Note that the default for performing scalar floating point arithmetic is to use SSE instructions on targets that support SSE1 and SSE2 See section 2 3 1 Scalar SSE Code Generation for a detailed discussion of this topic Xvi Preface Related Publications The following documents contain additional information related to the x86 and x64 architectures and the compilers and tools available from The Portland Group PGI Fortran Reference manual describes the FORTRAN 77 Fortran 90 95 and HPF statements data types input output format specifiers and additional reference material related to use of the PGI Fortran compilers System V Application Binary Interface Processor Supplement by AT amp T UNIX System Laboratories Inc Prentice Hall Inc System V Application Binary Interface X86 64 Architecture Processor Supplement http www x86 64 org abi pdf Fortran 95 Handbook Complete ISO ANSI Reference Adams et al The MIT Press Cambridge Mass 1997 Programming in VAX Fortran Version 4 0 Digital Equipment Corporation September 1984 IBM VS Fortran IBM Corporation Rev GC26 4119 The C Programming Language b
217. ilation semantics with regard to conformance to TEEE 754 L lt dirname gt Specifies a library directory I lt libname gt Loads a library M lt pegflag gt Selects variations for code generation and optimization m Displays a link map on the standard output mcmodel medium tp k8 64 and tp p7 64 targets only Generate code which supports the medium memory model in the linux86 64 environment module lt moduledir gt F90 F95 HPF only Save search for module files in directory lt mod uledir gt mp align no numa Interpret and process user inserted shared memory parallel program ming directives see Chapters 5 and 6 O lt level gt Specifies code optimization level where lt level gt is 0 1 2 or 3 0 Names the object file pc lt val gt tp px p5 p6 piii targets only Set precision globablly for x87 float ing point calculations must be used when compiling the main pro gram lt val gt may be one of 32 64 or 80 pg Instrument the generated executable to produce a gprof style gmon out sample based profiling trace file qp is also supported and is equivalent pgf77libs Append PGF77 runtime libraries to the link line pgf90libs Append PGF90 PGF95 runtime libraries to the link line Q Selects variations for compiler steps R lt directory gt Linux only Passed to the Linker Hard code lt directory gt into the search path for shared object files Tr
218. iler to produce all available information on why various optimizations are not performed instructs the compiler to produce all available information on why loops are not automatically parallelized In particular if a loop is not parallelized due to potential data dependence the variable s that cause the potential dependence will be listed in the Mneginfo messages instructs the compiler to produce information on why memory hierarchy optimizations on loops are not performed instructs the compiler to display error messages at the specified and higher levels where level is one of the following inform fatal severe warn inform instructs the compiler to display fatal error messages instructs the compiler to display severe and fatal error messages instructs the compiler to display warning severe and fatal error messages informinstructs the compiler to display all error messages inform warn severe and fatal instructs the compiler to keep the assembly file as compilation continues Normally the assembler deletes this file when it is finished The assembly file has the same filename as the source file but with a s extension 81 Mlist instructs the compiler to create a listing file The listing file is filename st where the name of the source file is filename f Mnolist the compiler does not create a listing file This is the default Mmaked11 Windows only generate a dynamic link library DLL
219. in the PARALLEL SECTIONS END SECTIONS directives must be a structured block and the code in each SECTION must be a structured block The available clauses are as defined in Section 5 2 PARALLEL END PARALLEL and Section 5 6 DO END DO 5 14 ORDERED The OpenMP ORDERED directive is supported using the following syntax SOMP ORDERED lt Fortran code block executed by processor gt SOMP END ORDERED The ORDERED directive can appear only in the dynamic extent of a DO or PARALLEL DO directive that includes the ORDERED clause The code block between the ORDERED END ORDERED directives is executed by only one thread at a time and in the order of the loop iterations This sequentializes the ordered code block while allowing parallel execution of statements outside the code block The following additional restrictions apply to the ORDERED directive The ORDERED code block must be a structured block It is illegal to branch into or out of the block A given iteration of a loop with a DO directive cannot execute the same ORDERED directive more than once and cannot execute more than one ORDERED directive 5 15 ATOMIC The OpenMP ATOMIC directive uses following syntax SOMP ATOMIC The ATOMIC directive is semantically equivalent to enclosing the following single statement in a CRITICAL END CRITICAL directive pair The statement must be of one of the following forms x X operator expr 120 OpenMP Directives
220. ine option While a command line option affects the entire source file that is being compiled pragmas apply the effects of a particular command line option to selected functions or to selected loops in the source file Pragmas may also toggle an option selectively enabling and disabling the option Pragmas let you tune selected functions or loops based on your knowledge of the code The general syntax of a pragma is pragma scope pragma body The optional scope field is an indicator for the scope of the pragma some pragmas ignore the scope indicator The valid scopes are global indicates the pragma applies to the entire source file routine indicates the pragma applies to the next function loop indicates the pragma applies to the next loop but not to any loop contained within the loop body Loop scoped pragmas are only applied to for and while loops If a scope indicator is not present the default scope if any is applied Whitespace must appear after the pragma keyword and between the scope indicator and the body of the pragma Whitespace may also surround any special characters such as a comma or an equal sign Case is significant for the names of the pragmas and any variable names that appear in the body of the pragma Optimization Directives and Pragmas 151 7 6 C C Pragma Summary The following table summarizes the supported pragmas The scope entry in the table indicates the permitted scope indicators for each pragm
221. inform option The compiler messages refer to a severity level a message number and the line number where the error occurs The compiler can also display internal error messages on standard errors If your compilation produces any internal errors contact you re The Portland Group s technical reporting service by sending e mail to trs pgroup com If you use the listing file option Mlist the compiler places diagnostic messages after the source lines in the listing file in the following format PGFTN etype enum message filename line Where etype is a character signifying the severity level enum is the error number message is the error message filename is the source filename line is the line number where the compiler detected an error 247 B 2 Phase Invocation Messages You can display compiler assembler and linker phase invocations by using the v command line option For further information about this option see Chapter 3 Command Line Options BA Fortran Compiler Error Messages This section presents the error messages generated by the PGF77 and PGF95 compilers The compilers display error messages in the program listing and on standard output and can also display internal error messages on standard error B 3 1 Message Format Each message is numbered Each message also lists the line and column number where the error occurs A dollar sign in a message represents information that is specifi
222. ing an object of one data type to an object of a larger signed data type instructs the compiler not to extend the sign bit that is set as the result of converting an object of one data type to an object of a larger data type In the case where a scalar is used after a loop but is not defined on every iteration of the loop the compiler does not by default parallelize the loop However this option tells the compiler it s safe to parallelize the loop For a given loop the last value computed for all scalars make it safe to parallelize the loop 63 instructs the compiler to inhibit certain optimizations and to allow for stride 0 array references This option may degrade performance and should only be used if zero stride induction variables are possible Mnostrideo instructs the compiler to perform certain optimizations and to disallow for stride 0 array references Munix use UNIX symbol and parameter passing conventions for Fortran subprograms pgf77 pgf95 and pghpf for Win32 only Mvarargs force Fortran program units to assume procedure calls are to C functions with a varargs type interface pgf77 and pgf95 only Default For arguments that you do not specify the default code generation controls are as follows nodaz noflushz norecursive nostrideO noreentrant noref_externals nosignextend nosecond_underscore M lt pgflag gt Environment Controls Syntax Mlfs 32 bit Linux only link in libraries that enable file I O
223. ink the resulting object files into an executable pgcc Mipa fast c filel c pgcce Mipa fast c file2 c pgcc Mipa fast c file3 c pgcc Mipa fast o a out filel o file2 o file3 o d i d i ol d i The pgcc driver invokes the compiler and assembler as required to process each source file and invokes the IPA linker for the final link command If you modify one of the source files the executable can be rebuilt by compiling just that file and then relinking pgce c filel c pgcc o a out filel o file2 o file3 o d i d i When the IPA linker is invoked it will determine that the IPA optimized object for Dlel o file 1_ipa5_a out o is stale since it is older than the object Dlel o and hence will need to be rebuilt and will reinvoke the compiler to generate it In addition depending on the nature of the changes to the source file file1 c the interprocedural optimizations previously performed for file2 and file3 may now be inaccurate For instance IPA may have propagated a constant argument value in a call from a function in file1 c to a function in file2 c if the value of the argument has changed any optimizations based on that constant value are invalid The IPA linker will determine which if any of any Optimization amp Parallelization 33 previously created IPA optimized objects need to be regenerated and will reinvoke the compiler as appropriate to regenerate them Only those objects that are stale or which have new or
224. instruction scheduling A loop with a constant count may be completely unrolled or partially unrolled A loop with a non constant count may also be unrolled A candidate loop must be an innermost loop containing one to four blocks of code The following shows the use of the Munroll option pgf95 Munroll prog f The Munroll option is included as part of fast and fastsse on all x86 and x64 targets The loop unroller expands the contents of a loop and reduces the number of times a loop is executed Branching overhead is reduced when a loop is unrolled two or more times since each iteration of the unrolled loop corresponds to two or more iterations of the original loop the number of branch instructions executed is proportionately reduced When a loop is unrolled completely the loop s branch overhead is eliminated altogether Loop unrolling may be beneficial for the instruction scheduler When a loop is completely unrolled or unrolled two or more times opportunities for improved scheduling may be presented The code generator can take advantage of more possibilities for instruction grouping or filling instruction delays found within the loop Examples 2 1 and 2 2 show the effect of code unrolling on a segment that computes a dot product Optimization amp Parallelization 19 Example 2 1 Dot Product Code REAL 4 A 100 B 100 Z INTEGER I DO I 1 100 Z Z A i B i END DO END Example 2 2 Unrolled Dot Product Code R
225. interprocedural analysis I398 IPA distribution target propagated to The distribution target for the dummy argument was changed as per interprocedural analysis I399 IPA common block not optimized Messages 283 The given common block was not optimized by interprocedural analysis either because it was not declared in the main program or because it was declared differently in different subprograms E400 IPA dummy argument is an asterisk but actual argument is not a label The subprogram expects an alternate return label for this argument E401 Actual argument is a subprogram but Dummy argument is not declared EXTERNAL The call statement passes a function or subroutine name to a dummy argument that is a scalar variable or array E402 Actual argument is illegal E403 Actual argument and formal argument have different ranks The actual and formal array arguments differ in rank which is allowed only if both arrays are declared with the HPF SEQUENCE attribute E404 Sequential array section of in argument is not contiguous When passing an array section to a formal argument that has the HPF SEQUENCE attribute the actual argument must be a whole array with the HPF SEQUENCE attribute or an array section of such an array where the section is a contiguous sequence of elements E405 Array expression argument may not be passed to sequential dummy argument When the dummy argument has the HPF SEQUENCE attribu
226. ions cncall Calls in parallel loops are safe to parallelize Loops containing calls are candidates for parallelization Also no minimum loop count threshold must be satisfied before parallelization will occur and last values of scalars are assumed to be safe dist block Parallelize with block distribution this is the default Contiguous blocks of iterations of a parallelizable loop are assigned to the available processors dist cyclic Parallelize with cyclic distribution The outermost parallelizable loop in any loop nest is parallelized If a parallelized loop is innermost its iterations are allocated to processors cyclically For example if there are 3 processors executing a loop processor 0 performs iterations 0 3 6 etc processor 1 performs iterations 1 4 7 etc and processor 2 performs iterations 2 5 8 etc no innermost Enable parallelization of innermost loops The default is to not parallelize innermost loops since it is usually not profitable on dual core processors noassoc Disables parallelization of loops with reductions When linking the Mconcur switch must be specified or unresolved references will result The NCPUS environment variable controls how many processors or cores are used to execute parallelized loops This option appli Note es only on shared memory multi processor SMP or dual core processor based systems Mcray option option Mdepchk Mnodepchk Command Li
227. ironment 235 Functions Returning Structures or Unions A function can use either registers or the stack to return a structure or union The size and type of the structure or union determine how it is returned A structure or union is returned in memory if it is larger than 8 bytes or if its size is 3 5 6 or 7 bytes A structure or union is returned in rax if its size is 1 2 4 or 8 bytes If a structure or union is to be returned in memory the caller provides space for the return value and passes its address to the function as a hidden first argument in rcx This address will also be returned in rax A 3 3 Argument Passing Integral and Pointer Arguments Integral and pointer arguments are passed to a function using the next available register of the sequence rcx Yordx r8 r9 After this list of registers has been exhausted all remaining integral and pointer arguments are passed to the function via the stack Floating Point Arguments Float and double arguments are passed to a function using the next available XMM register of the sequence xmm0 xmm1 xmm2 xmm3 After this list of registers has been exhausted all remaining XMM floating point arguments are passed to the function via the stack Array Structure and Union Arguments Arrays and strings are passed to functions using a pointer to caller allocated memory Structure and union arguments of size 1 2 4 or 8 bytes will be passed as if they were integers of
228. is not considered to define a scope The dependent statement may not be a declaration Any objects constructed within the dependent statement are destroyed at exit from the dependent statement Implicit conversion from integral types to enumeration types is allowed A non const member function may be called for a const object A warning is issued A const void value may be implicitly converted to a void value e g when passed as an argument When in determining the level of argument match for overloading a reference parameter is initialized from an argument that requires a non class standard conversion the conversion counts as a user defined conversion This is an outright bug which unfortunately happens to be exploited in the NIH class libraries When a builtin operator is considered alongside overloaded operators in overload resolution the match of an operand of a builtin type against the builtin type required by the builtin operator is considered a standard conversion in all cases e g even when the type is exactly right without conversion A reference to a non const type may be initialized from a value that is a const qualified version of the same type but only if the value is the result of selecting a member from a const class object or a pointer to such an object A cast to an array type is allowed it is treated like a cast to a pointer to the array element type A warning is issued C Dialect Supported 299 e When a
229. it clause is specified Nested single process sections are ignored Branching into or out of a single process section is not supported The private and firstprivate clauses are as described in Section 6 2 omp parallel The copyprivate clause causes the variables in list to be copied from the private copies in the single thread that executes the single region to the other copies in all other threads of the team at the end of the single region The copyprivate clause must not be used with nowait 6 6 omp for The OpenMP omp for pragma uses the following syntax pragma omp for Clauses lt C C for loop to be executed in parallel gt 130 OpenMP Pragmas for C and C Clauses private list firstprivate list lastprivate list reduction operator list schedule kind chunk ordered nowait The real purpose of supporting parallel execution is the distribution of work across the available threads You can explicitly manage work distribution with constructs such as if omp_get_thread_num 0 else if omp_get_thread_num 1 However these constructs are not in the form of pragmas The omp for pragma provides a convenient mechanism for the distribution of loop iterations across the available threads in a parallel region The following variables can be used e Variables declared in a private list are treated as private to each processor participating in parallel execution of the loop meaning that
230. ium Pro Pentium II or first generation AMD Athlon processors They will only execute correctly on Pentium III Pentium 4 Xeon EM64T AthlonXP Athlon64 and Opteron systems running an SSE enabled operating system 2 5 1 4 Prefetch Option The option Mvect prefetch instructs the vectorizer to automatically generate prefetch instructions when vectorizable loops are encountered even in cases where SSE or SSE2 instructions are not generated Usually explicit prefetching is not necessary on Pentium 4 Xeon and Opteron because these processors support hardware prefetching nonetheless it sometimes can be worthwhile to experiment with explicit prefetching Prefetching can be controlled on a loop by loop level using prefetch directives which are described in detail in Section 7 8 Prefetch Directives Note Program units compiled with Mvect prefetch will not execute correctly on Pentium Pentium Pro or Pentium II processors They will execute correctly only on Pentium III Pentium 4 Xeon EMO4T AthlonXP Athlon64 or Opteron systems In addition the prefetchw instruction is only supported on AthlonXP Athlon64 or 22 Optimization amp Parallelization Opteron systems and can cause instruction faults on non AMD processors For this reason the PGI compilers do not generate prefetchw instructions by default on any target In addition to these sub options to Mvect several other sub options are supported See the description of
231. ives exceeded the maximum currently 20 S262 Unterminated macro definition for A newline was encountered in the formal parameter list for the indicated macro 263 Unterminated string or character constant A newline with no preceding backslash was found in a quoted string I264 Possible nested comment The characters were found within a comment S265 lt reserved message number gt S266 lt reserved message number gt Messages 273 267 W268 W269 lt reserved message number gt Cannot inline subprogram common block mismatch Cannot inline subprogram argument type mismatch This message may be Severe if have gone too far to undo inlining process F270 W271 1272 273 F274 275 F276 I277 278 W279 274 Missing exlib option Can t inline wrong number of arguments Argument of inlined function not used Inline library not specified on command line inlib switch Unable to access file TOC Unable to open file while extracting or inlining Assignment to constant actual parameter in inlined subprogram Inlining of function may result in recursion lt reserved message number gt Possible use of before definition in Messages The optimizer has detected the possibility that a variable is used before it has been assigned a value The names of the variable and the function in which the use occurred are listed The line number if specified is the line number
232. k L RG fcon nofcon Do don t assume unsuffixed real constants nofcon R G are single precision invarif noinvarif Do don t remove invariant if constructs from invarif L RG loops Istval nolstval Do don t compute last values Istval L RG opt Select optimization level N A R G safe nosafe Do don t treat pointer arguments as safe safe R G safe_lastval Parallelize when loop contains a scalar used not enabled L outside of loop safeptr nosafeptr Do don t ignore potential data dependencies nosafeptr L R G to pointers single nosingle Do don t convert float parameters to double nosingle R G unroll nounroll Do don t unroll loops nounroll L RG vector novector Do don t perform vectorizations vector L RG vintr novintr Do don t recognize vector intrinsics vintr L RG fcon nofcon This pragma alters the effects of the Mfcon command line option a M Language control The pragma instructs the compiler to treat non suffixed floating point constants as float rather than double By default all non suffixed floating point constants are treated as double Optimization Directives and Pragmas safe nosafe By default the compiler assumes that all pointer arguments are unsafe That is the storage located by the pointer can be accessed by other pointers The forms of the safe pragma are pragma scope no safe pragma safe variable variable where scope is either global or routi
233. ke at command Jevel 1 GG aes oils cides eM tense XV Eegeregie ee XV PGE 9D reine area t XV PGHP EE XV CPP EE 4 D Data Ke 175 eline E E TET 184 C class and object layout 183 C ClASSOS occ hieis iieii 182 C C aggregate alignment s nssnss111 183 C C scalar data ines 180 C C StrUC a raae eeraa eaa RE ahe 182 C CE VOI o iea ai nsa niee asaidie 185 DEC structures 0 1ccsieieen lade 178 DEC Untons ae aaaeei earar rate nirien 178 F90 derived types seeren 180 Fortranecn isteach ne e 175 internal padding WEE tail padding 184 Directives FOrtran sete tess ehh Bi ee tees 3 OPTIMIZATION ni enasna 141 Parallelization ccccccccceeeeeeeeeeeeeeeeeeees 107 prefet h isar Srine snini tira aa 157 SCOPO A ATE nn ie 148 E Environment variables 0 cceeeeeeees 171 ME BIND tee r a erra aint 171 Index MP SPIN sis eias cae tdi te teed and 171 MP EE 172 MPSTKZ sieictciesces AEAT 9 NOPU EE 172 NCPUS_ MAX cceeccccceeeeeeeeteneeteeeeeetaes 172 NO GTOp MEGGAOE 172 tal eifreg ea einen oad se 172 PGI CONTINUE 172 173 STATIC_RANDOM_SEED eeeee 173 IMP DIR WE 173 KEEN 173 F Filename Conventions ccsceeceeeseseeeeeseeenees 3 OXTOMSIONS au bade neike ateena kine E apinain ak 3 Input FileStream aaa a aaae aagi 3 O tp t EE 5 Floating point stack AA 86 Fortran directive summanm 142 named common Dlocke AAA 191 Fortran Parallelization Directives ATOMIG eet aa eia eni i a aa i
234. l booll TRUE letterl v numintl 11 numint2 44 numfloatl 39 6 numdoubl 39 2 numshorl 981 Compile and execute the program fmain f with the call to cfunc_ using the following command lines pgcc c cfunc c pgf95 cfunc o fmain f Executing the a out file should produce the following output Tov ILI 44 39 6 39 2 981 10 9 Example C Calling Fortran Example 10 6 C Main Program cmain c shows a C main program that calls the Fortran subroutine shown in Example 10 5 Fortran Subroutine forts f Notice that each call uses the E operator to pass by reference Also notice that the call to the Fortran subroutine uses all lower case and a trailing wow 194 Inter language Calling Example 10 5 Fortran Subroutine forts f subroutine forts booll letterl numinti amp numint2 numfloati numdoub1 numshor1 logical 1 booll character letterl integer numintl numint2 double precision numdoub1 real numfloatl integer 2 numshorl booll true letterl v numintl 11 numint2 44 numdoubl1 902 numfloatl 39 6 numshorl 299 return end Example 10 6 C Main Program cmain c main char booll letterl1 int numintl numint2 float numfloatil double numdoub1 short numshor1 extern void forts forts amp booll amp letterl amp numint1 amp numint2 amp numfloatli amp numdoubl amp numshorl 1 printf Sa Sc d Sd 3 1f 0 d n bool1
235. l data items in each DLL are resolved to separate memory locations In short global data in a DLL cannot be directly accessed from outside the DLL The PGI runtime DLLs can be used to create both executables and other DLLs The following switches apply Mdll Link with the DLL version of the runtime libraries This flag is required when linking with any DLL built by the PGI compilers Mmakedll Generate a dynamic link library or DLL Mnopgdllmain Do not link the module containing the default DllMain into the DLL This flag applies to building DLLs with the PGF95 and PGHPF compilers If you want to replace the default DlIMain routine with a custom DIIMain Q use this flag and add the object containing the custom DllMain to the link line The latest version of the default DlIMain is included in the Release Notes the code in this routine specific to PGF95 and PGHPF must be incorporated into the custom version of DllMain to ensure the appropriate function of your DLL o lt file gt Passed to the linker Name the DLL lt file gt output def lt file gt Passed to linker Generate a def named lt file gt for the DLL The def file contains the symbols exported by the DLL Generating a def file is not required when building a DLL but can be a useful debugging tool if the DLL Libraries and Environment Variables 161 out implib lt file gt export all symbols no default excludes does not contain the s
236. l for programs written for earlier versions of Fortran the compiler assumes that all local variables are subject to the SAVE statement Note that this may allow older Fortran programs to run but it can greatly reduce performance the compiler does not assume that all local variables are subject to the SAVE statement 67 Mstandard Munixlogical Mnounixlogical Mupcase Mnoupcase the compiler flags non ANSI conforming source code directs the compiler to treat logical values as true if the value is non zero and false if the value is zero UNIX F77 convention When Munixlogical is enabled a logical value or test that is non zero is TRUE and a value or test that is zero is FALSE In addition the value of a logical expression is guaranteed to be one 1 when the result is TRUE Idirects the compiler to use the VMS convention for logical values for true and false Even values are true and odd values are false the compiler allows uppercase letters in identifiers With Mupcase the identifiers X and x are different and keywords must be in lower case This selection affects the linking process if you compile and link the same source code using Mupcase on one occasion and Mnoupcase on another you may get two different executables depending on whether the source contains uppercase letters The standard libraries are compiled using the default Mnoupcase the compiler converts all identifiers to low
237. lag and the carry flag The direction flag must be set to the forward i e zero direction before entry and upon exit from a function Other user flags have no specified role in the standard calling sequence and are not reserved Floating Point Control Word The control word contains the floating point flags such as the rounding mode and exception masking This register is initialized at process initialization time and its value must be preserved Signals can interrupt processes Functions called during signal handling have no unusual restriction on their use of registers Moreover if a signal handling function returns the process resumes its original execution path with registers restored to their original values Thus programs and compilers may freely use all registers without danger of signal handlers changing their values A 1 2 Function Return Values Functions Returning Scalars or No Value e A function that returns an integral or pointer value places its result in register eax e A function that returns a long long integer value places its result in the registers edx and eax The most significant word is placed in edx and the least significant word is placed in eax e A floating point return value appears on the top of the floating point stack The caller must then remove the value from the floating point stack even if it does not use the value Failure of either side to meet its obligations leads to undefined program be
238. lar target processor s architecture as well as replacements that take advantage of the x86 or x64 architecture instruction set and registers For the discussion in this and the following chapters optimization is divided into the following categories Local Optimization This optimization is performed on a block by block basis within a program s basic blocks A basic block is a sequence of statements in which the flow of control enters at the beginning and leaves at the end without the possibility of branching except at the end The PGI compilers perform many types of local optimization including algebraic identity removal constant folding common sub expression elimination pipelining redundant load and store elimination scheduling strength reduction and peephole optimizations 13 Global Optimization This optimization is performed on a program unit over all its basic blocks The optimizer performs control flow and data flow analysis for an entire program unit All loops including those formed by IFs and GOTOs are detected and optimized Global optimization includes constant propagation copy propagation dead store elimination global register allocation invariant code motion and induction variable elimination Loop Optimization Unrolling Vectorization and Parallelization The performance of certain classes of loops may be improved through vectorization or unrolling options Vectorization transforms loops to improve memory a
239. le and example are adapted from System V Application Binary Interface AMD64 Architecture Processor Supplement Example A 2 Parameter Passing typedef struct int a b double d structparm structparm s int e f 9 H 134k float flt double m n extern void func int e int f structparm s int g int h float flt double m double n int i int j int k func er fy Su Gr h flt mr ny ty e OK Run time Environment 223 Table A 9 Register Allocation for Example A 2 General Purpose Registers Floating Point Registers Se Jordi e oxmm0 s d 0 j rsi f xmml1 flt 8 k rdx s a s b xmm2 m Zack g xmm3 n r8 h r9 i Implementing a Stack In general compilers and programmers must maintain a software stack The stack pointer register orsp is set by the operating system for the application when the program is started The stack must grow downwards from high addresses A separate frame pointer enables calls to routines that change the stack pointer to allocate space on the stack at run time e g alloca Some languages can also return values from a routine allocated on stack space below the original top of stack pointer Such a routine prevents the calling function from using Yorsp relative addressing for values on the stack If the compiler does not call routines that leave rsp in an altered state when they return a frame pointer is not needed and may not be used if t
240. le processing group 230 scale factor out of range 128 to 127 Fortran P edit descriptor scale factor not within range of 128 to 127 231 error on data conversion 233 too many constants to initialize group item 234 invalid edit descriptor An invalid edit descriptor has been found in a format statement 235 edit descriptor does not match item type Data types specified by I O list item and corresponding edit descriptor conflict 236 formatted record longer than 2000 characters 292 Messages 237 quad precision type unsupported 238 tab value out of range A tab value of less than one has been specified 239 entity name is not member of group 242 illegal operation on direct access file 243 format parentheses nesting depth too great 244 syntax error entity name expected 245 syntax error within group definition 246 infinite format scan for edit descriptor 248 illegal subscript or substring specification 249 error in format illegal E F G or D descriptor 250 error in format number missing after or 251 illegal character in format string 252 operation attempted after end of file Messages 293 253 attempt to read non existent record direct access 254 illegal repeat count in format 294 Messages Appendix C C Dialect Supported The PGC compiler accepts the C language as defined by The Annotated C Reference Manual ARM by Ellis and Stroustrup Addison Wesley 1990 includin
241. led using C conventions More than one subroutine may be listed separated by commas See above for a complete description of the implementation of the C calling convention NOTE The directive prefix DEC is also supported but requires a space between the prefix and the directive keyword ATTRIBUTES The must begin the prefix when compiling using Fortran 90 freeform format The characters C or can be used in place of in either form of the prefix when compiling used fixed form F 77 style format The directives are completely case insensitive 10 14 6 Using the UNIX Calling Convention Using the UNIX calling convention is straightforward Any program unit compiled using Munix compilation flag will use the UNIX convention Inter language Calling 203 204 Inter language Calling Chapter 11 C Name Mangling Name mangling transforms the names of entities so that the names include information on aspects of the entity s type and fully qualified name This is necessary since the intermediate language into which a program is translated contains fewer and simpler name spaces than there are in the C language Specifically e Overloaded function names are not allowed in the intermediate language Classes have their own scopes in C but not in the generated intermediate language For example an entity x from inside a class must not conflict with an entity x from the file scope e External names in the object code form a comple
242. les that can be used to specify shared memory parallelism in Fortran C and C programs The OpenMP C C pragmas include a parallel region construct for writing coarse grain SPMD programs work sharing constructs which specify that C C for loop iterations should be split among the available threads of execution and synchronization constructs The data environment is controlled using clauses on the pragmas or with additional pragmas Run time library functions are provided to query the parallel runtime environment for example to determine how many threads are participating in execution of a parallel region Finally environment variables are provided to control the execution behavior of parallel programs For more information on OpenMP and a complete copy of the OpenMP C C API specification see http www openmp org 6 1 Parallelization Pragmas Parallelization pragmas are pragma statements in a C or C program that are interpreted by the PGCC C and C compilers when the option mp is specified on the command line The form of a parallelization pragma is pragma omp pragmas name clauses The pragmas follow the conventions of the C and C standards Whitespace can appear before and after the Preprocessing tokens following the pragma omp are subject to macro replacement The order in which clauses appear in the parallelization pragmas is not significant Spaces separate clauses within the pragmas Clauses on pragmas may be repeated as
243. letter pass identifier between the identifier and the comma or between the comma and the option Usage In the following example the linker loads the text segment at address Oxffc00000 and the data segment at address Oxffe00000 pgf95 W1 k t 0Oxffc00000 d 0xffe00000 myprog f W Do not print warning messages 3 2 C and C specific Compiler Options The following options are specific to PGCC C and or C A pgCC only Using this option the PGC compiler accepts code conforming to the proposed ANSI C standard It issues errors for non conforming code 94 Command Line Options Default By default the compiler accepts code conforming to the standard C Annotated Reference Manual Usage The following command line requests ANSI conforming C pgCC A hello cc Cross references b and p no_lalternative_tokens pgCC only Enable or disable recognition of alternative tokens These are tokens that make it possible to write C without the use of the amp and characters The alternative tokens include the operator keywords e g and bitand etc and digraphs The default behavior is no_alternative_tokens B pgec and pgCC only Enable use of C style comments starting with in C program units Default The PGCC ANSI and K amp R C compiler does not allow C style comments Usage In the following example the compiler accepts C style comments pgcc B myprog cc b
244. lex function returns its value in memory The caller provides space for the return value and passes the address of this storage as if it were the first argument to the function Alternate return specifiers of a Fortran function are not passed as arguments by the caller The alternate return function passes the appropriate return value back to the caller in rax The handling of the following Fortran 90 features is implementation defined internal procedures pointer arguments assumed shape arguments functions returning arrays and functions returning derived types A 3 4 4 Interlanguage Calling Inter language calling between Fortran and C C is possible if function subroutine parameters and return values match types If a C C function returns a value call it from Fortran as a function otherwise call it as a subroutine If a Fortran function has type CHARACTER or COMPLEX call it from C C as a void function If a Fortran subroutine has alternate returns call it from C C as a Run time Environment 241 function returning int the value of such a subroutine is the value of the integer expression specified in the alternate RETURN statement If a Fortran subroutine does not contain alternate returns call it from C C as a void function The following table provides the C C data type corresponding to each Fortran data type 242 Run time Environment Arrays Table A 17 Fortran and C C Data Type Compatibility
245. lication you should refer to Chapter 5 OpenMP Directives for Fortran or Chapter 6 OpenMP Pragmas for C and C to see if insertion of explicit parallelization directives or pragmas and use of the mp compiler option enables the application to run in parallel 2 6 2 Loops That Fail to Parallelize In spite of the sophisticated analysis and transformations performed by the compiler programmers will often note loops that are seemingly parallel but are not parallelized In this subsection we ll look at some examples of common situations where parallelization does not occur 2 6 2 1 Innermost Loops As noted earlier in this chapter the PGI compilers will not parallelize innermost loops by default because it is usually not profitable You can override this default using the command line option Mconcur innermost 2 6 2 2 Timing Loops Often loops will occur in programs that are similar to timing loops The outer loop in the following example is one such loop Optimization amp Parallelization 27 do 1 j 1 2 do t T thy on a i b i c i lcontinue The outer loop above is not parallelized because the compiler detects a cross iteration dependence in the assignment to a i Suppose the outer loop were parallelized Then both processors would simultaneously attempt to make assignments into a 1 n Now in general the values computed by each processor for a 1 n will differ so that simultaneous assignment into a 1
246. licit instantiation of templates is implemented typename keyword is implemented explicit is accepted to declare Non converting constructors The scope of a variable declared in a for init statement of a loop is the scope of the loop not the surrounding scope Member templates are implemented The new specialization syntax using template lt gt is implemented Cy qualifiers are retained on rvalues in particular on function return values The distinction between trivial and nontrivial constructors has been implemented as has the distinction between PODs and non PODs with trivial constructors The linkage specification is treated as part of the function type affecting function overloading and implicit conversions extern inline functions are supported and the default linkage for inline functions is external A typedef name may be used in an explicit destructor call Placement delete is implemented An array allocated via a placement new can be deallocated via delete Covariant return types on overriding virtual functions are supported enum types are considered to be non integral types Partial specialization of class templates is implemented Partial ordering of function templates is implemented C Dialect Supported 297 e Function declarations that match a function template are regarded as independent functions not as guiding declarations that are instances of the template e It is possible to overload opera
247. ling Command Line Options Mrecursive Mnorecursive Mref_externals Mnoref_externals Mreentrant Mnoreentrant mpi perform MPI profiling available only in PGI CDK Cluster Development Kit configurations time Sample based instruction level profiling instructs the compiler to allow Fortran subprograms to be called recursively Fortran subprograms may not be called recursively force references to names appearing in EXTERNAL statements pgf77 pgf95 and pghpf only do not force references to names appearing in EXTERNAL statements pgf77 pgf95 and pghpf only instructs the compiler to avoid optimizations that can prevent code from being reentrant instructs the compiler not to avoid optimizations that can prevent code from being reentrant Msecond_underscore instructs the compiler to add a second underscore to the name of a Fortran global symbol if its name already contains an underscore This option is useful for maintaining compatibility with object code compiled using g77 which uses this convention by default pgf77 pgf95 and pghpf only Mnosecond_underscore Msignextend Mnosignextend Msafe_lastval Mstride0d Command Line Options instructs the compiler not to add a second underscore to the name of a Fortran global symbol if its name already contains an underscore pgf77 pgf95 and pghpf only instructs the compiler to extend the sign bit that is set as a result of convert
248. llow in names xh pgCC only Enable exception handling EDG option suffix see P pgCC only Use with E F or P to save intermediate file in a file with the specified suffix 3 1 Generic PGI Compiler Options Use the option to display the invocations of the compiler assembler and linker These invocations are command lines created by the driver from your command line input and the default values Default The compiler does not display individual phase invocations Usage The following command line requests verbose invocation information pgf95 prog f Cross reference Minfo V v HHH Use the option to display the invocations of the compiler assembler and linker but do not execute them These invocations are command lines created by the compiler driver from the PGIRC files and the command line options Default The compiler does not display individual phase invocations Usage The following command line requests verbose invocation information pgf95 myprog f Command Line Options 45 Cross reference Minfo V dryrun byteswapio Use the byteswapio option to swap the byte order of data in unformatted Fortran data files on input output When this option is used the order of bytes is swapped in both the data and record control words the latter occurs in unformatted sequential files Specifically this option can be used to convert big endian format
249. logical_ expression NUM_THREADS scalar_integer_expression SCHEDULE type chunk ORDERED The semantics of the PARALLEL WORKSHARE directive are identical to those of a parallel region containing a single WORKSHARE construct Note that the END PARALLEL WORKSHARE directive is optional and that NOWAIT may not be specified on an END PARALLEL WORKSHARE directive The available clauses are as defined in Section 5 2 PARALLEL END PARALLEL 5 12 SECTIONS END SECTIONS The OpenMP SECTIONS END SECTIONS directive pair uses the following syntax Syntax 118 OpenMP Directives for Fortran SOMP SECTIONS Clauses SOMP SECTION lt Fortran code block executed by processor i gt SOMP SECTION lt Fortran code block executed by processor j gt SOMP END SECTIONS NOWAIT Clauses PRIVATE list FIRSTPRIVATE list LASTPRIVATE list REDUCTION operator intrinsic list The SECTIONS END SECTIONS directives define a non iterative work sharing construct within a parallel region Each section is executed by a single processor If there are more processors than sections some processors will have no work and will jump to the implied barrier at the end of the construct If there are more sections than processors one or more processors will execute more than one section A SECTION directive may only appear within the lexical extent of the enclosing SECTIONS END SECTIONS directives In addition the co
250. long paral float para2 double para3 f return sum_3 1l_ paral f_para2 d_para3 j printf Parameter one type long 08x n 1 paral printf Parameter two type float f n f para2 printf Parameter three type double g n d_para3 printf The sum after conversion f n f return File sum_3 s Computes paral para2 para3 text align 16 globl sum_3 sum_3 pushq Srbp movq rsp rbp cvtsi2ssq rdi xmm2 addss xmm0 xmm2 cvtss2sd Sxmm2 xmm2 addsd Sxmm1 xmm2 cvtsd2ss Sxmm2 xmm2 movaps Sxmm2 xmm0 popg srbp ret type sum_3 function size sum_3 sum_3 Run time Environment 225 A ZA Linux86 64 Fortran Supplement Sections A2 4 1 through A2 4 4 define the Fortran supplement to the ABI for x64 Linux The register usage conventions set forth in that document remain the same for Fortran 226 Run time Environment A 2 4 1 Fortran Fundamental Types Table A 10 Linux86 64 Fortran Fundamental Types Fortran Type Te Ee Sa INTEGER 4 4 INTEGER 1 1 1 INTEGER 2 2 2 INTEGER 4 4 4 INTEGER 8 8 8 LOGICAL 4 4 LOGICAL 1 1 1 LOGICAL 2 2 2 LOGICAL 4 4 4 LOGICAL 8 8 8 BYTE 1 1 CHARACTER n n 1 REAL 4 4 REAL 4 4 4 REAL 8 8 8 DOUBLE PRECISION 8 8 COMPLEX 8 4 COMPLEX 8 8 4 COMPLEX 16 16 8 DOUBLE COMPLEX 16 8 A logical constant is one of TRUE FALSE Run time
251. lt omp h gt void omp_unset_lock omp_lock_t lock void omp_unset_nest_lock omp_nest_lock_t lock causes the calling thread to release ownership of the lock associated with lock It is illegal to make a call to this routine if lock has not been associated with a lock include lt omp h gt int omp_test_lock omp_lock_t lock int omp_test_nest_lock omp_nest_lock_t lock causes the calling thread to try to gain ownership of the lock associated with lock The function returns non zero if the thread gains ownership of the lock and zero otherwise It is illegal to make a call to this routine if lock has not been associated with a lock 6 16 Environment Variables OMP_NUM_THREADS specifies the number of threads to use during execution of parallel regions The default value for this variable is 1 For historical reasons the environment variable NCPUS is supported with the same functionality In the event that both OMP_NUM_THREADS and NCPUS are defined the value of OMP_NUM_THREADS takes precedence NOTE OMP_NUM_THREADS threads will be used to execute the program regardless of the number of physical processors available in the system As a result you can run programs using more threads than physical processors and they will execute correctly However performance of programs executed in this manner can be unpredictable and oftentimes will be inefficient OMP_SCHEDULE specifies the type of iteration scheduling to use for omp for
252. lt reserved message number gt F90 F90 F90 F90 COMMON 266 extension use of structure field reference extension nonstandard form of constant extension amp alternate return extension mixed non character and character elements in Messages W183 IS SI W184 W185 186 187 188 189 F90 extension mixed non character and character EQUIVALENCE Mixed type elements numeric and or character types in COMMON Mixed numeric and or character type EQUIVALENCE Argument missing for formal argument Too many arguments specified for Argument number to type mismatch Argument number to association of scalar actual argument to array dummy argument 190 191 192 W193 Argument number to non conformable arrays Argument number to cannot be an assumed size array Argument number to must be a label Argument number to does not match INTENT OUT Messages 267 W194 INTENT IN argument cannot be defined S195 Statement may not appear in an INTERFACE block 196 Deferred shape specifiers are required for 197 Invalid qualifier or qualifier value in OPTIONS statement An illegal qualifier was found or a value was specified for a qualifier which does not expect a value In either case the qualifier for which the error occurred is indicated in the error message S198 in ALLOCATE DEALLOCATE W199 Unaligned memory reference
253. lternate code altcode generation for vectorized loops For each loop the compiler decides whether to generate altcode and what type s to generate which may be any or all of altcode without iteration peeling altcode with non temporal stores and other data cache optimizations and altcode based on array alignments calculated dynamically at runtime The compiler also determines suitable loop count and array alignment conditions for executing the alternate code For a vectorized loop if possible generate an alternate vectorized loop containing additional aligned moves which is executed if a runtime array alignment test is passed For each auto parallelized loop generate an alternate serial loop to be executed if the loop count is less than or equal to n If n is omitted or n is O the compiler determines a suitable value of n for each loop This directive sets the loop count threshold for parallelization of reduction loops to n For each auto parallelized reduction loop generate an alternate serial loop to be executed if the loop count is less than or equal to n If n is omitted or n is 0 the compiler determines a suitable value of n for each loop For a vectorized loop if possible generate an alternate vectorized loop containing non temporal stores and other cache optimizations to be executed if the loop count is greater than n If n is omitted or n is 1 the compiler determines a suitable value of n for each loop The alternate code is
254. ly 20 source lines in the source file myprog f pgf95 Minfo inline Minline 20 myprog f In the following example the assembler does not delete the assembly file myprog s after the assembly pass pgf95 Mkeepasm myprog f In the following example the compiler creates the listing file myprog Ist pgf95 Mlist myprog f 82 Command Line Options In the following example array bounds checking is enabled pgf95 Mbounds myprog f Cross reference m S V v mcmodel medium For use only on 64 bit Linux targets Generate code for the medium memory model in the linux86 64 execution environment Implies Mlarge_arrays The default small memory model of the linux86 64 environment limits the combined area for a user s object or executable to 1GB with the Linux kernel managing usage of the second 1GB of address for system routines shared libraries stacks etc Programs are started at a fixed address and the program can use a single instruction to make most memory references The medium memory model allows for larger than 2GB data areas or bss sections Program units compiled using either mcmodel medium or fpic require additional instructions to reference memory The effect on performance is a function of the data use of the application The memodel medium switch must be used at both compile time and link time to create 64 bit executables Program units compiled for the default small memory model can be li
255. m You can specify more than one path filename 1 2 2 Command line Options The command line options control various aspects of the compilation process For a complete alphabetical listing and a description of all the command line options refer to Chapter 3 Command Line Options 1 2 3 Fortran Directives and C C Pragmas Fortran directives or C C pragmas inserted in program source code allow you to alter the effects of certain command line options and control various aspects of the compilation process for a specific routine or a specific program loop For a complete alphabetical listing and a description of all the Fortran directives and C C pragmas refer to Chapter 5 OpenMP Directives for Fortran Chapter 6 OpenMP Pragmas for C and C and Chapter 7 Optimization Directives and Pragmas 1 3 Filename Conventions The PGI compilers use the filenames that you specify on the command line to find and to create input and output files This section describes the input and output filename conventions for the phases of the compilation process 1 3 1 Input Files You can specify assembly language files preprocessed source files Fortran C C source files object files and libraries as inputs on the command line The compiler driver determines the type of each input file by examining the filename extensions The drivers use the following conventions filename f indicates a Fortran source file Getting Started 3
256. m_threads is used to change the number of processors subsequent calls to omp_get_max_threads will return the new value This function returns the maximum value whether executing from a parallel or serial region of code include lt omp h gt int omp_get_num_procs void returns the number of processors that are available to the program include lt omp h gt int omp_in parallel void returns non zero if called from within a parallel region and zero if called outside of a parallel region When called from within a parallel region that is serialized for example in the presence of an if clause evaluating to zero the function will return zero include lt omp h gt void omp_set_dynamic int dynamic_threads is designed to allow automatic dynamic adjustment of the number of threads used for execution of parallel regions This function is recognized but currently has no effect include lt omp h gt int omp get _dynamic void OpenMP Pragmas for C and C 137 is designed to allow the user to query whether automatic dynamic adjustment of the number of threads used for execution of parallel regions is enabled This function is recognized but currently always returns zero include lt omp h gt void omp_set_nested int nested is designed to allow enabling disabling of nested parallel regions This function is recognized but currently has no effect include lt omp h gt int omp_get_nested void is designed
257. may be a sub region of code that should execute only on the master thread Instead of ending the parallel region before this subregion and then starting it up again after this subregion the omp master pragma allows the user to conveniently designate code that executes on the master thread and is skipped by the other threads There is no implied barrier on entry to or exit from a master section Nested master sections are ignored Branching into or out of a master section is not supported OpenMP Pragmas for C and C 129 include lt stdio h gt include lt omp h gt main int a 2 1 1 pragma omp parallel alomp get _thread_num omp_get_thread_num pragma omp master printf YOU SHOULD ONLY SEE THIS ONCE n printf a 0 d a 1 d n a 0 a 1 6 5 omp single The OpenMP omp single pragma uses the following syntax pragma omp single Clauses lt C C structured block gt Clauses private list firstprivate list copyprivate list nowait In a parallel region of code there may be a subregion of code that will only execute correctly on a single thread Instead of ending the parallel region before this subregion and then starting it up again after this subregion the omp single pragma allows the user to conveniently designate code that executes on a single thread and is skipped by the other threads There is an implied barrier on exit from a single process section unless the optional nowa
258. mbers For example the union un1 below aligns on a 4 byte boundary since the alignment of c the most restrictive element is four union unl short a 2 bytes char b 1 byte int c 4 bytes Fortran C and C Data Types 183 Structure alignment can result in unused space called padding Padding between members of a structure is called internal padding Padding between the last member and the end of the space occupied by the structure is called tail padding Figure 9 1 Internal Padding in a Structure illustrates structure alignment Consider the following structure struct strci char a occupies byte 0 short b occupies bytes 2 and 3 char c occupies byte 4 int d occupies bytes 8 through 11 Figure 9 1 Internal Padding in a Structure Figure 9 2 Tail Padding in a Structure shows how tail padding is applied to a structure aligned on a doubleword boundary struct strce2 int m1 4 occupies bytes 0 through 15 double m2 occupies bytes 16 through 23 short m3 occupies bytes 24 and 25 st 9 2 5 Bit field Alignment Bit fields have the same size and alignment rules as other aggregates with several additions to these tules S Bit fields are allocated from right to left S A bit field must entirely reside in a storage unit appropriate for its type Bit fields never cross unit boundaries Bit fields may share a storage unit with other structure unio
259. mplements a floating point stack using 8 80 bit registers Each register uses bits 0 63 as the significand bits 64 78 for the exponent and bit 79 is the sign bit This 80 bit real format is the default format called the extended format When values are loaded into the floating point stack they are automatically converted into extended real format The precision of the floating point stack can be controlled however by setting the precision control bits bits 8 and 9 of the floating control word appropriately In this way you can explicitly set the precision to standard IEEE double precision using 64 bits or to single precision using 32 bits The default precision is system dependent To alter the precision in a given program unit the main program must be compiled with the same pc option The command line option pc val lets the programmer set the compiler s precision preference Valid values for val are 32 single precision 64 double precision e 80 extended precision 1 According to Intel documentation this only affects the x87 operations of add subtract multiply divide and square root In particular it does not appear to affect the x87 tran scendental instructions 86 Command Line Options Extended Precision Option Operations performed exclusively on the floating point stack using extended precision without storing into or loading from memory can cause problems with accumulated values within the extra 16 bits of
260. mutex Monetrip Mnoonetrip Msave Mnosave Command Line Options the compiler recognizes a backslash as an escape character in quoted strings in accordance with standard C usage the compiler requires that all program variables be declared the compiler does not require that all program variables be declared the compiler treats as a synonym for standard input for reading and standard output for writing the compiler treats as a synonym for unit 5 on input and unit 6 on output the compiler treats lines containing D in column 1 as executable statements ignoring the D the compiler does not treat lines containing D in column 1 as executable statements does not ignore the D char specifies the character to which the compiler maps the dollar sign The compiler allows the dollar sign in names with Mextend the compiler accepts 132 column source code otherwise it accepts 72 column code with Mfixed the compiler assumes input source files are in FORTRAN 77 style fixed form format with Mfree the compiler assumes the input source files are in Fortran 90 95 freeform format the compiler generates critical section calls around Fortran I O statements the compiler does not generate critical section calls around Fortran I O statements the compiler forces each DO loop to execute at least once the compiler does not force each DO loop to execute at least once This option is usefu
261. my argument The call statement passes a subroutine name to a dummy argument that is used as a function W360 FUNCTION passed to FUNCTION dummy argument with different result type The call statement passes a function argument to a function dummy argument but the dummy has a different result type W361 FUNCTION passed to SUBROUTINE dummy argument The call statement passes a function name to a dummy argument that is used as a subroutine W362 Argument has a different type than dummy argument The type of the actual argument is different than the type of the corresponding dummy argument W363 Dummy argument is a POINTER but actual argument is not The dummy argument is a pointer so the actual argument must be also W364 Array or array expression passed to scalar dummy argument The actual argument is an array but the dummy argument is a scalar variable W365 Scalar or scalar expression passed to array dummy argument The actual argument is a scalar variable but the dummy argument is an array F366 Internal error interprocedural analysis fails An internal error occurred during interprocedural analysis please report this to the compiler maintenance group If user errors were reported when collecting IPA information or during IPA analysis correcting them may avoid this error I367 Array bounds cannot be matched to formal argument Passing a nonsequential array to a sequential dummy argument may require
262. n extern float sum_3 long paral float para2 double para3 Z return sum_3 1 paral f _para2 d_para3 j printf Parameter one type long 08x n 1 paral printf Parameter two type float f n f para2 printf Parameter three type double Sonn d_para3 printf The sum after conversion f n f return File sum_3 s Computes paral para2 para3 etext align 4 Long EN1 sum_3 0xc8000000 align 16 globl sum_3 sum_3 pushl ebp movl esp Sep subl 8 esp ENT fildl 8 ebp fadds 12 ebp faddl 16 ebp fstps 4 ebp flds 4 ebp leave ret type sum_3 function Size sum_3 sum_3 Run time Environment 217 A 2 Linux86 64 Programming Model This section defines compiler and assembly language conventions for the use of certain aspects of an x64 processor running a linux86 64 operating system These standards must be followed to guarantee that compilers application programs and operating systems written by different people and organizations will work together The conventions supported by the PGCC ANSI C compiler implement the application binary interface ABI as defined in the System V Application Binary Interface AMD64 Architecture Processor Supplement and the System V Application Binary Interface listed in the Related Publications section in the Preface A 2 1 Function Calling Sequence This section describes the standard function calling sequence including the stack frame
263. n will produce values different from sequential execution of the loops In this example values computed for a 1 n don t depend on j so that simultaneous assignment by both processors will not yield incorrect results However it is beyond the scope of the compilers dependence analysis to determine that values computed in one iteration of a loop don t differ from values computed in another iteration So the worst case is assumed and different iterations of the outer loop are assumed to compute different values for a 1 n Is this assumption too pessimistic If j doesn t occur anywhere within a loop the loop exists only to cause some delay most probably to improve timing resolution And it s not usually valid to parallelize timing loops to do so would distort the timing information for the inner loops 2 6 2 3 Scalars Quite often scalars will inhibit parallelization of non innermost loops There are two separate cases that present problems In the first case scalars appear to be expandable but appear in non innermost loops as in the following example a i j x c i j 1 continue There are a number of technical problems to be resolved in order to recognize expandable scalars in non innermost loops Until this generalization occurs scalars like x above will inhibit parallelization of loops in which they are assigned In the following example scalar k is not expandable and it is not an accumulator for a reduction 2 if
264. n a parallel region or within a subroutine or function that is called from within a parallel region the results are undefined This subroutine has precedence over the OMP_NUM_THREADS environment variable integer omp_get_thread_num returns the thread number within the team The thread number lies between 0 and omp_get_num_threads 1 When called from a serial region this function returns 0 A nested parallel region is the same as a single parallel region integer function omp_get_max_threads returns the maximum value that can be returned by calls to omp_get_num_threads If omp_set_num_threads is used to change the number of processors subsequent calls to omp_get_max_threads will return the new value This function returns the maximum value whether executing from a parallel or serial region of code integer function omp_get_num procs returns the number of processors that are available to the program logical function omp_in parallel returns TRUE if called from within a parallel region and FALSE if called outside of a parallel region When called from within a parallel region that is serialized for example in the presence of an IF clause evaluating FALSE the function will return FALSE subroutine omp_set_dynamic scalar_logical_exp is designed to allow automatic dynamic adjustment of the number of threads used for execution of parallel regions This function is recognized but currently has no effect logic
265. n another common block 059 Conflicting equivalence between and The indicated equivalence implies a storage layout inconsistent with other equivalences 060 Illegal equivalence of structure variable Messages 255 STRUCTURE and UNION variables may not appear in EQUIVALENCE statements S061 Equivalence of and extends common block backwards W062 Equivalence forces to be unaligned EQUIVALENCE statements have defined an address for the variable which has an alignment not optimal for variables of its data type This can occur when INTEGER and CHARACTER data are equivalenced for instance I063 Gap in common block before 064 Illegal use of in DATA statement implied DO loop The indicated variable is referenced where it is not an active implied DO index variable S065 Repeat factor less than zero S066 Too few data constants in initialization statement S067 Too many data constants in initialization statement 068 Numeric initializer for CHARACTER out of range 0 through 255 A CHARACTER 1 variable or character array element can be initialized to an integer octal or hexadecimal constant if that constant is in the range 0 through 255 S069 Illegal implied DO expression The only operations allowed within an implied DO expression are integer and S070 Incorrect sequence of statements The statement order is incorrect For instance an IMPLICIT NONE statement must precede a specification statem
266. n array is selected from a class the type qualifiers on the class object if any are not preserved in the selected array In the normal mode any type qualifiers on the object are preserved in the element type of the resultant array e An identifier in a function is allowed to have the same name as a parameter of the function A warning is issued e An expression of type void may be supplied on the return statement in a function with a void return type A warning is issued e cfront has a bug that causes a global identifier to be found when a member of a class or one of its base classes should actually be found This bug is not emulated in cfront compatibility mode A parameter of type const void is allowed on operator delete it is treated as equivalent to void Zen e A period may be used for qualification where should be used Only may be used as a global qualifier Except for the global qualifier the two kinds of qualifier operators may not be mixed in a given name i e you may say A B C or A B C but not A B C or A B C A period may not be used in a vacuous destructor reference nor in a qualifier that follows a template reference such as A lt T gt B cfront 2 1 does not correctly look up names in friend functions that are inside class definitions In this example function f should refer to the functions and variables e g f1 and a1 from the class declaration Instead the global definitions are used
267. n members including members that are not bit fields Unnamed bit field s types do not affect the alignment of a structure or union Items of signed unsigned long long type may not appear in field declarations 184 Fortran C and C Data Types Figure 9 2 Tail Padding in a Structure st m1 0 st m1 1 st ml 2 st ml 3 m2 m2 XXXX 9 2 6 Other Type Keywords in C and C byte 0 byte 4 byte 8 byte 12 byte 16 byte 20 byte 24 byte 28 The void data type is neither a scalar nor an aggregate You can use void or void as the return type of a function to indicate the function does not return a value or as a pointer to an unspecified data type respectively The const and volatile type qualifiers do not in themselves define data types but associate attributes with other types Use const to specify that an identifier is a constant and is not to be changed Use volatile to prevent optimization problems with data that can be changed from outside the program such as memory mapped I O buffers Fortran C and C Data Types 185 186 Fortran C and C Data Types Chapter 10 Inter language Calling This chapter describes inter language calling conventions for C C and Fortran programs using the PGI compilers The following sections describe how to call a Fortran function or subroutine from aC or C program and how to call a C or C function from a Fortran program For information on calling assembly languag
268. n procedures and does not change any known global variables Without Mipa safe any unknown procedures will cause IPA to fail declares that all unknown procedures are safe see Mipa safe default is nosafeall perform Fortran 90 array shape propagation default is noshape only collect IPA summary information when compiling this prevents IPA optimization of this file but allows optimization for other files linked with this file remove uncalled vestigial functions default is novestigial M1 re array assoc noassoc Enables loop carried redundancy elimination an optimization that can reduce the number of arithmetic operations and memory references in loops Mnolre Mnoframe 74 array assoc noassoc treat individual array element references as candidates for possible loop carried redundancy elimination The default is to eliminate only redundant expressions involving two or more operands allow expression re association specifying this sub option can increase opportunities for loop carried redundancy elimination but may alter numerical results disallow expression re association Disables loop carried redundancy elimination Eliminates operations that set up a true stack frame pointer for every function With this option enabled you cannot perform a traceback on the generated code and you cannot access local variables Command Line Options Mnoi4 Mpfi pgf77 and pgf95 only the c
269. n with one or more additional options usage information for those options is displayed to standard output Usage In the following example usage information for Minline is printed to standard output pgcc help Minline Minline lib lt inlib gt lt func gt except lt func gt name lt func gt size lt n gt levels lt n gt Enable function inlining lib lt extlib gt Use extracted functions from extlib lt func gt Inline function func except lt func gt Do not inline function func name lt func gt Inline function func size lt n gt Inline only functions smaller than n levels lt n gt Inline n levels of functions Minline Inline all functions that were extracted In the following example usage information for help shows how groups of options can be listed or examined according to function pgcc help help help groups asm debug language linker opt other overall phase prepro suffix switch target variable Show compiler switches Cross reference 2038 show V flags I Adds a directory to the search path for files that are included using the INCLUDE statement or the preprocessor directive include Use the I option to add a directory to the list of where to search for the included files The compiler searches the directory specified by the I option before the default directories stdinc Syntax Idirectory Where directory is the name of the directory added to the
270. nMP Pragmas for C and C It should be emphasized that by default there is no work distribution in a parallel region Each active thread executes the entire region redundantly until it encounters a directive that specifies work distribution For work distribution see the omp for pragma include lt stdio h gt include lt omp h gt main int a 2 1 1 pragma omp parallel afomp_get_thread_num omp_get_thread_num printf a 0 sd all d a 0 a 1 The variables specified in a private list are private to each thread in a team In effect the compiler creates a separate copy of each of these variables for each thread in the team When an assignment to a private variable occurs each thread assigns to its local copy of the variable When operations involving a private variable occur each thread performs the operations using its local copy of the variable Other important points to note about private variables are the following e Variables declared private in a parallel region are undefined upon entry to the parallel region If the first use of a private variable within the parallel region is in a right hand side expression the results of the expression will be undefined i e this is probably a coding error e Likewise variables declared private in a parallel region are undefined when serial execution resumes at the end of the parallel region The variables specified in a shared list are shared between
271. nal argument is the length of the return value If a Fortran function is declared to return a character value of constant length for example CHARACTER 4 FUNCTION CHF the second extra parameter representing the length of the return value must still be supplied A Fortran complex function returns its value in memory The caller provides space for the return value and passes the address of this storage as if it were the first argument to the function Alternate return specifiers of a Fortran function are not passed as arguments by the caller The alternate return function passes the appropriate return value back to the caller in rax The handling of the following Fortran 90 features is implementation defined internal procedures pointer arguments assumed shape arguments functions returning arrays and functions returning derived types A 2 4 4 Inter language Calling Inter language calling between Fortran and C C is possible if function subroutine parameters and return values match types If a C C function returns a value call it from Fortran as a function otherwise call it as a subroutine If a Fortran function has type CHARACTER or COMPLEX call it from C C as a void function If a Fortran subroutine has alternate returns call it from C C as a 228 Run time Environment function returning int the value of such a subroutine is the value of the integer expression specified in the alternate RETURN statement If a Fortran subroutin
272. nd In this case where the variables are implicitly real 4 operations are performed on the floating point stack where optimization removed unnecessary loads and stores from memory The general case of copy propagation being performed follows this pattern a x y 2 0 a Instead of storing x into a then loading a to perform the addition the value of x can be left on the floating point stack and added to 2 0 Thus memory accesses in some cases can be avoided leaving answers in the extended real format If copy propagation is disabled stores of all left hand sides will be performed automatically and reloaded when needed This will have the effect of rounding any results to their declared sizes Command Line Options 87 For the above program w has a value of 1 8446744E 19 when executed using default extended precision If however Kieee is set the value becomes 1 6777216E 07 single precision This difference is due to the fact that Kieee disables copy propagation so all intermediate results are stored into memory then reloaded when needed Copy propagation is only disabled for floating point operations not integer With this particular example setting the pc switch will also adjust the result The switch Kieee also has the effect of making function calls to perform all transcendental operations Although the function still produces the x86 machine instruction for computation unless in C the Mnobuiltin switch is set arguments are
273. ndary LOGICAL 8 8 byte boundary BYTE 1 byte boundary INTEGER 2 2 byte boundary INTEGER 4 4 byte boundary INTEGER 8 8 byte boundary REAL 4 4 byte boundary REAL 8 8 byte boundary COMPLEX 8 4 byte boundary COMPLEX 16 8 byte boundary 9 1 2 FORTRAN 77 Aggregate Data Type Extensions The PGF77 compiler supports de facto standard extensions to FORTRAN 77 that allow for aggregate data types An aggregate data type consists of one or more scalar data type objects You can declare 178 Fortran C and C Data Types the following aggregate data types array consists of one or more elements of a single data type placed in contiguous locations from first to last structure is a structure that can contain different data types The members are allocated in the order they appear in the definition but may not occupy contiguous locations union is a single location that can contain any of a specified set of scalar or aggregate data types A union can have only one value at a time The data type of the union member to which data is assigned determines the data type of the union after that assignment The alignment of an array a structure or union an aggregate affects how much space the object occupies and how efficiently the processor can address members Arrays use the alignment of their members Array types align according to the alignment of the array elements For example an array of INTEGER 2 data aligns on a 2 byte boundary Structure
274. ne When the pragma safe is not followed by a variable name or name list all pointer arguments appearing in a routine if scope is routine or all routines if scope is global will be treated as safe If variable names occur after safe each name is the name of a pointer argument in the current function The named argument is considered to be safe Note that if just one variable name is specified the surrounding parentheses may be omitted There is no command line option corresponding to this pragma safeptr nosafeptr The pragma safeptr directs the compiler to treat pointer variables of the indicated storage class as safe The pragma nosafeptr directs the compiler to treat pointer variables of the indicated storage class as unsafe This pragma alters the effects of the Msafeptr command line option The syntax of this pragma is pragma scope value where value is no safeptr arg local auto global static all Note that the values local and auto are equivalent For example in a file containing multiple functions the command line option Msafeptr might be helpful for one function but can t be used because another function in the file would produce incorrect results In such a file the safeptr pragma used with routine scope could improve performance and produce correct results single nosingle The pragma single directs the compiler not to convert float parameters to double in non prototyped functions This can r
275. ne Options pgf77 and pgf95 only Force Cray Fortran CF77 compatibility with respect to the listed options Possible values of option include pointer for purposes of optimization it is assumed that pointer based variables do not overlay the storage of any other variable instructs the compiler to assume unresolved data dependencies actually conflict 71 instructs the compiler to assume potential data dependencies do not conflict However if data dependencies exist this option can produce incorrect code Mfprelaxed option Mnofprelaxed Mi4 instructs the compiler to use relaxed precision in the calculation of some intrinsic functions Can result in improved performance at the expense of numerical accuracy The possible values for option are div Perform divide using relaxed precision sqrt Perform square root with relaxed precision rsqrt Perform reciprocal square root 1 sqrt using relaxed precision With no options Mfprelaxed will choose generate relaxed precision code for those operations that generate a significant performance improvement depending on the target processor default instructs the compiler not to use relaxed precision in the calculation of intrinsic functions pgf77 and pgf95 only the compiler treats INTEGER variables as INTEGER 4 Mipa lt option gt lt option gt Pass options to the interprocedural analyzer Note Mipa implies O2 72 and the minimum optimiza
276. ng against any PGI compiled DLL such as obj2 dll The defaultlib switch is used to specify that obj2 lib the DLL s import library should be used to resolve the imported data and subroutine in prog c Step 3 Ensure that obt dl is in your path then run the executable prog2 to determine if the DLL was successfully created and linked o prog2 in func2 data 11 Should you wish to change obj2 dll without changing the subroutine or function interfaces no rebuilding of prog2 is necessary Just recreate obj2 dll and the new obt dl will be loaded at runtime Example 3 Build two DLLs when each DLL is dependent on the other and use them to build the main program In the following source files object3 c makes calls to routines defined in object4 c and vice versa This situation of mutual imports requires two steps to build each DLL object3 c void _ declspec dllimport func_4b void void _ declspec dllexport func_3a void printf func 3a calling a routine in obj4 dl11l n func _4b void _ declspec dllexport func_3b void printf func_3b n 168 Libraries and Environment Variables object4 c void _ declspec dllimport func_3b void void _ declspec dllexport func_4a void printf func_4a calling a routine in obj3 d11 n func _3b void _ declspec dllexport func_4b void printf func _4b n prog3 c void _ declspec dllimport func_3a void vo
277. nked into medium memory model executables as long as they are compiled fpic or position independent The linux86 64 environment provides static libxxx a archive libraries that are built with and without fpic and dynamic libxxx so shared object libraries that are compiled fpic The mcmodel medium linkswitch implies the fpic switch and will utilize the shared libraries by default Similarly the PGI linux86 64 lt rel gt lib directory contains the libraries for building small memory model codes and the PGI linux86 64 lt rel gt libso directory contains shared libraries for building mcmodel medium and fpic executables Note It appears from the GNU tools and documentation that creation of medium memory model shared libraries is not supported However you can create static archive libraries a that are fpic Default The compiler generates code for the small memory model Usage The following command line requests position independent code be generated and the mcmodel medium option be passed to the assembler and linker pgf95 mcmodel medium myprog f module lt moduledir gt Use the module option to specify a particular directory in which generated intermediate mod files should be placed If the module lt moduledir gt option is present and USE statements are present in a compiled program unit lt moduledir gt will search for mod intermediate files prior to the search in the default local directory
278. nneren xi le INECC xii Hardware and Software Constraints c cccceceeeeeeeeeeeeeeeseeeseeeeeeeenaeeeeeas xiii GONVENTIONS erst std adnan aie ae eee ieee ee xiii Related Pubcatons AA xvii Getti g Started E 1 OETAN EE E TE E T TE 1 Invoking the Command level PGI Compilers cccecceeeeeeeeeeeeeteeeeeeeees 1 Command line Syntax ccccccccceeeeeseneeeeeeeeeeeeeeeeseaeeeeeeeeeseeaeeeseaeeeeesaeeeenaeees 2 Command line Options ccccccecceeeeeeeeeeeeeeeeneeeseaaeeeeeeeeseeaeesseaeeeeecaeeeeaeees 3 Fortran Directives and C C Pragmas cccceceesceeeeeeeeeeeeeeseeeeessaaeeeenes 3 Filename Conventions sirmis aehan eutanaasia avia eadra 3 delt e UE 3 lu e IGS sack ei Ae ae eh retin eit he et abe 5 Parallel Programming Using the PGI Compilers csseeeeeeeeeeeteeeeeeees 7 Running SMP Parallel Programs cccccceceeeeeeeeeeeeecaeeeeeaeeeeeeaeeeseaeeeeaas 7 Running Data Parallel HPF Programs cccceceeeeeeeeeeeeeeneeeeeeeeenaaeeeeeeeaaas 8 Using the PGI Compilers on Linux 00 cececceeeceeeeeeeeeeeeceeeeaeeeseaeeeesaeeeeaes 9 Linux Header FIE e vice a EEEE nis 9 Running Parallel Programs On LINUX eee ceeeeeeeeeee eee eeeeaeeeeeeeeaaeeeeeeeaees 9 Using the PGI Compilers On Windows cccceceeeeeeeeeeeeeeeeeeeeseeeeeeeneeeees 10 BASH Shell Environment c cccccceeeeeceeceeeeeeeeeeeeeeeeeeaeeseeeeeesaeeseeeeeees 10 Command Prompt Window cceccceeceeeeeee
279. noptimized code compiled using the option O0 can be significantly slower than code generated at other optimization levels The M lt opt gt option where lt opt gt is vect concur unroll or ipa sets the optimization level to level 2 if no O options are supplied The fast and fastsse options set the optimization level to a target dependent optimization level if no O options are supplied 2 10 Local Optimization Using Directives and Pragmas Command line options let you specify optimizations for an entire source file Directives supplied within a Fortran source file and pragmas supplied within a C or C source file provide information to the compiler and alter the effects of certain command line options or default behavior of the compiler many directives have a corresponding command line option While a command line option affects the entire source file that is being compiled directives and pragmas let you do the following e Apply or disable the effects of a particular command line option to selected subprograms or to selected loops in the source file for example an optimization Globally override command line options e Tune selected routines or loops based on your knowledge or on information obtained through profiling Chapter 7 Optimization Directives and Pragmas provides details on how to add directives and pragmas to your source files Optimization amp Parallelization 37 2 11 Execution Timing and
280. nual for list of allowed compiler directives S021 Label field of continuation line is not blank Messages 251 The first five characters of a continuation line must be blank S022 Unexpected end of file missing END statement 023 Syntax error unbalanced Unbalanced parentheses or brackets W024 CHARACTER or Hollerith constant truncated to fit data type A character or hollerith constant was converted to a data type that was not large enough to contain all of the characters in the constant This type conversion occurs when the constant is used in an arithmetic expression or is assigned to a non character variable The character or hollerith constant is truncated on the right that is if 4 characters are needed then the first 4 are used and the remaining characters are discarded w025 Illegal character ignored The current line contains a character possibly non printing which is not a legal Fortran character characters inside of character or Hollerith constants cannot cause this error As a general rule all non printing characters are treated as white space characters blanks and tabs no error message is generated when this occurs If for some reason a non printing character is not treated as a white space character its hex representation is printed in the form dd where each d is a hex digit S026 Unmatched quote 027 Illegal integer constant Integer constant is too large for 32 bit word 028 Illegal real o
281. o the PGI compiler commands referenced above within BASH you have access to over 100 common commands and utilities including but not limited to the following vi emacs make tar untar gzip gunzip ftp sed grep egrep fgrep awk cat cksum cp date diff du find kill Is more less mv printenv env rm rmdir touch wc If you are familiar with program development in a Linux environment editing compiling and executing programs within BASH will be very comfortable If you have not previously used such an environment you should take time to familiarize yourself with either the vi or emacs editors and with makefiles The emacs editor has an extensive online tutorial which you can start by bringing up emacs and selecting the appropriate option under the pull down help menu You can get a thorough introduction to the construction and use of makefiles through the online Makefile User s Guide 10 Getting Started 1 6 2 Command Prompt Window The PGI submenu in the Windows Start menu contains an item named PVF Command Prompt This is used to launch a Windows Command Prompt window that has an environment pre initialized for usage of the PGI compilers and tools On x64 systems there are two selections one is tagged x64 to denote that its environment is pre initialized for the 64 bit compilers and tools Getting Started 12 Getting Started Chapter 2 Optimization amp Parallelization Source code that is readable maintainabl
282. of bits in the bit field bit field 2 signed value 1 to 32 two s complement 28ize to 2812 11 where bits integer size is the number of bits in the bit field pointer 4 address 0 to 232 1 enum 4 two s complement 281 to 231 1 integer Fortran C and C Data Types 181 1 Approximate value 2 Bit fields occupy as many bits as you assign them up to 4 bytes and their length need not be a multiple of 8 bits 1 byte Table 9 5 Scalar Alignment Data Type Alignment char is aligned on a 1 byte boundary short is aligned on a 2 byte boundary long int is aligned on a 4 byte boundary enum is aligned on a 4 byte boundary pointer is aligned on a 4 byte boundary float is aligned on a 4 byte boundary double is aligned on an 8 byte boundary long double is aligned on an 8 byte boundary signed or unsigned 9 2 2 Cand C Aggregate Data Types An aggregate data type consists of one or more scalar data type objects You can declare the following aggregate data types array consists of one or more elements of a single data type placed in contiguous locations from first to last class C only is a class that defines an object and its member functions The object can contain fundamental data types or other aggregates including other classes The class members are allocated in the order they appear in the definition but may not occupy contiguous locations struct is a structure t
283. of pgftn as or Id Qpathpathname The third Q variety using the path keyword lets you supply an additional pathname to the search path for the compiler s required o files Qproducesourcetype The fourth Q variety using the produce keyword lets you choose a stop after location for the compilation based on the supplied sourcetype parameter Valid sourcetypes are i c s and o These indicate respectively stop after preprocessing compiling assembling or linking Usage The following examples show the different Q options pgf95 Qproduce s hello f pgf95 Qoption ld s hello f pgf95 Qpath home test hello f pgf95 Qdir home comp new hello f Cross reference p R lt directory gt Valid only on Linux and is passed to the linker Instructs the linker to hard code the pathname lt directory gt into the search path for generated shared object dynamically linked library files Note that there cannot be a space between R and lt directory gt Cross reference fpic shared G r4 and r8 Interpret DOUBLE PRECISION variables as REAL r4 or REAL variables as DOUBLE PRECISION r8 Usage pgf95 r4 myprog f Cross reference i2 i4 i8 nor8 Command Line Options 89 rc Specifies the name of the driver startup configuration file If the file or pathname supplied is not a full pathname the path for the configuration file loaded is relative to the DRIVER path the path
284. ograms in which there is known to be no pointer aliasing However for obvious reasons this command line option must be used carefully Three other options which are extremely useful are help Minfo and dryrun You can see a specification of any command line option by invoking any of the PGI compilers with help in combination with the option in question without specifying any input files For example Optimization amp Parallelization 15 pgf95 help fastsse Reading rcfile usr pgi_rel linux86 64 6 0 bin pgf95re fastsse fast Mvect sse Mcache align Mflushz fast Common optimizations 02 Munroll c 1 Mnoframe Mlre Or to see the full functionality of help itself which can return information on either an individual option or groups of options by type pgf95 help help Reading rcfile usr pgi_rel linux86 64 6 0 bin pgf95re help groups asm debug language linker opt other overal1 phase prepro suffix switch target variable The Minfo option can be used to display compile time optimization listings When this option is used the PGI compilers will issue informational messages to stdout as compilation proceeds From these messages you can determine which loops are optimized using unrolling SSE instructions vectorization parallelization interprocedural optimizations and various miscellaneous optimizations You can also see where and whether functions are inlined The Mneginfo option can be u
285. omp for loops must be executed by all threads participating in the parallel region or none at all e The for loop must be a structured block and its execution must not be terminated by break OpenMP Pragmas for C and C 131 Values of the loop control expressions and the chunk expressions must be the same for all threads executing the loop include lt stdio h gt include lt math h gt main float a 1000 b 1000 int i for i 0 i lt 1000 i b i i pragma omp parallel pragma omp for for i 0 i lt 1000 i ali sqrt b il l The schedule clause specifies how iterations of the for loop are divided up between processors Given a schedule kind chunk clause kind can be static dynamic guided or runtime These are defined as follows 6 7 When schedule static chunk is specified iterations are allocated in contiguous blocks of size chunk The blocks of iterations are statically assigned to threads in a round robin fashion in order of the thread ID numbers The chunk must be a scalar integer expression If chunk is not specified a default chunk size is chosen equal to number_of_iterations omp_num_threads 1 omp_num_threads When schedule dynamic chunk is specified iterations are allocated in contiguous blocks of size chunk As each thread finishes a piece of the iteration space it dynamically obtains the next set of iterations The chunk must be a scalar integer expression If no
286. ompatibility with programs parallelized using legacy SGI style directives Syntax 116 OpenMP Directives for Fortran CSDOACROSS Clauses lt Fortran DO loop to be executed in parallel gt Clauses PRIVATE LOCAL list SHARED SHARE list MP_SCHEDTYPE SIMPLE INTERLEAVE CHUNK lt integer_expression gt IF logical _expression The C DOACROSS directive has the effect of a combined parallel region and parallel DO loop applied to the loop immediately following the directive It is very similar to the OpenMP PARALLEL DO directive but provides for backward compatibility with codes parallelized for SGI systems prior to the OpenMP standardization effort The CSDOACROSS directive must not appear within a parallel region It is a shorthand notation that tells the compiler to parallelize the loop to which it applies even though that loop is not contained within a parallel region While this syntax is more convenient it should be noted that if multiple successive DO loops are to be parallelized it is more efficient to define a single enclosing parallel region and parallelize each loop using the OpenMP DO directive A variable declared PRIVATE or LOCAL to a CSDOACROSS loop is treated the same as a private variable in a parallel region or DO see above A variable declared SHARED or SHARE to a C DOACROSS loop is shared among the threads meaning that only 1 copy of the variable exists to be used and or modifie
287. ompilation If pch automatic PCH mode appears on the command line following this option its effect is erased no_ using_std pgCC only Enable or disable implicit use of the std namespace when standard header files are included Default The default is using_std Usage The following command line disables implicit use of the std namespace pgCC no_using std hello cc t pgCC only Control instantiation of template functions Syntax t arg where arg is one of the following all Instantiates all functions whether or not they are used Command Line Options 99 local Instantiates only the functions that are used in this compilation and forces those functions to be local to this compilation Note This may cause multiple copies of local static variables If this occurs the program may not execute correctly none Instantiates no functions this is the default used Instantiates only the functions that are used in this compilation Usage In the following example all templates are instantiated pgCC tall myprog cc 100 Command Line Options Chapter A Function Inlining Function inlining replaces a call to a function or a subroutine with the body of the function or subroutine This can speed up execution by eliminating parameter passing and function subroutine call and return overhead It also allows the compiler to optimize the function with the rest of the code Note that using function inlining indi
288. ompiler treats INTEGER variables as INTEGER 2 generate profile feedback instrumentation this includes extra code to collect run time statistics and dump them to a trace file for use in a subsequent compilation Mpfi must also appear when the program is linked When the resulting program is executed a profile feedback trace file pgfi out is generated in the current working directory see Mpfo Note compiling and linking with Mpfi adds significant runtime overhead to almost any executable you should use executables compiled with Mpfi only for execution of training runs Mpfo enable profile feedback optimizations requires the presence of a pgfi out profile feedback trace file in the current working directory See Mpfi Mprefetch option option enables generation of prefetch instructions on processors where they Mnoprefetch Mr8 Mnors Mr8sintrinsics Command Line Options are supported Possible values for option include d m set the fetch ahead distance for prefetch instructions to m cache lines n p set the maximum number of prefetch instructions to generate for a given loop to p nta use the prefetchnta instruction plain use the prefetch instruction default t0 use the prefetchtO instruction w use the AMD specific prefetchw instruction Disables generation of prefetch instructions pgf77 pgf95 and pghpf only the compiler promotes REAL variables and constants to DOUBL
289. on by the C C caller A Fortran COMPLEX function returns its value in memory the first argument passed to the function must contain the address of the storage for this value A Fortran CHARACTER function adds two arguments to the beginning of its argument list The following example of calling a Fortran CHARACTER function from C C illustrates these caller provided extra parameters CHARACTER FUNCTION CHF C1 I CHARACTER C1 INTEGER I END extern void chf_ char tmp 10 char c1 9 int i chf tmp 10 cl Si 9 The extra parameters tmp and 10 are supplied for the return value while 9 is supplied as the length of cl Refer to Section 2 8 Argument Passing and Return Conventions for additional information A A Win64 Programming Model This section defines compiler and assembly language conventions for the use of certain aspects of an x64 processor running a Win64 operating system These standards must be followed to guarantee that compilers application programs and operating systems written by different people and organizations will work together The conventions supported by the PGCC ANSI C compiler implement the application binary interface ABI as defined in the AMD64 Software Conventions document A 3 1 Function Calling Sequence This section describes the standard function calling sequence including the stack frame register usage and parameter passing Register Usage Conventions The following tabl
290. onconst_ref_anachronism is enabled a reference to a nonconst class can be bound to a class rvalue of the same type or a derived type thereof struct A A int A operator A amp A operator const A amp main A b 1 b A 1 A 2 Allowed as anachronism C 2 New Language Features Accepted The following features not in the ARM but in the X3J16 WG21 Working paper are accepted e The dependent statement of an if while do while or for is considered to be a scope and the restriction on having such a dependent statement be a declaration is removed e The expression tested in an if while do while or for as the first operand of a operator or as an operand of the EST or operators may have a pointer to member type or a class type that can be converted to a pointer to member type in addition to the scalar cases permitted by the ARM e Qualified names are allowed in elaborated type specifiers Use of a global scope qualifier in member references of the form x A B and p gt A B e The precedence of the third operand of the operator is changed e If control reaches the end of the main routine and main has an integral return type it is treated as if a return 0 statement were executed e Pointers to arrays with unknown bounds as parameter types are diagnosed as errors e A functional notation cast of the form AQ can be used even if A is a class without a nontrivial constructor The tempo
291. one command o pgcc Mipa fast o a out filel c file2 c file3 c Just like compiling without Mipa the driver executes several steps to produce the assembly and object files to create the executable 32 Optimization amp Parallelization d i pgcc Mipa fast S o filel s filel c as o filel o filel s pgcc Mipa fast S o file2 s file2 c as o file2 o file2 s pgcc Mipa fast S o file3 s file3 c as o file3 o file3 s pgcc Mipa fast o a out filel o file2 o file3 o d d dp dp d d In the last step an IPA linker is invoked to read all the IPA summary information and perform the interprocedural propagation The IPA linker reinvokes the compiler on each of the object files to recompile them with interprocedural information This creates three new objects with mangled names filel ipa5 a out o file2 ipa5 a out o file2_ipa5 a out o The system linker is then invoked to link these IPA optimized objects into the final executable Later if one of the three source files is edited the executable can be rebuilt with the same command line o pgcc Mipa fast o a out filel c file2 c file3 c This will work but again has the side effect of compiling each source file and recompiling each object file at link time 2 7 6 Building a Program with IPA Several Steps Just by adding the Mipa command line switch it is possible to use individual pgcc commands to compile each source file followed by a command to l
292. ons to perform scalar floating point arithmetic this option is valid only on tp p7 k8 32 k8 64 targets Mnoscalarsse Do not use SSE SSE2 instructions to perform scalar floating point arithmetic use x87 instructions instead this option is not valid in combination with the tp k8 64 option Msmart instructs the compiler driver to invoke an AMD64 specific post pass assembly optimization utility Mnosmart instructs the compiler not to invoke an AMD64 specific post pass assembly optimization utility Munro11 option option invokes the loop unroller This also sets the optimization level to 2 if the level is set to less than 2 The option is one of the following 76 Command Line Options Mnounroll cm instructs the compiler to completely unroll loops with a constant loop count less than or equal to m a supplied constant If this value is not supplied the m count is set to 4 nu instructs the compiler to unroll u times a loop that is not completely unrolled or has a non constant loop count If u is not supplied the unroller computes the number of times a candidate loop is unrolled instructs the compiler not to unroll loops M no vect option option disable enable the code vectorizer where option is one of the Command Line Options following altcode noaltcode assoc noassoc cachesize n no sizelimit Instructs the vectorizer to generate alternate code altcode for vectori
293. onstants as float data types instead of double data types This option can improve the performance of single precision code Mschar specifies signed char characters The compiler treats plain char declarations as signed char Msingle do not to convert float parameters to double parameters in non prototyped functions This option can result in faster code if your program uses only float parameters However since ANSI C specifies that routines must convert float parameters to double parameters in non prototyped functions this option results in non ANSI conformant code Mnosingle instructs the compiler to convert float parameters to double parameters in non prototyped functions Muchar instructs the compiler to treat plain char declarations as unsigned char Default For arguments that you do not specify the defaults are as follows noasmkeyword nosingle dollar schar Usage In this example the compiler allows the asm keyword in the source file pgcc Masmkeyword myprog c In the following example the compiler maps the dollar sign to the dot character Command Line Options 69 pgcc Mdollar myprog c In the following example the compiler treats floating point constants as float values pgcc Mfcon myprog c In the following example the compiler does not convert float parameters to double parameters pgcc Msingle myprog c Without Muchar or with Mschar the variable ch is a signed charac
294. opagate the program will work correctly but without the benefit of IPA I382 IPA subprograms analyzed Interprocedural analysis succeeded in finding and analyzing this many subprograms in the whole program I383 IPA dummy arguments replaced by constants Interprocedural analysis has found this many dummy arguments in the whole program that can be replaced by constants I384 IPA dummy arguments changed to INTENT IN Interprocedural analysis has found this many dummy arguments in the whole program that are not modified and can be declared as INTENT N W385 IPA INTENT IN dummy arguments should be INTENT INOUT Interprocedural analysis has found this many dummy arguments in the whole program that were declared as INTENT IN but should be INTENT INOUT I386 IPA array alignments propagated Interprocedural analysis has found this many array dummy arguments that could have the inherited array alignment replaced by a descriptive alignment I387 IPA array alignments verified Interprocedural analysis has verified that the prescriptive or descriptive alignments of this many array dummy arguments match the alignments of the actual argument I388 IPA array distribution formats propagated 282 Messages Interprocedural analysis has found this many array dummy arguments that could have the transcriptive distribution format replaced by a descriptive format I389 IPA array distribution formats verified Interpro
295. optimized for the case when the data referenced in the loop does not all fit in level 2 cache For a vectorized loop where iteration peeling is performed by default if possible generate an alternate vectorized loop without iteration peeling to be executed if the loop count is less than or equal to n If nis omitted or n is 1 the compiler determines a suitable value of n for each loop and in some cases it may decide not to generate an alternate unpeeled loop For each vectorized loop generate an alternate scalar loop to be executed if the loop count is less than or equal to n If n is omitted or nis 1 the compiler determines a suitable value of n for each loop cpgi noaltcodeThis directive sets the loop count thresholds for parallelization of all innermost loops to 0 and disables alternate code generation for vectorized loops Optimization Directives and Pragmas assoc noassoc This directive toggles the effects of the Mvect noassoc command line option an Optimization M control By default when scalar reductions are present the vectorizer may change the order of operations so that it can generate better code e g dot product Such transformations change the result of the computation due to roundoff error The noassoc directive disables these transformations This directive affects the compiler only when Mvect is enabled on the command line bounds nobounds This directive alters the effects of the Mbounds command lin
296. options is important relative to l options supplied Syntax Ldirectory Where directory is the name of the library directory Default Search the standard library directory Usage In the following example the library directory is lib and the linker links in the standard libraries required by PGF95 from lib pgf95 L lib myprog f In the following example the library directory lib is searched for the library file libx a and both the directories lib and libz are searched for liby a pgf95 L lib 1x L libz ly myprog f Command Line Options 53 l lt library gt Loads a library The linker searches lt library gt in addition to the standard libraries Libraries specified with Jare searched in order of appearance and before the standard libraries Syntax llibrary Where library is the name of the library to search The compiler prepends the characters lib to the library name and adds the a extension following the library name Usage In the following example if the standard library directory is lib the linker loads the library lib libmylib a in addition to the standard libraries pgf95 myprog f lmylib M lt pgflag gt Selects options for code generation The options are divided into the following categories Code generation Environment Inlining Fortran Language Controls C C Language Controls Optimization Miscellaneous The following table lists and briefly describes the options alphabe
297. options outlined in the table above These options are grouped according the category that appears in column 3 of the table above and are listed with exact syntax defaults and notes concerning similar or related options For the latest information and description of a given option or to see all available options use the help command line option to any of the PGI compilers M lt pgflag gt Syntax Mdaz Mnodaz Mdwarfl1 Mdwarf2 Command Line Options Code Generation Controls Set IEEE denormalized input values to zero there is a performance benefit but misleading results can occur such as when dividing a small normalized number by a denormalized number This option must be set for the main program to take effect Do not treat denormalized numbers as zero This option must be set for the main program to take effect Generate DWARF format debug information must be used in combination with g 61 Mdwar f3 Mflushz Mnoflushz Mfunc32 Mlarge_ arrays Mnolarge_arrays Mnomain M no movnt Mprof option option 62 Generate DWARF format debug information must be used in combination with g Generate DWARF3 format debug information must be used in combination with g Set SSE flush to zero mode if a floating point underflow occurs the value is set to zero This option must be set for the main program to take effect Do not set SSE flush to zero mode generate underfl
298. or C and C cccccesseeeeeeeeeeeeeeeseeeeeeeeeeeneess 125 Parallelization PraQmas oiis aaa a a 125 omp parallel eranmi a a a eae a a arash eeieeie 126 omp crica serai ee a ti ea E N need lbiane 128 ell MEET 129 lu Rule EE 130 OMP Stee 130 omp Ne 132 omp parallel te 133 lu Be Le 133 omp Parallel SEUNS iriiritia rene detente dua invaapcevcentephasdevvenctaweanneynee 134 OMp ordered EE 135 eil e un E 135 Omp IUSR asioi utei t Sac ce a a ce A ee Ae ge 136 OMPp threadprivate aerea eeni aae a R eA a aA E A E 136 Run time Library Routines sesessssersssserrsssrirrrssritrrsssinrrnsrrnnnsnsrennas 136 Environment Variables A 139 Optimization Directives and PragmaSs csssseecesssseeeneeeeeees 141 Adding Directives to Fortran ceeeeecceceeeeeeeeceeeeeeeeeaeeeeeeeeaeeeeeeeaaeeeeeeeaas 141 Fortran Directive Summary cccccceeseeceeeeeeeeeeeeeeeeeeeaeeseeeeeetiaeeneeeend 142 Scope of Directives and Command Line options s s s 148 IDEC directive extensions c cccccceeccceeeceeeeeeeeeeeeeeseeeeeeeeeeeeeeaeeeneeeaaes 150 Adding Pragmas to C and C oo ieieeeceeccesseeeneeeseeeeeeeeeeaeesenesneaeeeaaes 151 C C Pragma Summary ccceccccececeeeeeeeeeeceaeeeeeaeeeseaeeeeaeeessneeeesnaeeeeaes 152 Scope of C C Pragmas and Command Line Options e 155 Prefetch DIreCtives Zeg ic Secvandece pesssncdenensied ccveesancedeuessauevebenduaiddevbsuatees 157 Libraries and Environment Variables sss
299. ore the parameters can be used The va_arg macro must be used to access the successive parameters For unprototyped functions or functions that use varargs floating point arguments passed in registers must be passed in both an XMM register and its corresponding general purpose register C Parameter Conversion In C for a called prototyped function the parameter type in the called function must match the argument type in the calling function If the called function is not prototyped the calling convention uses the types of the arguments but promotes char or short to int and unsigned char or unsigned short to unsigned int and promotes float to double unless you use the Msingle option For more information on the Msingle option refer to Chapter 3 If the called function is prototyped the unused bits of a register containing a char or short parameter are undefined and the called function must extend the sign of the unused bits when needed Calling Assembly Language Programs Example A 5 C Program Calling an Assembly language Routine File testmain c main long 1 paral 0x3 800000 float para2 1 0 double d_para3 0 5 float f return extern float sum_3 long paral float para2 double para3 f return sum_3 1 paral f _para2 d_para3 j printf Parameter one type long 08x n 1 paral printf Parameter two type float f n f para2 printf Parameter three type double Sonn d_para3 printf
300. orresponding scalar_logical_expression evaluates to TRUE Otherwise the code within the region will be executed by a single processor regardless of the value of the environment variable OMP_NUM_THREADS If the NUM_THREADS clause is present the corresponding scalar_integer_expression must evaluate to a positive integer value This value sets the maximum number of threads used during execution of the parallel region A NUM_THREADS clause overrides either a previous call to the library routine omp_set_num_threads or the setting of the OMP_NUM_THREADS environment variable 5 3 CRITICAL END CRITICAL The OpenMP END CRITICAL directive uses the following syntax SOMP CRITICAL name lt Fortran code executed in body of critical section gt SOMP END CRITICAL name Within a parallel region you may have code that will not execute properly when multiple threads act upon the same sub region of code This is often due to a shared variable that is written and then read again The CRITICAL END CRITICAL directive pair defines a subsection of code within a parallel region referred to as a critical section which will be executed one thread at a time The optional name argument identifies the critical section The first thread to arrive at a critical section will be the first to execute the code within the section The second thread to arrive will not begin execution of statements in the critical section until the first thread has exi
301. orted by the PGI Fortran Compilers Convention Default STDCALL C UNIX Case of symbol name Upper Lower Lower Lower Leading underscore Yes Yes Yes Yes Trailing underscore No No No Yes Argument byte count added Yes Yes No No Arguments passed by refer Yes No No Yes ence Character argument byte After each No No End of argu counts passed char argument ment list Character strings truncated No Yes Yes No to first character and passed by value varargs support No No Yes No Caller cleans stack No No Yes Yes Except arrays which are always passed by reference even in the STDCALL and C conventions NOTE While it is compatible with the Fortran implementations of Microsoft and several other vendors the C calling convention supported by the PGI Fortran compilers for Windows is not strictly compatible with the C calling convention used by most C C compilers In particular symbol names produced by PGI Fortran compilers using the C convention are all lower case The standard C convention is to preserve mixed case symbol names You can cause any of the PGI Fortran compilers to preserve mixed case symbol names using the Mupcase option but be aware that this could have other ramifications on your program 10 14 2 Symbol Name Construction and Calling Example This section presents an example of the rules outlined in table 10 3 In the pseudocode used below addr refers to the address of a
302. ory clusters of workstations or clusters of SMP workstations or servers Coding a data parallel version of an application can be more work than using OpenMP directives but has the advantage that the resulting executable is usable on all types of parallel systems regardless of whether shared memory is available See the PGHPF User s Guide for a complete description of how to build and execute data parallel HPF programs In this manual the first two types of parallel programs are collectively referred to as SMP parallel programs The third type is referred to as a data parallel program or simply as an HPF program Some newer CPUs incorporate two or more complete processor cores functional units registers level 1 cache level 2 cache etc on a single silicon die These are referred to as multi core processors For purposes of HPF threads or OpenMP parallelism these cores function as 2 or more distinct processors However the processing cores are on a single chip occupying a single socket on a system motherboard For purposes of PGI software licensing a multi core processor is treated as a single CPU 1 4 1 Running SMP Parallel Programs When you execute an SMP parallel program by default it will use only processor To run on more than one processor set the NCPUS environment variable to the desired number of processors subject to a maximum of 4 for PGI s workstation class products You can set this environment variable by issuing t
303. osition Contents Frame 4n 8 ebp argument word n previous 8 ebp argument word 0 4 ebp return address 0 ebp caller s ebp current 4 ebp n bytes of local n ebp variables and temps Several key points concerning the stack frame e The stack is kept double word aligned e Argument words are pushed onto the stack in reverse order i e the rightmost argument in C call syntax has the highest address A dummy word may be pushed ahead of the rightmost argument in order to preserve doubleword alignment All incoming arguments appear on the stack residing in the stack frame of the caller e An argument size is increased if necessary to make it a multiple of words This may require tail padding depending on the size of the argument All registers on an x86 system are global and thus visible to both a calling and a called function Registers ebp ebx edi Gest and esp are non volatile across function calls Therefore a function must preserve these registers values for its caller Remaining registers are volatile scratch If a calling function wants to preserve such a register value across a function call it must save its value explicitly Some registers have assigned roles in the standard calling sequence Joesp The stack pointer holds the limit of the current stack frame which is the address of the stack s bottom most valid word At all times the stack pointer should point to a word aligned area
304. ot propagate distribution target from to The most common cause is when passing an array with a transcriptive distribution target to a dummy argument with prescriptive or descriptive distribution target I375 Distribution format mismatch between and Usually this arises when the actual and dummy arguments are distributed in different dimensions I376 Alignment stride mismatch between and This may arise when the actual argument has a different stride in its alignment to its template than does the dummy argument I377 Alignment offset mismatch between and This may arise when the actual argument has a different offset in its alignment to its template than does the dummy argument I378 Distribution target mismatch between and Messages 281 This may arise when the actual and dummy arguments have different distribution target sizes I379 Alignment of is too complex The alignment specification of the array is too complex for interprocedural analysis to verify or propagate the program will work correctly but without the benefit of IPA I380 Distribution format of is too complex The distribution format specification of the array is too complex for interprocedural analysis to verify or propagate the program will work correctly but without the benefit of IPA I381 Distribution target of is too complex The distribution target specification of the array is too complex for interprocedural analysis to verify or pr
305. ou know that the scalar is assigned on the last iteration of the loop making it safe to parallelize the loop a pragma is available to let the compiler know the loop is safe to parallelize Use the following C pragma to tell the compiler that for a given loop the last value computed for all scalars make it safe to parallelize the loop cpgi l safe _lastval In addition a command line option Msafe_lastval provides this information for all loops within the routines being compiled essentially providing global scope Optimization Directives and Pragmas 147 unroll nounroll The directive nounroll is used to disable loop unrolling and unroll to enable unrolling The directive takes arguments c and n A c specifies that c complete unrolling should be turned on or off An n specifies that n count unrolling should be turned on or off In addition the following arguments may be added to the unroll directive cpgi unroll c v This sets the threshold to which c unrolling applies v is a constant a loop whose constant loop count is lt v is completely unrolled cpgi unroll n v This adjusts threshold to which n unrolling applies v is a constant a loop to which n unrolling applies is unrolled v times The directives unroll and nounroll only apply if Munroll is selected on the command line vector novector The directive novector is used to disable vectorization The directive vector is used to re enable vectorization after
306. ows This option must be set for the main program to take effect Align functions on 32 byte boundaries Enable support for 64 bit indexing and single static data objects larger than 2GB in size This option is default in the presence of mcmodel medium Can be used separately together with the default small memory model for certain 64 bit applications that manage their own memory space Disable support for 64 bit indexing and single static data objects larger than 2GB in size When placed after mcmodel medium on the command line disables use of 64 bit indexing for applications that have no single data object larger than 2GB instructs the compiler not to include the object file that calls the Fortran main program as part of the link step This option is useful for linking programs in which the main program is written in C C and one or more subroutines are written in Fortran pgf77 pgf95 and pghpf only instructs the compiler to generate nontemporal move and prefetch instructions even in cases where the compiler cannot determine statically at compile time that these instructions will be beneficial Set profile options option can be any of the following dwarf generate limited DWARF information to enable source correlation by 3rd party profiling tools func perform PGI style function level profiling hwcts Use PAPI based profiling with hardware counters linux86 64 platforms only lines perform PGI style line level profi
307. p If PGI_CONTINUE is set to verbose the stack will be automatically cleaned up and execution will continue after printing of a warning message instructs the compiler to check for pointers that are de referenced while initialized to NULL pgf95 and pghpf only instructs the compiler to check the stack for available space in the prologue of a function and before the start of a parallel region Prints a warning message and aborts the program gracefully if stack space is insufficient Useful when many local and private variables are declared in an OpenMP program Mcpp option option Command Line Options 79 Md11 Mgccbug s run the PGI cpp like pre processor without execution of any subsequent compilation steps This option is useful for generating dependence information to be included in makefiles option is one or more of the following Note only one of the m md mm or mmd options can be present if multiple of these options are listed the last one listed is accepted and the others are ignored m print makefile dependencies to stdout md print makefile dependencies to filename d where filename is the root name of the input file being processed mm print makefile dependencies to stdout ignoring system include files mmd print makefile dependencies to filename d where filename is the root name of the input file being processed ignoring system include files no comment don t retain comments in ed output
308. parallelized program with NCPUS or OMP_NUM_THREADS set to a value larger than the number of physical processors in the system For example if you produce a parallelized executable a out and execute as follows on a system with only one processor d i setenv NCPUS 2 a out Warning OMP NUM THREADS or NCPUS 2 greater than available cpus 1 FORTRAN STOP d i Setting MP_WARN to no will eliminate these warning messages NCPUS The NCPUS environment variable can be used to set the number of processes or threads used in parallel regions The default is to use only one process or thread serial mode If both OMP_NUM_THREADS and NCPUS are set the value of OMP_NUM_THREADS takes precedence Warning setting NCPUS to a value larger than the number of physical processors or cores in your system can cause parallel programs to run very slowly NCPUS_MAX The NCPUS_MAX environment variable can be used to limit the maximum number of processes or threads used in a parallel program Attempts to dynamically set the number of processes or threads to a higher value for example using set_omp_num_threads will cause the number of processes or threads to be set at the value of NCPUS_MAX lt rather than the value specified in the function call NO_STOP_MESSAGE If this variable exists the execution of a plain STOP statement does not produce the message FORTRAN STOP The default behavior of the PGI Fortran compilers is to issue this message PGI
309. pgCC only Enable compilation of C with cfront 2 1 compatibility This causes the compiler to accept language constructs that while not part of the C language definition are accepted by the AT amp T C Language System cfront release 2 1 This option also enables acceptance of anachronisms Default The compiler does not accept cfront language constructs that are not part of the C language definition Usage In the following example the compiler accepts cfront constructs pgCC b myprog cc Cross references cfront2 1 b3 cfront3 0 p A Command Line Options 95 b3 pgCC only Enable compilation of C with cfront 3 0 compatibility This causes the compiler to accept language constructs that while not part of the C language definition are accepted by the AT amp T C Language System cfront release 3 0 This option also enables acceptance of anachronisms Default The compiler does not accept cfront language constructs that are not part of the C language definition Usage In the following example the compiler accepts cfront constructs pgCC b3 myprog cc Cross references cfront2 1 b cfront3 0 p A no_ bool pgCC only Enable or disable recognition of bool The default value is bool cfront_2 1 pgCC only Enable compilation of C with cfront 2 1 compatibility This causes the compiler to accept language constructs that while not part of the C language definition a
310. pgi g novector integer maxtime time parameter n 1000 maxtime 10 double precision a n n b n n c n n do time 1 maxtime do i 1 n dogi yn Gite A Ae BCL yy enddo enddo enddo end In this version the compiler disables vectorization for the entire source file Another use of the directive scoping mechanism turns an option on or off locally either for a specific procedure or for a specific loop integer maxtime time parameter n 1000 maxtime 10 double precision a n n b n n c n n cpgi l novector do time 1 maxtime do i 1 n do j 1 n e i j a i j b i j enddo enddo enddo end Loop level scoping does not apply to nested loops That is the directive only applies to the following loop In this example the directive turns off vector transformations for the top level loop If the outer loop were a timing loop this would be a practical use for a loop scoped directive Optimization Directives and Pragmas 149 7 4 DEC directive extensions These extensions are only enabled on Windows platforms Syntax ATTRIBUTES Clause DECS ATTRIBUTES lt attr option gt where lt attr option gt is one of ALIAS alias_name routine_name Specifies an alternative name with which to resolve routine_name C Same as STDCALL on Win64 DLLEXPORT name Specifies that name is being exported to other applications or DLL s DLLIMPORT name Specifies that name is being imported from other
311. pilers save the last value of the scalar However if the loop is parallelized and the scalar is not assigned on every iteration it may be difficult without resorting to costly critical sections to determine on what iteration t is last assigned Analysis allows the compiler to determine that a scalar is assigned on each iteration and hence that the loop is safe to parallelize if the scalar is used later For example for i 1 i lt n i if x i gt 0 0 Eer 20 else 33 20 yli Es GE eer AE where t is assigned on every iteration of the loop However there are cases where a scalar may be privatizable but if it is used after the loop it is unsafe to parallelize Examine this loop for Las La Rz i if x i gt 0 0 Es zl Optimization amp Parallelization 29 yli Vi Cy where each use of t within the loop is reached by a definition from the same iteration Here t is privatizable but the use of t outside the loop may yield incorrect results since the compiler may not be able to detect on which iteration of the parallelized loop t is last assigned The compiler detects the above cases Where a scalar is used after the loop but is not defined on every iteration of the loop parallelization will not occur When the programmer knows that the scalar is assigned on the last iteration of the loop the programmer may use a directive or pragma to let the compiler know the loop is safe to parallelize
312. ple Intel Pentium HI and AMD AthlonXP MP processors scalar floating point arithmetic as generated by the PGI Workstation compilers is performed using x87 floating point stack instructions With the advent of SSE SSE2 instructions on Intel Pentium 4 Xeon and AMD Opteron Athlon64 it is possible to perform all scalar floating point arithmetic using SSE SSE2 instructions In most cases this is beneficial from a performance standpoint The default on 32 bit Intel Pentium I II tp p6 tp piii etc or AMD AthlonXP MP tp k7 is to use x87 instructions for scalar floating point arithmetic The default on Intel Pentium 4 Xeon or Intel EM64T running a 32 bit operating system tp p7 AMD Opteron Athlon64 running a 32 bit operating system tp k8 32 or AMD Opteron Athlon64 or Intel EM64T processors running a 64 bit operating system tp k8 64 and tp p7 64 respectively is to use SSE SSE2 instructions for scalar floating point arithmetic The only way to override this default on AMD Opteron Athlon64 or Intel EM64T processors running a 64 bit operating system is to specify an older 32 bit target for example tp k7 or tp piii 18 Optimization amp Parallelization Note that there can be significant arithmetic differences between calculations performed using x87 instructions versus SSE SSE2 By default all floating point data is promoted to IEEE 80 bit format when stored on the x87 floating point stack and all x87 operations are performed regis
313. r S462 Arguments and return values of ELEMENATAL procedures cannot have the POINTER attribute S463 Arguments of ELEMENATAL procedures cannot be procedures S464 An ELEMENTAL procedure cannot be passed as argument Messages 289 BA Fortran Runtime Error Messages This section presents the error messages generated by the runtime system The runtime system displays error messages on standard output B 4 1 Message Format The messages are numbered but have no severity indicators because they all terminate program execution B 4 2 Message List Here are the runtime error messages 201 illegal value for specifier An improper specifier value has been passed to an I O runtime routine Example within an OPEN statement form unknown 202 conflicting specifiers Conflicting specifiers have been passed to an I O runtime routine Example within an OPEN statement form unformatted blank null 203 record length must be specified A recl specifier required for an I O runtime routine has not been passed Example within an OPEN statement access direct has been passed but the record length has not been specified recl specifier 204 illegal use of a readonly file Self explanatory Check file and directory modes for readonly status 205 SCRATCH and SAVE KEEP both specified In an OPEN statement a file disposition conflict has occurred Example within an OPEN statement status scratch and dispose keep hav
314. r elif directive was found after a else directive only endif is allowed in this context S222 else after else A preprocessor else directive was found after a else directive only endif is allowed in this context 223 if directives too deeply nested Preprocessor if directive nesting exceeded the maximum allowed currently 10 224 Actual parameters too long for The total length of the parameters in a macro call to the indicated macro exceeded the maximum allowed currently 2048 W225 Argument mismatch for The number of arguments supplied in the call to the indicated macro did not agree with the number of parameters in the macro s definition F226 Can t find include file The indicated include file could not be opened 227 Definition too long for The length of the macro definition of the indicated macro exceeded the maximum allowed currently 2048 S228 EOF in comment The end of a file was encountered while processing a comment S229 EOF in macro call to 270 Messages The end of a file was encountered while processing a call to the indicated macro 230 EOF in string The end of a file was encountered while processing a quoted string 231 Formal parameters too long for The total length of the parameters in the definition of the indicated macro exceeded the maximum allowed currently 2048 232 Identifier too long The length of an identifier exceeded the maximum allowed curren
315. r double precision constant 029 Illegal constant Illegal hexadecimal octal or binary constant A hexadecimal constant consists of digits 0 9 and letters A F or a f any other character in a hexadecimal constant is illegal An octal constant consists of digits 0 7 any other digit or character in an octal constant is illegal A binary constant consists of digits 0 or 7 any other digit or character in a binary constant is illegal S030 Explicit shape must be specified for 252 Messages 031 Illegal data type length specifier for The data type length specifier e g 4 in INTEGER 4 is not a constant expression that is a member of the set of allowed values for this particular data type W032 Data type length specifier not allowed for The data type length specifier e g 4 in INTEGER 4 is not allowed in the given syntax e g DIMENSION A 10 4 033 Illegal use of constant A constant was used in an illegal context such as on the left side of an assignment statement or as the target of a data initialization statement 034 Syntax error at or near I035 Predefined intrinsic loses intrinsic property An intrinsic name was used in a manner inconsistent with the language definition for that intrinsic The compiler based on the context will treat the name as a variable or an external function 036 Illegal implicit character range First character must alphabetically precede second 037 Contradictory da
316. rallel sections 134 SOCON Sinin arooda dee here ess 133 SINQIO sissies ids eek e starea ae aain 130 threadprivate csecccesecceeeseeeeeseeeeeeeeenes 136 OpenMP C C Support Routines omp destroy Joch 138 omp_get read num 137 omp get dvnamiet 138 omp_get_max_threads eene 137 OMp_Get_nested eeeeeeeeeeeeseeeeeneees 138 OMp_Get_NUM_PIOCS neee 137 omp_get_num_threads ssssseeesseeeeen 136 omp_get_wtick eseese 138 OMP_JEt_WHiME ieee 138 OMpP_IN_parallel 0 eeseeeeseeeeeeeeeeneeees 137 omp mit Joch 138 omp_set_dynamic e er 137 omp set oeh gege dee wate 138 omp_set_nested eeen 138 omp_set_num_threads eee 137 omp test lckin 139 OMp_UNSEet_lOCK eeen 139 OpenMP environment variables MES eiae iunie iharap 124 140 171 OMP D YNAMIC 124 139 OMP NESTED Zo nisrarnerermei inian 124 139 OMP NUM THREADS 124 139 OMP GCHEDULE cc eeeeeeeeeeeeeeereees 124 OpenMP Fortran Directives ceeeeeee 107 ATOMIC 8 eege SEENEN 120 BARRIER E 116 GRITIGAL inenen ae i nen teres 111 eer 121 ET ON geht mniscacncscote eels Seat NA oe cet 25 MAER be tbe cues ee oat exe 112 auto parallelization EH 25 ORDERED E EEEE 120 failed auto parallelization EE 27 81 PARAL E eii c eeu ees a ae en ee 108 Mconcur auto parallelization 8 Ze EN PARALMEEE DO ng ae eebe 117 NCPUS environment variable Ot PARALLEL SECTIONS nenne 119 Sale Mast al naea te rae ee eee ate 2
317. rams or to selected loops in the source file for example an optimization Use directives to tune selected routines or loops 7 1 Adding Directives to Fortran Directives may have any of the following forms cpgi g directive cpgi r directive cpgi l directive cpgi directive The C must be in column 1 Either or is allowed in place of C The scope indicator occurs after the this indicator controls the scope of the directive Some directives ignore the scope indicator The valid scopes as shown above are g global indicates the directive applies to the end of the source file T routine indicates the directive applies to the next subprogram l loop indicates the directive applies to the next loop but not to any loop contained within the loop body Loop scoped directives are only applied to DO loops blank indicates that the default scope for the directive is applied The body of the directive may immediately follow the scope indicator Alternatively any number of blanks may precede the name of the directive Any names in the body of the directive including the directive name may not contain embedded blanks Blanks may surround any special characters such as a comma or an equal sign The directive name including the directive prefix may contain upper or lower case letters case is not significant Case is significant for any variable names that appear in the body of the directive if the command line option Mupcase is
318. ran Version 4 0 Digital Equipment Corporation September 1984 e IBM VS Fortran IBM Corporation Rev GC26 4119 Military Standard Fortran DOD Supplement to American National Standard Programming Language Fortran ANSI x 3 1978 MIL STD 1753 November 9 1978 e American National Standard Programming Language C ANSI X3 159 1989 ISO IEC 9899 1999 Information technology Programming Languages C Geneva 1999 C99 Organization This manual is divided into the following chapters and appendices Chapter 1 Getting Started provides an introduction to the PGI compilers and describes their use and overall features Chapter 2 Optimization amp Parallelization describes standard optimization techniques that with little effort allow users to significantly improve the performance of programs Chapter 3 Command Line Options provides a detailed description of each command line option Chapter 4 Function Inlining describes how to use function inlining and shows how to create an inline library Chapter 5 OpenMP Directives for Fortran provides a description of the OpenMP Fortran parallelization directives and shows examples of their use Chapter 6 OpenMP Pragmas for C and C provides a description of the OpenMP C and C parallelization pragmas and shows examples of their use Chapter 7 Optimization Directives and Pragmas provides a description of each Fortran optimization dire
319. ranscriptive distribution formats replaced Messages 277 1333 1334 I335 1336 1337 1338 339 340 341 342 343 344 IPA IPA IPA IPA IPA IPA transcriptive distribution targets replaced descriptive prescriptive array alignments verified descriptive prescriptive distribution formats verified descriptive prescriptive distribution targets verified common blocks optimized common blocks not optimized Bad IPA contents file Bad IPA file format Unable Unable Unable Unable to to to to create file while analyzing IPA information open file while analyzing IPA information open IPA contents file create file while collecting IPA information F345 Internal error in table overflow Analysis failed due to a table overflowing its maximum size 278 Messages W346 Subprogram appears twice The subprogram appears twice in the same source file IPA will ignore the first appearance F347 Missing ipalib option Interprocedural analysis enabled with the ipacollect ipaanalyze or ipapropagate options requires the ipalib option to specify the library directory W348 Common has different distribution target The array was declared in a common block with a different distribution target in another subprogram W349 Common has different distribution format The array was declared in a common block with a diff
320. rary created gets the same default initialization to zero as a static object of the class type e A cast can be used to select one out of a set of overloaded functions when taking the address of a function e Template friend declarations and definitions are permitted in class definitions and class template definitions e Type template parameters are permitted to have default arguments 296 C Dialect Supported Function templates may have nontype template parameters A reference to const volatile cannot be bound to an rvalue Qualification conversions such as conversion from T to T const const are allowed Digraphs are recognized Operator keywords e g and bitand etc are recognized Static data member declarations can be used to declare member constants wchar_t is recognized as a keyword and a distinct type bool is recognized RTTI runtime type identification including dynamic_cast and the typeid operator are implemented Declarations in tested conditions in if switch for and while statements are supported Array new and delete are implemented New style casts static_cast reinterpret_cast and const_cast are implemented Definition of a nested class outside its enclosing class is allowed mutable is accepted on nonstatic data member declarations Namespaces are implemented including using declarations and directives Access declarations are broadened to match the corresponding using declarations Exp
321. re accepted by the AT amp T C Language System cfront release 2 1 This option also enables acceptance of anachronisms Default The compiler does not accept cfront language constructs that are not part of the C language definition Usage In the following example the compiler accepts cfront constructs pgCC cfront_2 1 myprog cc Cross references b b3 cfront3 0 p A cfront_3 0 pgCC only Enable compilation of C with cfront 3 0 compatibility This causes the compiler to accept language constructs that while not part of the C language definition are accepted by the AT amp T C Language System cfront release 3 0 This option also enables acceptance of anachronisms Default The compiler does not accept cfront language constructs that are not part of the C language definition Usage In the following example the compiler accepts cfront constructs 96 Command Line Options pge cfront_3 0 myprog cc Cross references cfront2 1 b b3 p A create_pch filename pgCC only If other conditions are satisfied create a precompiled header file with the specified name If pch automatic PCH mode appears on the command line following this option its effect is erased diag_suppress tag pgCC only Override the normal error severity of the specified diagnostic messages The message s may be specified using a mnemonic error tag or using an error number diag_remark tag pgCC
322. re is no command line option corresponding to this directive Normally the compiler removes certain invariant if constructs from within a loop and places them outside of the loop The directive noinvarif directs the compiler to not move such constructs The directive invarif toggles a previous noinvarif ivdep The ivdep directive is equivalent to the directive nodepchk opt The syntax of this directive is cpgi lt scope gt opt lt level gt where the optional lt scope gt is r or g and lt level gt is an integer constant representing the optimization level to be used when compiling a subprogram routine scope or all subprograms in a file global scope The opt directive overrides the value specified by the command line option On Istval nolstval There is no command line option corresponding to this directive The compiler determines whether the last values for loop iteration control variables and promoted scalars need to be computed In certain cases the compiler must assume that the last values of these variables are needed and therefore computes their last values The directive nolstval directs the compiler not to compute the last values for those cases safe_lastval During parallelization scalars within loops need to be privatized Problems are possible if a scalar is accessed outside the loop For example do i 1 N if x i gt 5 0 t x i enddo ME tS le creates a problem since the value of t may not b
323. reported to trs pgroup com F001 Source input file name not specified On the command line source file name should be specified either before all the switches or after them F002 Unable to open source input file Source file name misspelled file not in current working directory or file is read protected F003 Unable to open listing file Probably user does not have write permission for the current working directory F004 Generic message for file errors F005 Unable to open temporary file Compiler uses directory usr tmp or tmp in which to create temporary files If neither of these directories is available on the node on which the compiler is being used this error will occur s006 Input file empty Source input file does not contain any Fortran statements other than comments or compiler directives F007 Subprogram too large to compile at this optimization level Internal compiler data structure overflow working storage exhausted or some other non recoverable problem related to the size of the subprogram If this error occurs at opt 2 reducing the opt level to 1 may work around the problem Moving the subprogram being compiled to its own source file may eliminate the problem If this error occurs while compiling a subprogram of fewer than 2000 statements it should be reported to the compiler maintenance group as a possible compiler problem F008 Error limit exceeded The compiler gives up because too many se
324. ross a function call it must save its value explicitly Registers are used extensively in the standard calling sequence The first six integer and pointer arguments are passed in these registers listed in order rdi rsi rdx Yorcx r8 r9 The first eight floating point arguments are passed in the first eight XMM registers xmm0 xmm_1 xmm7 The registers rax and rdx are used to return integer and pointer values The registers xmm0 and xmm1 are used to return floating point values Additional registers with assigned roles in the standard calling sequence 220 Run time Environment rsp rbp RFLAGS The stack pointer holds the limit of the current stack frame which is the address of the stack s bottom most valid word The stack must be 16 byte aligned The frame pointer holds a base address for the current stack frame Consequently a function has registers pointing to both ends of its frame Incoming arguments reside in the previous frame referenced as positive offsets from rbp while local variables reside in the current frame referenced as negative offsets from rbp A function must preserve this register value for its caller The flags register contains the system flags such as the direction flag and the carry flag The direction flag must be set to the forward i e zero direction before entry and upon exit from a function Other user flags have no specified role in the standard calling sequence
325. ructure W137 Field namelist is required in nested structures W138 Multiply defined STRUCTURE member name A member name was used more than once within a structure W139 Structure in RECORD statement not defined A RECORD statement contains a reference to a STRUCTURE that has not yet been defined S140 Variable is not a RECORD 141 RECORD required on left of 142 is not a member of this RECORD 143 requires initializer W144 NEED ERROR MESSAGE This is used as a temporary message for compiler development W145 FILL only valid within STRUCTURE block The FILL special name was used outside of a STRUCTURE multiline statement It is only valid when used within a STRUCTURE multiline statement even though it is ignored 146 Expression must be character type 147 Character expression not allowed in this context 148 Reference to required Messages 263 An aggregate reference to a record was expected during statement compilation but another data type was found instead 149 Record where arithmetic value required An aggregate record reference was encountered when an arithmetic expression was expected 150 Structure Record derived type or member not allowed in this context A structure record or member reference was found in a context which is not supported For example the use of structures records or members within a data statement is disallowed S151 Empty TYPE STRUCTURE UNION or MAP T
326. s inform specifies the minimum level of error severity that Miscellaneous the compiler displays inline invokes the function inliner Inlining no ipa invokes inter procedural analysis and optimiza Optimization tion no iomutex determines whether critical sections are generated Fortran Language no large_arrays enable support for 64 bit indexing and single static data objects of size larger than 2GB Code Generation only lfs link in libraries that allow file T O to files of size Environment larger than 2GB on 32 bit systems 32 bit Linux only no lre Disable enable loop carried redundancy elimina Optimization tion keepasm instructs the compiler to keep the assembly file Miscellaneous nolist specifies whether the compiler creates a listing Miscellaneous file makedll Generate a dynamic link library DLL Windows Miscellaneous Command Line Options 57 pgflag Description Category no movnt disable force generation of non temporal moves Code Generation and prefetching neginfo instructs the compiler to produce information on Miscellaneous why certain optimizations are not performed noframe eliminates operations that set up a true stack Optimization frame pointer for functions nomain when the link step is called don t include the Code Generation object file that calls the Fortran main program pgf77 pgf95
327. s corresponding to each source file and subsequently to link the object files together into a single executable file Thus the command above is roughly equivalent to the following commands performed individually 30 Optimization amp Parallelization d i pgcc S o filel s filel c as o filel o filel s pgcc S o file2 s file2 c as o file2 o file2 s pgcc S o file3 s file3 c as o file3 o file3 s pgcc o a out filel o file2 o file3 o A d dp dp d d If any of the three source files is edited the executable can be rebuilt with the same command line pgcc o a out filel c file2 c file3 c This always works as intended but has the side effect of recompiling all of the source files even if only one has changed For applications with a large number of source files this can be time consuming and inefficient 2 7 2 Building a Program Without IPA Several Steps It is also possible to use individual pgcc commands to compile each source file into a corresponding object file and one to link the resulting object files into an executable pgce c filel c pgce c file2 c pgcce c file3 c pgcc o a out filel o file2 o file3 o d i d i ol d i The pgcc driver invokes the compiler and assembler as required to process each source file and invokes the linker for the final link command If you modify one of the source files the executable can be rebuilt by compiling just that file and then relinking pgce
328. s too cannot be promoted to vectors 101 Vector operation not allowed on Record and character typed entities may only be referenced as scalar quantities 102 Arithmetic IF expression has wrong data type The parenthetical expression of an arithmetic if statement must be an integer real or double precision scalar expression 103 Type conversion of subscript expression for The data type of a subscript expression must be integer If it is not it is converted Messages 259 104 Illegal control structure This message is issued for a number of errors involving IF THEN statements and DO loops If the line number specified is the last line END statement of the subprogram the error is probably an unterminated DO loop or IF THEN statement S105 Unmatched ELSEIF ELSE or ENDIF statement An ELSEIF ELSE or ENDIF statement cannot be matched with a preceding IF THEN statement 106 DO index variable must be a scalar variable The DO index variable cannot be an array name a subscripted variable a PARAMETER name a function name a structure name etc 107 Illegal assigned goto variable 108 Illegal variable in NAMELIST group A NAMELIST group can only consist of arrays and scalars which are not dummy arguments and pointer based variables I109 Overflow in constant constant truncated at left A non decimal hexadecimal octal or binary constant requiring more than 64 bits produces an overflow The constan
329. s and pragmas Usage The following command line requests processing of any shared memory directives present in myprog f pgf95 mp myprog f Cross reference Mconcur and Mvect O lt level gt Invokes code optimization at the specified level Syntax O level Where level is one of the following 0 creates a basic block for each statement Neither scheduling nor global optimization is done To specify this level supply a 0 zero argument to the O option 1 schedules within basic blocks and performs some register allocations but does no global optimization 84 Command Line Options H performs all level 1 optimizations and also performs global scalar optimizations such as induction variable elimination and loop invariant movement 3 level three specifies aggressive global optimization This level performs all level one and level two optimizations and enables more aggressive hoisting and scalar replacement optimizations that may or may not be profitable Default This table shows the interaction between the O option g option and Mvect options Table 3 4 Optimization and O g Mvect and Mconcur Options Optimize Debug M Optimization Option Option Option Level none none none 1 none none Mvect 2 none none Mconcur 2 none g none 0 O none or g none 2 Olevel none or g none level Olevel lt 2 none or g Mvect 2 Olevel lt 2 none or g
330. s and Unions align according to the alignment of the most restricted data type of the structure or union In the next example the union aligns on a 4 byte boundary since the alignment of c the most restrictive element is four STRUCTURE astr UNION MAP INTEGER 2 a 2 bytes END MAP BYTE b 1 byte END MAP MAP INTEGER 4 c 4 bytes END MAP END UNION END STRUCTURE Structure alignment can result in unused space called padding Padding between members of the structure is called internal padding Padding between the last member and the end of the space is called tail padding The offset of a structure member from the beginning of the structure is a multiple of the member s alignment For example since an INTEGER 2 aligns on a 2 byte boundary the offset of an INTEGER 2 member from the beginning of a structure is a multiple of two bytes Fortran C and C Data Types 179 9 1 3 Fortran 90 Aggregate Data Types Derived Types The Fortran 90 standard added formal support for aggregate data types The TYPE statement begins a derived type data specification or declares variables of a specified user defined type For example the following would define a derived type ATTENDEE TYPE ATTENDEE CHARACTER LEN 30 NAME CHARACTER LEN 30 ORGANIZATION CHARACTER LEN 30 EMAIL END TYPE ATTENDEE In order to declare a variable of type ATTENDEE and access the contents of such a variable code such as the following would be use
331. s are comments in a program that are interpreted by the PGI Fortran compilers when the option mp is specified on the command line The form of a parallelization directive is sentinel directive _name clauses With the exception of the SGI compatible DOACROSS directive the sentinel must be OMP C OMP or OMP must start in column 1 one and must appear as a single word without embedded white space The sentinel marking a DOACROSS directive is C Standard Fortran syntax restrictions line length case insensitivity etc apply to the directive line Initial directive lines must have a space or zero in column six and continuation directive lines must have a character other than space or zero in column six Continuation lines for CSDOACROSS directives are specified using the C amp sentinel 107 The order in which clauses appear in the parallelization directives is not significant Commas separate clauses within the directives but commas are not allowed between the directive name and the first clause Clauses on directives may be repeated as needed subject to the restrictions listed in the description of each clause The compiler option mp enables recognition of the parallelization directives The use of this option also implies Mreentrant local variables are placed on the stack and optimizations that may result in non reentrant code are disabled e g Mnoframe Miomutex critical sections are generated around Fortran I O statem
332. s that are mangled and provides some details on the mangling algorithm For more details refer to The Annotated C Reference Manual 11 2 1 Type Name Mangling Using PGC each type has a corresponding mangled encoding For example a class type is represented as the class name preceded by the number of characters in the class name as in Sabcde for abcde Simple types are encoded as lower case letters as in i for int or f for float Type modifiers and declarators are encoded as upper case letters preceding the types they modify as in U for unsigned or P for pointer 206 C Name Mangling 11 2 2 Nested Class Name Mangling Nested class types are encoded as a Q followed by a digit indicating the depth of nesting followed by a _ followed by the mangled form names of the class types in the fully qualified name of the class from outermost to innermost class A class B Q2_1A1B 1 11 2 3 Local Class Name Mangling The name of the nested class itself is mangled to the form described above with a prefix __ which serves to make the class name distinct from all user names Local class names are encoded as L followed by a number which has no special meaning it s just an identifying number assigned to the class followed by __ followed by the mangled name of the class this is not in the ARM and cfront encodes local class names slightly differently void class A L1_1A This form is used when encoding the local class name
333. same variable is passed in both argument positions or one formal argument and a global or COMMON variable may be aliased because the global or COMMON variable is passed as an actual argument If either alias is assigned in the subroutine unexpected results may occur this message alerts the user that this situation is disallowed by the Fortran standard F415 IPA fails incorrect IPA file Interprocedural analysis saves its information in special IPA files in the specified IPA directory One of these files has been renamed or corrupted This can arise when there are two files with the same prefix such as a hpf and a f90 E416 Argument has the SEQUENCE attribute but the dummy parameter does not When an actual argument is an array with the SEQUENCE attribute the dummy parameter must have the SEQUENCE attribute or an INTERFACE block must be used E417 Interface block for is a SUBROUTINE but should be a FUNCTION Messages 285 E418 E419 W420 W421 E422 E423 E424 E425 E426 E427 E428 E429 286 Interface block for is a FUNCTION but should be a SUBROUTINE Interface block for is a FUNCTION has wrong result type Earlier directive overrides directive directive can only appear in a function or subroutine Nonconstant DIM argument is not supported Constant DIM argument is out of range Equivalence using substring or vector triplets is not allowed A record is not allowed in
334. scriminately can result in much larger code size and no increase in execution speed The PGI compilers provide two categories of inlining e Automatic inlining During the compilation process a hidden pass precedes the compilation pass This hidden pass extracts functions that are candidates for inlining The inlining of functions occurs as the source files are compiled e Inline libraries You create inline libraries for example using the pgf95 command and the Mextract and o options There is no hidden extract pass but you must ensure that any files that depend on the inline library use the latest version of the inline library There are important restrictions on inlining Inlining only applies to certain types of functions Refer to Section 4 5 Restrictions on Inlining at the end of this chapter for more details on function inlining limitations 4 1 Invoking Function Inlining To invoke the function inliner use the Minline option If you do not specify an inline library the compiler performs a special prepass on all source files named on the compiler command line before it compiles any of them This pass extracts functions that meet the requirements for inlining and puts them in a temporary inline library for use by the compilation pass Several Minline options let you determine the selection criteria for functions to be inlined These selection criteria include except func Inline all eligible functions except
335. sed to display informational messages listing why certain optimizations are inhibited The dryrun option can be useful as a diagnostic tool if you need to see the steps used by the compiler driver to pre process compile assemble and link in the presence of a given set of command line inputs When you specify the dryrun option these steps will be printed to stdout but will not actually be performed For example this allows inspection of the default and user specified libraries that are searched during the link phase and the order in which they are searched by the linker The remainder of this chapter describes the O options the loop unroller option Munroll the vectorizer option Mvect the auto parallelization option Mconcur and the inter procedural analysis optimization Mipa and the profile feedback instrumentation Mpfi and optimization Mpfo options Usually you should be able to get very near optimal compiled performance using some combination of these switches The following overview will help if you are just getting started with one of the PGI compilers or wish to experiment with individual optimizations Complete specifications of each of these options are listed in Chapter 3 Command Line Options The chapters that follow provide more detailed information on other M lt pgflag gt options that control specific optimizations including function inlining Explicit parallelization through the use of OpenMP dire
336. ssseeeeeeeeeeeeeeeeeees 159 Using builtin Math Functions in C C ooo eeeeeeeeeeeeeeeeeeneeseneeenaees 159 Creating and Using Shared Object Files on Linux s 159 Creating and Using Dynamic Link Libraries on Win32 eeeeeeee 161 Creating and Using Dynamic Link Libraries on Windows n 165 NV 8 5 8 6 8 7 8 8 Chapter 9 9 1 9 1 1 9 1 2 9 1 3 9 2 9 2 1 9 2 2 9 2 3 9 2 4 9 2 5 9 2 6 Chapter 10 10 1 10 2 10 3 10 4 10 5 10 5 1 10 6 10 6 1 10 6 2 10 6 3 10 7 10 8 10 9 10 10 10 11 10 12 10 13 10 14 10 14 1 10 14 2 10 14 3 10 14 4 10 14 5 10 14 6 Chapter 11 vi Usmo NEE 170 LAPACK the BLAS and FFTs c ccceececeeeeeeeeeeeeceeeeeeeeeaeseeneeeesaeessenees 170 The C Standard Template Library 170 Environment Variables Auen 171 Fortran C and C Data Types c cccccssseeeeeseeeeeeeneeeneeeeeeeeeeeeens 175 Fortran RENERT 175 Fortran SCalars i s casvieicwisn ttle Mian eh Een 175 FORTRAN 77 Aggregate Data Type Extensions ccscceeteeeeereeeees 178 Fortran 90 Aggregate Data Types Derived Types eessen 180 C and C Data Types 0 ceceeeececeececeeeeeeeeeeeeaeeeeeeeeeseaaeeeseaeeeseeeeesaeeeeaes 180 Cang C LEE 180 C and C Aggregate Data Types ccceeecceeeeceeeeeeeeeeseeeeeeeeeeetaeeeenes 182 Class and Object Data Layout ccccceeeeceeeseeeeeeeeeeceeaeeeeeeeesetaeeeteneeeesas 183 Aggregate Alignment 00 0 0 eccccccceeseccceceeeeeeeeeee
337. stination directory the compiler overwrites it The following example demonstrates the use of output filename extensions pgf95 c proto f protol F This produces the output files proto o and protol o both of which are binary object files Prior to compilation the file protol F is pre processed because it has a F filename extension 6 Getting Started 1 4 Parallel Programming Using the PGI Compilers The PGI compilers support three styles of parallel programming e Automatic shared memory parallel programs compiled using the Mconcur option to pgf77 pgf95 pgcc or pgCC parallel programs of this variety can be run on shared memory parallel SMP systems such as dual core or multi processor workstations OpenMP shared memory parallel programs compiled using the mp option to pgf77 pgf95 pgcc or pgCC parallel programs of this variety can be run on SMP systems Carefully coded user directed parallel programs using OpenMP directives can often achieve significant speed ups on dual core workstations or large numbers of processors on SMP server systems Chapter 5 OpenMP Directives for Fortran and Chapter 6 OpenMP Pragmas for C and C contain complete descriptions of user directed parallel programming e Data parallel shared or distributed memory parallel programs compiled using the PGHPF High Performance Fortran compiler parallel programs of this variety can be run on SMP workstations or servers distributed mem
338. t lt lt func a lt lt numl lt lt b lt lt num2 lt lt ptr c lt lt res lt lt endl res numl1 num2 cout lt lt func res lt lt res lt lt endl Example 10 10 C Main Program cmain c Calling a C Function extern void cpfunc int a int b int c include lt stdio h gt main int a b c a 8 b 2 printf main a d b d ptr c x n a b amp c cpfunc a b amp c printf main res d n c To compile this C function and C main program use the following commands pgcc c cmain c pgCC cmain o cpsub C Executing the resulting a out file should produce the following output main a b 2 ptr c Oxbffffb94 func a b 2 ptr c bffffb94 func res 4 main res 4 8 8 Note that you cannot use the extern C form of declaration for an object s member functions 10 12 Example Fortran Calling C The Fortran main program shown in Example 10 11 Fortran Main Program fmain f calling a C function calls the C function shown in Example 10 12 C function cpfunc C Notice that each argument is defined as a pointer in the C function since Fortran passes by reference Also notice that the C function name uses all lower case and a trailing _ Inter language Calling 197 Example 10 11 Fortran Main Program fmain f calling a C function logical 1 booll character letterl integer 4 numintl numint2 real numfloatl double precision numdo
339. t is truncated at left e g 1234567890abcdef1 x will be 234567890abcdef1 x I110 lt reserved message number gt I111 Underflow of real or double precision constant I112 Overflow of real or double precision constant S113 Label is referenced but never defined S114 Cannot initialize 260 Messages W115 Assignment to DO variable in loop 116 Illegal use of pointer based variable S117 Statement not allowed within a definition The statement may not appear ina STRUCTURE or derived type definition S118 Statement not allowed in DO IF or WHERE block I119 Redundant specification for Data type of indicated symbol specified more than once I120 Label is defined but never referenced I121 Operation requires logical or integer data types An operation in an expression was attempted on data having a data type incompatible with the operation For example a logical expression can consist of only logical elements of type integer or logical Real data would be invalid I122 Character string truncated Character string or Hollerith constant appearing in a DATA statement or PARAMETER statement has been truncated to fit the declared size of the corresponding identifier W123 Hollerith length specification too big reduced The length specifier field of a hollerith constant specified more characters than were present in the character field of the hollerith constant The length specifier was reduced to agree wit
340. t operations link in faster non IEEE libraries if available and disable underflow traps PIC Linux only Generate position independent code Equivalent to fpic Provided for compatibility with other compilers pic Linux only Generate position independent code Equivalent to fpic Provided for compatibility with other compilers 52 Command Line Options trap option option Controls the behavior of the processor when floating point exceptions occur Possible options include D fp e align ignored e inv e denorm e divz e ovf e unf e inexact Ktrap is only processed by the compilers when compiling main functions programs The options inv denorm divz ovf unf and inexact correspond to the processor s exception mask bits invalid operation denormalized operand divide by zero overflow underflow and precision respectively Normally the processor s exception mask bits are on floating point exceptions are masked the processor recovers from the exceptions and continues If a floating point exception occurs and its corresponding mask bit is off or unmasked execution terminates with an arithmetic exception C s SIGFPE signal Ktrap fp is equivalent to Ktrap inv divz ovf Default The default is Knoieee L Specifies a directory to search for libraries Use L to add directories to the search path for library files Multiple L options are valid However the position of multiple L
341. ta type specified for The indicated identifier appears in more than one type specification statement and different data types are specified for it S038 Symbol has not been explicitly declared The indicated identifier must be declared in a type statement this is required when the IMPLICIT NONE statement occurs in the subprogram W039 Symbol appears illegally in a SAVE statement An identifier appearing in a SAVE statement must be a local variable or array 040 Illegal common variable Indicated identifier is a dummy variable is already in a common block or has previously been defined to be something other than a variable or array W041 Illegal use of dummy argument Messages 253 This error can occur in several situations It can occur if dummy arguments were specified on a PROGRAM statement It can also occur if a dummy argument name occurs ina DATA COMMON SAVE or EQUIVALENCE statement A program statement must have an empty argument list 042 is a duplicate dummy argument 043 Illegal attempt to redefine An attempt was made to define a symbol in a manner inconsistent with an earlier definition of the same symbol This can happen for a number of reasons The message attempts to indicate the situation that occurred intrinsic An attempt was made to redefine an intrinsic function A symbol that represents an intrinsic function may be redefined if that symbol has not been previously verified to be an in
342. te but does not E445 Array bounds must be integer The expressions in the array bounds must be integer 446 Argument number to rank mismatch The number of dimensions in the array or array expression does not match the number of dimensions in the dummy argument 447 Argument number to must be a subroutine or function name 448 Argument number to must be a subroutine name S449 Argument number to must be a function name 450 Argument number to kind mismatch S451 Arrays of derived type with a distributed member are not supported 452 Assumed length character is not a dummy argument 288 Messages S453 Derived type variable with pointer member not allowed in IO S454 Subprogram is not a module procedure Only names of module procedures declared in this module or accessed through USE association can appear ina MODULE PROCEDURE statement S455 A derived type array section cannot appear with a member array section A reference like A B where A is a derived type array and B is a member array is not allowed a section subscript may appear after A or after B but not both S456 Unimplemented for data type for MATMUL S457 Illegal expression in initialization S458 Argument to NULL must be a pointer S459 Target of NULL assignment must be a pointer S460 ELEMENTAL procedures cannot be RECURSIVE S461 Dummy arguements of ELEMENATAL procedures must be scala
343. te the actual argument must be a whole array with the HPF SEQUENCE attribute or a contiguous array section of such an array unless an INTERFACE block is used E406 Actual argument and formal argument have different character lengths The actual and formal array character arguments have different character lengths which is allowed only if both character arrays are declared with the HPF SEQUENCE attribute unless an INTERFACE block is used W407 Argument has a different character length than dummy argument The character length of the actual argument is different than the length specified for the corresponding dummy argument 284 Messages W408 Specified main program is not a PROGRAM The main program specified on the command line is a subroutine function or block data subprogram W409 More than one main program in IPA directory and There is more than one main program analyzed in the IPA directory shown The first one found is used W410 No main program found IPA analysis fails The main program must appear in the IPA directory for analysis to proceed W411 Formal argument is DYNAMIC but actual argument is an expression W412 Formal argument is DYNAMIC but actual argument is not I413 Formal argument has two reaching distributions and may be a candidate for cloning I414 and may be aliased and one of them is assigned Interprocedural analysis has determined that two formal arguments because the
344. te DWARF 1 format Code Generation debug information dwarf2 when used with g generate DWARF2 format Code Generation debug information dwarf3 when used with g generate DWARF3 format Code Generation debug information extend the compiler accepts 132 column source code Fortran Language otherwise it accepts 72 column code pgf77 pgf95 and pghpf only extract invokes the function extractor Inlining fcon instructs the compiler to treat floating point con C C Language stants as float data types pgcc and pgCC only fixed the compiler assumes F77 style fixed format Fortran Language source code pgf95 and pghpf only no flushz do don t set SSE flush to zero mode Code Generation 56 Command Line Options pgflag Description Category no fprelaxed option Perform certain floating point intrinsic functions using relaxed precision Optimization around Fortran I O calls pgf77 pgf95 and pghpf only free the compiler assumes F90 style free format Fortran Language source code pgf95 and pghpf only func32 the compiler aligns all functions to 32 byte Code Generation boundaries gccbug s match behavior of certain gcc bugs Miscellaneous noi4 determines how the compiler treats INTEGER Optimization variables pgf77 pgf95 and pghpf only info prints informational messages regarding optimi Miscellaneous zation and code generation to standard output as compilation proceed
345. ted in Normal C Mode ccccccccsseeeeeessstteeeeseeaes 298 cfront 2 1 Compatibility Mode ou eecccceceeeeeeeeeeeeeeeeeeeeeeeeeaeeeeeeeeessaaeeeeaes 299 cfront 2 1 3 0 Compatibility Mode Contents Tables Table P 1 Table P 2 Table 1 1 Table 2 1 Table 3 1 Table 3 2 Table 3 3 Table 3 4 Table 5 1 Table 6 1 Table 7 1 Table 7 2 Table 9 1 Table 9 2 Table 9 3 Table 9 4 Table 9 5 Table 10 1 Table 10 2 Table 10 3 Table A 1 Table A 2 Table A 3 Table A 4 Table A 5 Table A 6 Table A 7 Table A 8 Table A 9 Table A 10 Table A 11 Table A 12 Table A 13 Table A 14 Table A 15 Table A 16 Table A 17 Table A 18 PGI Compilers and Commande retr setnnrnnnsnnnsssrnesns XV Processor OPtiONs TE xvi Stop after Options Inputs and Outputs 000 0 eee eesseceeeeeeteeeeeeseeseeeeeeeaaes 6 Optimization and O g and M lt opt gt Options ccececeeseeeeteeeeeeteeeeees 37 Generic PGI Compiler Options ccccceceeenceeeeeeeeeeeeeeeeaaeeeeaaeesecaeeesaeeeeaes 40 C and C specific Compiler Options ccccceeesceeeeceeeeeeeeeeeteeeseaeeeeees 43 M Options Summary ENEE 55 Optimization and O g Mvect and Mconcur Options ssssssseneeeanene 85 Initialization of REDUCTION Variables ccceceeeeeeceeeeeeneeeeeteeeeeneeeeeees 110 Initialization of Reduction Vaables AA 128 Fortran Optimization Directive Gummanm 143 C C Pragma SUMIMAL
346. ted the critical section Likewise each of the remaining threads will wait its turn to execute the statements in the critical section Critical sections cannot be nested and any such specifications are ignored Branching into or out of a critical section is illegal If a name argument appears on a CRITICAL directive the same name must appear on the END CRITICAL directive PROGRAM CRITICAL USE REAL A 100 100 MX LMX INTEGER I J MX 1 0 LMX 1 0 CALL RANDOM SEED CALL RANDOM NUMBER A SOMP PARALLEL PRIVATE I FIRSTPRIVATE LMX SOMP DO OpenMP Directives for Fortran 111 DO J 1 100 DO I 1 100 LMX MAX A I J LMX ENDDO ENDDO SOMP CRITICAL MX MAX MX LMX SOMP END CRITICAL SOMP END PARALLEL PRINT MAX VALUE OF A IS MX END Note that this program could also be implemented without the critical region by declaring MX as a reduction variable and performing the MAX calculation in the loop using MX directly rather than using LMX See Section 5 2 PARALLEL END PARALLEL and Section 5 6 DO END DO for more information on how to use the REDUCTION clause on a parallel DO loop 5 4 MASTER END MASTER The OpenMP END MASTER directive uses the following syntax SOMP MASTER lt Fortran code in body of MASTER section gt SOMP END MASTER In a parallel region of code there may be a sub region of code that should execute only on the master thread Instead of ending the parall
347. tely flat name space The names of entities with external linkage must be projected onto that name space so that they do not conflict with one another A function f from a class A for example must not have the same external name as a function f from class B Some names are not names in the conventional sense of the word they re not strings of alphanumeric characters for example operator We can see that there are two problems here 1 Generating external names that will not clash 2 Generating alphanumeric names for entities with strange names in C Name mangling solves these problems by generating external names that will not clash and alphanumeric names for entities with strange names in C It also solves the problem of generating hidden names for some behind the scenes language support in such a way that they will match up across separate compilations You will see mangled names if you view files that are translated by PGC and you do not use tools that demangle the C names Intermediate files that use mangled names include the assembly and object files created by the pgCC command and the C like file that can be viewed as output from pgCC using the i command line option The name mangling algorithm for the PGC compiler is the same as that for cfront and also matches the description in Section 7 2 Function Name Encoding of The Annotated C Reference Manual except for some minor details Refer to the ARM for a complet
348. tem_clock count_rate hz H call system_clock count clock0 lt do work gt call system_clock count clock1 t clockl1 clock0 time real sum t real hz size t 38 Optimization amp Parallelization Chapter 3 Command Line Options This chapter describes the syntax and operation of each compiler option The options are arranged in alphabetical order On a command line options need to be preceded by a hyphen If the compiler does not recognize an option it passes the option to the linker This chapter uses the following notation item Square brackets indicate that the enclosed item is optional item item Braces indicate that you must select one and only one of the enclosed items A vertical bar I separates the choices Horizontal ellipses indicate that zero or more instances of the preceding item are valid NOTE Some options do not allow a space between the option and its argument or within an argument This fact is noted in the syntax section of the respective option 39 Table 3 1 Generic PGI Compiler Options Option Description Display invocation information HHH Show but do not execute the driver commands same as dryrun byteswapio Fortran only Swap bytes from big endian to little endian or vice versa on input output of unformatted data C Instrument the generated executable to perform array bounds check
349. ter char ch signed char sch If Muchar is specified on the command line pgcc Muchar myprog c char ch above is equivalent to unsigned char ch M lt pgflag gt Optimization Controls Syntax Mcache_ align Align unconstrained objects of length greater than or equal to 16 bytes on cache line boundaries An unconstrained object is a data object that is not a member of an aggregate structure or common block This option does not affect the alignment of allocatable or automatic arrays cache_align Note To effect cache line alignment of stack based local variables the main program or function must be compiled with Mcache_align Mconcur option option Instructs the compiler to enable auto concurrentization of loops If Mconcur is specified multiple processors will be used to execute loops that the compiler determines to be parallelizable Where option is one of the following noJaltcode n Instructs the parallelizer to generate alternate serial code for parallelized loops If altcode is specified without arguments the parallelizer determines an appropriate cutoff length and generates serial code to be executed whenever the loop count is less than or equal to that length If altcode n is specified the serial altcode is executed whenever the loop count is less than or equal to n If noaltcode is specified the parallelized version of the loop is always executed regardless of the loop count 70 Command Line Opt
350. ter to register in this same format Values are converted back to IEEE 32 bit or IEEE 64 bit when stored back to memory for REAL float and DOUBLE PRECISION double data respectively The default precision of the x87 floating point stack can be reduced to IEEE 32 bit or IEEE 64 bit globally by compiling the main program with the pc 32 64 option to the PGI Workstation compilers which is described in detail in Chapter 3 Command Line Options However there is no way to ensure that operations performed in mixed precision will match those produced on a traditional load store RISC UNIX system which implements IEEE 64 bit and IEEE 32 bit registers and associated floating point arithmetic instructions In contrast arithmetic results produced on Intel Pentium 4 Xeon AMD Opteron Athlon64 or Intel EM64T processors will usually closely match or be identical to those produced on a traditional RISCH UNIX system if all scalar arithmetic is performed using SSE SSE2 instructions You should keep this in mind when porting applications to and from systems which support both x87 and full SSE SSE2 floating point arithmetic Many subtle issues can arise which affect your numerical results sometimes to several digits of accuracy 2 4 Loop Unrolling using Munroll This optimization unrolls loops executing multiple instances of the loop during each iteration This reduces branch overhead and can improve execution speed by creating better opportunities for
351. terrupt processes Functions called during signal handling have no unusual restriction on their use of registers Moreover if a signal handling function returns the process resumes its original execution path with registers restored to their original values Thus programs and compilers may freely use all registers without danger of signal handlers changing their values A 3 2 Function Return Values Functions Returning Scalars or No Value A function that returns an integral or pointer value that fits in 64 bits places its result in rax A function that returns a floating point value that fits in the XMM registers returns this value in xmm0 A function that returns a value in memory via the stack places the address of this memory passed to the function as a hidden first argument in rcx in rax Functions that return no value also called procedures or void functions put no particular value in any register A call instruction pushes the address of the next instruction the return address onto the stack The return instruction pops the address off the stack and effectively continues execution at the next instruction after the call instruction A function that returns a scalar or no value must preserve the caller s registers as described above Additionally the called function must remove the return address from the stack leaving the stack pointer rsp with the value it had before the call instruction was executed Run time Env
352. th Linux and Windows operating systems See the PGI Fortran Reference manual for a complete list of available routines in the PGI implementation of LIB3F 8 6 LAPACK the BLAS and FFTs Pre compiled versions of the public domain LAPACK and BLAS libraries are included with the PGI compilers on Linux and Windows systems in the files PGI lt target gt lib lapack a and PGI lt target gt lib blas a respectively where lt target gt is replaced with the appropriate target name linux86 linux86 64 win64 or nt86 To use these libraries simply link them in using the l option when linking your main program pgf95 myprog f lblas llapack Highly optimized assembly coded versions of the BLAS and certain FFT routines may be available for your platform In some cases these are shipped with the PGI compilers See the current release notes for the PGI compilers you are using to determine if these optimized libraries exist where they can be downloaded if necessary and how to incorporate them into your installation as the default 8 7 The C Standard Template Library The PGC compiler includes a bundled copy of the STLPort Standard C Library See the online Standard C Library tutorial and reference manual at http www stlport com for further details and licensing 170 Libraries and Environment Variables 8 8 Environment Variables Several environment variables can be used to alter the default behavior of the PGI compilers and t
353. the Minfo option is present on the compile line See Chapter 3 Command Line Options for a complete specification of Mconcur Optimization amp Parallelization 25 A loop is considered parallelizable if doesn t contain any cross iteration data dependencies Cross iteration dependencies from reductions and expandable scalars are excluded from consideration enabling more loops to be parallelizable In general loops with calls are not parallelized due to unknown side effects Also loops with low trip counts are not parallelized since the overhead in setting up and starting a parallel loop will likely outweigh the potential benefits In addition the default is to not parallelize innermost loops since these often by definition are vectorizable using SSE instructions and it is seldom profitable to both vectorize and parallelize the same loop especially on multi core processors Compiler switches and directives are available to let you override most of these restrictions on auto parallelization 2 6 1 Auto parallelization Sub options The parallelizer performs various operations that can be controlled by arguments to the Mconcur command line option The following sections describe these arguments that affect the operation of the vectorizer In addition these vectorizer operations can be controlled from within code using directives and pragmas For details on the use of directives and pragmas refer to Chapter 7 Optimization Directives an
354. the same size Structures and unions of other sizes will be passed as a pointer to a temporary allocated by the caller and whose value contains the value of the argument The caller allocated temporary memory used for arguments of aggregate type must be 16 byte aligned Passing Arguments on the Stack Registers are assigned using the argument s ordinal position in the argument list For example if a function s first argument is an integral type and its second argument is a floating point type the first argument will be passed in the first general purpose register Yorcx and the second argument will be passed in the second XMM register xmm1 the first XMM register and second general purpose register are ignored Arguments after the first four are passed on the stack they are pushed on the stack in reverse order with the last argument pushed first 236 Run time Environment Table A 15 Register Allocation for Example A 4 shows the register allocation and stack frame offsets for the function declaration and call shown in the following example Example A 4 Parameter Passing typedef struct int i float f structl int i float f double d long 1 long long 11 SETUCEL s1 extern void func int i float f structl1 sl double d long long 11 long 1 fune a E et dy EI Et Table A 15 Register Allocation for Example A 4 General Purpose d f Stack Frame Registers Floating Point Registers Offset
355. this context WORD type cannot be converted Interface block for has wrong number of arguments Interface block for should have Interface block for should not have Messages E430 Interface block for has wrong W431 Program is too large for Interprocedural Analysis to complete W432 Illegal type conversion E433 Subprogram called within INDEPENDENT loop not LOCAL W434 Incorrect home array specification ignored S435 Array declared with zero size An array was declared with a zero or negative dimension bound as real a 1 or an upper bound less than the lower bound as real a 4 2 W436 Independent loop not parallelized W437 Type will be mapped to Where DOUBLE PRECISION is not supported it is mapped to REAL and similarly for COMPLEX 16 or COMPLEX 372 E438 not supported on this platform This construct is not supported by the compiler for this target S439 An internal subprogram cannot be passed as argument S440 Defined assignment statements may not appear in WHERE statement or WHERE block S441 may not appear in a FORALL block Messages 287 E442 Adjustable length character type not supported on this host S443 EQUIVALENCE of derived types not supported on this host S444 Derived type in EQUIVALENCE statement must have SEQUENCE attribute A variable or array with derived type appears in an EQUIVALENCE statement The derived type must have the SEQUENCE attribu
356. tically and includes a field showing the category 54 Command Line Options Table 3 3 M Options Summary pgflag Description Category anno annotate the assembly code with source code Miscellaneous no autoinline C C when a function is declared with the inline Inlining keyword inline it at O2 and above no asmkeyword specifies whether the compiler allows the asm C C Language keyword in C C source files pgcc and pgCC only no backslash determines how the backslash character is treated Fortran Language in quoted strings pgf77 pgf95 and pghpf only no bounds specifies whether array bounds checking is Miscellaneous enabled or disabled no builtin Do don t compile with math subroutine builtin Optimization support which causes selected math library rou tines to be inlined pgcc and pgCC only byteswapio Swap byte order big endian to little endian or Miscellaneous vice versa during I O of Fortran unformatted data cache_align where possible align data objects of size greater Optimization than or equal to 16 bytes on cache line bound aries chkfpstk check for internal consistency of the x87 FP stack Miscellaneous in the prologue of a function and after returning from a function or subroutine call tp px p5 p6 piii targets only chkptr check for NULL pointers pgf95 and pghpf only Miscellaneous chkstk check the stack for available space upon entry to Miscellan
357. ting Started 5 Table 1 1 Stop after Options Inputs and Outputs Stop Option after Input Output E prepro Source files must have F extension for Fortran preprocessed file to cessing standard out F prepro Source files must have F extension this option is not preprocessed file f cessing valid for pgcc or pgCC P prepro Source files this option is not valid for pgf77 pgf95 preprocessed file i cessing or pghpf S compi Source files or preprocessed files assembly language lation file s c assem Source files preprocessed files or assembly language unlinked object file bly files 0 none linking Source files preprocessed files assembly language executable files a out files object files or libraries If you specify multiple input files or do not specify an object filename the compiler uses the input filenames to derive corresponding default output filenames of the following form where filename is the input filename without its extension filename f indicates a preprocessed file if you compiled a Fortran file using the P option Dlename stindicates a listing file from the Mlist option filename o indicates an object file from the c option filename s indicates an assembly language file from the S option Note Unless you specify otherwise the destination directory for any output file is the current working directory If the file exists in the de
358. tion level that can be specified in combination with Mipa is O2 For example if you specify Mipa O1 on the command line the optimization level will automatically be elevated to O2 by the compiler driver It is typical and recommended to use Mipa fast Many of the following sub options can be prefaced with no which reverses or disables the effect of the sub option if it s included in an aggregate sub option like Mipa fast The choices of option are no align recognize when targets of a pointer dummy are aligned default is noalign no arg remove arguments replaced by const ptr default is noarg no cg generate call graph information for viewing using the pgicg command line utility default is nocg no const perform interprocedural constant propagation default is const Command Line Options Command Line Options except lt func gt no f90ptr fast force no globals inline n ipofile no keepobj no libc no libinline no libopt no localarg main lt func gt no ptr no pure used with inline to specify functions which should not be inlined default is to inline all eligible functions according to internally defined heuristics F90 F95 pointer disambiguation across calls default is nof90ptr choose IPA options generally optimal for the target Use help to see the settings for Mipa fast on a given target force all objects to re compile regardless of
359. tional NOWAIT clause is specified Nested single process sections are ignored Branching into or out of a single process section is not supported PROGRAM SINGLE USE INTEGER A 0 1 INTEGER omp_get_thread_num SOMP PARALLEL A omp_get_thread_num omp_get_thread_num SOMP SINGLE PRINT YOU SHOULD ONLY SEE THIS ONCE SOMP END SINGLE SOMP END PARALLEL PRINT A O A O A 1 A 1 END The PRIVATE and FIRSTPRIVATE clauses are as described in Section 5 2 PARALLEL END PARALLEL The COPYPRIVATE clause causes the variables in list to be copied from the private copies in the single thread that executes the SINGLE region to the other copies in all other threads of the team at the end of the SINGLE region The COPYPRIVATE clause must not be used with NOWAIT 5 6 DO END DO The OpenMP DO END DO directive uses the following syntax Syntax OpenMP Directives for Fortran 113 SOMP DO Clauses lt Fortran DO loop to be executed in parallel gt SOMP END DO NOWAIT Clauses PRIVATE list FIRSTPRIVATE list LASTPRIVATE list REDUCTION operator intrinsic list SCHEDULE type chunk ORDERED The real purpose of supporting parallel execution is the distribution of work across the available threads You can explicitly manage work distribution with constructs such as IF omp_get_thread_num EQ 0 THEN ELSE l IF omp_get_thread_num EQ 1 THEN ENDIF However these cons
360. tive was not correctly formed W248 Syntax error in line A line directive was not correctly formed W249 Syntax error in module A module directive was not correctly formed W250 Syntax error in undef A undef directive was not correctly formed W251 Token after ifdef must be identifier The ifdef directive was not followed by an identifier W252 Token after ifndef must be identifier The ifndef directive was not followed by an identifier 253 Too many actual parameters to The number of actual arguments to the indicated macro exceeded the maximum allowed currently 31 S254 Too many formal parameters to 272 Messages The number of formal arguments to the indicated macro exceeded the maximum allowed currently 31 F255 Too much pushback The preprocessor ran out of space while processing a macro expansion The macro may be recursive W256 Undefined directive The identifier following a was not a directive name S257 EOF in include directive End of file was encountered while processing a include directive 258 Unmatched elif A elif directive was encountered with no preceding if or elif directive S259 Unmatched else A else directive was encountered with no preceding if or elif directive 260 Unmatched endif A endif directive was encountered with no preceding if ifdef or ifndef directive S261 Include files nested too deeply The nesting depth of include direct
361. tly 2048 S233 lt reserved message number gt W234 Illegal directive name The sequence of characters following a sign was not an identifier W235 Illegal macro name A macro name was not an identifier 236 Illegal number The indicated number contained a syntax error F237 Line too long The input source line length exceeded the maximum allowed currently 2048 W238 Missing endif End of file was encountered before a required endif directive was found W239 Missing argument list for A call of the indicated macro had no argument list S240 Number too long The length of a number exceeded the maximum allowed currently 2048 W241 Redefinition of symbol The indicated macro name was redefined Messages 271 I242 Redundant definition for symbol A definition for the indicated macro name was found that was the same as a previous definition F243 String too long The length of a quoted string exceeded the maximum allowed currently 2048 244 Syntax error in define formal not identifier A formal parameter that was not an identifier was used in a macro definition W245 Syntax error in define missing blank after name or arglist There was no space or tab between a macro name or argument list and the macro s definition 246 Syntax error in if A syntax error was found while parsing the expression following a if or elif directive 247 Syntax error in include The include direc
362. to allow the user to query whether dynamic adjustment of the number of threads available for execution of parallel regions is enabled This function is recognized but currently always returns zero include lt omp h gt double omp_get_wtime returns the elapsed wall clock time in seconds as a floating point double value Times returned are per thread times and are not necessarily globally consistent across all threads include lt omp h gt double omp_get_wtick returns the resolution of omp_get_wtime in seconds as a floating point double value include lt omp h gt void omp_init_lock omp_lock_t lock void omp_init_nest_lock omp_nest_lock_t lock initializes a lock associated with the variable lock for use in subsequent calls to lock routines This initial state of lock is unlocked It is illegal to make a call to this routine if lock is already associated with a lock include lt omp h gt void omp_destroy_lock omp_lock_t lock void omp_destroy_nest_lock omp_nest_lock_t lock disassociates a lock associated with the variable lock include lt omp h gt void omp_set_lock omp_lock_t lock void omp_set_nest_lock omp_nest_lock_t lock causes the calling thread to wait until the specified lock is available The thread gains ownership of the lock when it is available It is illegal to make a call to this routine if lock has not been associated with a lock 138 OpenMP Pragmas for C and C include
363. to files larger than 2GB Large File Support Mnostartup instructs the linker not to link in the standard startup routine that contains the entry point _start for the program Note If you use the Mnostartup option and do not supply an entry point the linker issues the following error message Warning cannot find entry symbol _start Mnostddef instructs the compiler not to predefine any macros to the preprocessor when compiling a C program Mnostdlib 64 Command Line Options instructs the linker not to link in the standard libraries libpgftnrtl a libm a libc a and libpgc a in the library directory lib within the standard directory You can link in your own library with the l option or specify a library directory with the L option Default For arguments that you do not specify the default environment option depends on your configuration Cross reference D I L 1 U M lt pgflag gt Inlining Controls This section describes the M lt pgflag gt options that control function inlining Syntax Mext ract option option Extracts functions from the file indicated on the command line and creates or appends to the specified extract directory where option can be any of name func instructs the extractor to extract function func from the file size number instructs the extractor to extract functions with number or fewer statements from the file lib filename ext Use directory filename ext as t
364. tors using functions that take enum types and no class types e Explicit specification of function template arguments is supported Unnamed template parameters are supported e The new lookup rules for member references of the form x A PB and p gt A B are supported e The notation template and gt template etc is supported C 3 The following language features are not accepted The following feature of the ISO IEC 14882 1998 C standard is not supported e Exported templates are not implemented CA Extensions Accepted in Normal C Mode The following extensions are accepted in all modes except when strict ANSI violations are diagnosed as errors see the A option e A friend declaration for a class may omit the class keyword class A friend B Should be friend class B Constants of scalar type may be defined within classes class A const int size 10 int al size e In the declaration of a class member a qualified name may be used struct A int A Should be int e The preprocessing symbol c_plusplus is defined in addition to the standard __cplusplus e An assignment operator declared in a derived class with a parameter type matching one of its base classes is treated as a default assignment operator that is such a declaration blocks the implicit generation of a copy assignment operator This is cfront behavior that is known to be relied upon in at least one wi
365. tran PVF and The Portland Group are trademarks and PGI PGHPF PGF77 PGCC PGPROF and PGDBG are registered trademarks of STMicroelectronics Inc Other brands and names are the property of their respective owners The use of STLport a C Library is licensed separately and license distribution and copyright notice can be found in the online documentation for a given release of the PGI compilers and tools PGI User s Guide Copyright 1998 2000 The Portland Group Inc Copyright 2000 2006 STMicroelectronics Inc All rights reserved Printed in the United States of America First Printing Release 1 7 Jun 1998 Second Printing Release 3 0 Jan 1999 Third Printing Release 3 1 Sep 1999 Fourth Printing Release 3 2 Sep 2000 Fifth Printing Release 4 0 May 2002 Sixth Printing Release 5 0 Jun 2003 Seventh Printing Release 5 1 Nov 2003 Eight Printing Release 5 2 Jun 2004 Ninth Printing Release 6 0 Mar 2005 Tenth Printing Release 6 1 Dec 2005 Eleventh Printing Release 6 2 Aug 2006 Technical support trs pgroup com Sales sales pgroup com Web http www pgroup com ii Contents Preface Chapter 1 1 1 1 2 1 6 1 6 1 1 6 2 Chapter 2 2 1 2 2 2 3 2 3 1 2 4 2 5 2 5 1 2 5 1 1 2 5 1 2 2 5 1 3 2 5 1 4 2 5 2 2 6 2 6 1 E a a T XI Audience Description ssesssssessssrnesesrennnesrnnnnestnnnnernnnnnnnnnnneannnnnnnnnnnnnennnnna xi Compatibility and Conformance to Gtandarde eee een ae
366. trinsic function For example the intrinsic sin can be defined to be an integer array If a symbol is verified to be an intrinsic function via the INTRINSIC statement or via an intrinsic function reference then it must be referred to as an intrinsic function for the remainder of the program unit symbol An attempt was made to redefine a symbol that was previously defined An example of this is to declare a symbol to be a PARAMETER which was previously declared to be a subprogram argument S044 Multiple declaration for symbol A redundant declaration of a symbol has occurred For example an attempt was made to declare a symbol as an ENTRY when that symbol was previously declared as an ENTRY S045 Data type of entry point disagrees with function The current function has entry points with data types inconsistent with the data type of the current function For example the function returns type character and an entry point returns type complex S046 Data type length specifier in wrong position The CHARACTER data type specifier has a different position for the length specifier from the other data types Suppose we want to declare arrays ARRAYA and ARRAYB to have 8 elements each having an element length of 4 bytes The difference is that ARRAYA is character and ARRAYB is integer The declarations would be CHARACTER ARRAYA 8 4 and INTEGER ARRAYB 4 8 047 More than seven dimensions specified for array 048 Illegal use of si in declar
367. tructs are not in the form of directives The DO END DO directive pair provides a convenient mechanism for the distribution of loop iterations across the available threads in a parallel region Items to note about clauses are Variables declared in a PRIVATE list are treated as private to each processor participating in parallel execution of the loop meaning that a separate copy of the variable exists on each processor Variables declared in a FIRSTPRIVATE list are PRIVATE and in addition are initialized from the original object existing before the construct Variables declared in a LASTPRIVATE list are PRIVATE and in addition the thread that executes the sequentially last iteration updates the version of the object that existed before the construct The REDUCTION clause is as described in Section 5 2 PARALLEL END PARALLEL The SCHEDULE clause is explained in the following section If ORDERED code blocks are contained in the dynamic extent of the DO directive the ORDERED clause must be present For more information on ORDERED code blocks see Section 5 14 ORDERED The DO END DO directive pair directs the compiler to distribute the iterative DO loop immediately following the OMP DO directive across the threads available to the program The DO loop is executed in parallel by the team that was started by an enclosing parallel region If the OMP END DO directive is not specified the OMP DO is assumed to end with the enclosed
368. tterl integer numintl numint2 double precision numdoub1 real numfloatl integer 2 numshor1 booll true letterl v numinti 11 Mumint2 44 numdoubl 902 numfloatl 39 6 numshorl 299 return end Example 10 14 C main program cpmain C include lt iostream gt extern C extern void forts char char int int float double short main char booll letterl int numintl numint2 float numfloati1 double numdoubl1 short numshor1 forts amp booll1 amp letterl amp numint1 amp numint2 amp numfloatl amp numdoubl1 amp numshor1 cout lt lt booll booll cout lt lt TRUE cout lt lt FALSE cout lt lt endl cout lt lt letterl lt lt letterl lt lt endl cout lt lt numintl lt lt numinti lt lt endl cout lt lt numint2 lt lt numint2 lt lt endl cout lt lt numfloatl lt lt numfloatl lt lt endl cout lt lt numdoubl lt lt numdoubl lt lt endl cout lt lt numshorl lt lt numshorl lt lt endl Inter language Calling 199 To compile this Fortran subroutine and C program use the following command lines pgf95 c forts f pgCC forts o cpmain C lpgf95 lpgf95 rpml lpgf952 lpgf95rt1l lpgftnrtl Executing this C main should produce the following output bool1 TRUE letterl Sy numint1 11 numint2 44 numfloatl 39 6 numdoubl 902 numshor1 2
369. u can also use a size prefix followed by a number If size is specified what follows is always taken as a number levels number instructs the inliner to perform number levels of inlining The default number is 1 If you specify both func and number the compiler inlines functions that match the function name or have number or fewer statements For examples of inlining functions see Function Inlining Usage In the following example the compiler extracts functions that have 500 or fewer statements from the source file myprog f and saves them in the file extract il pgf95 Mextract 500 oextract il myprog f In the following example the compiler inlines functions with fewer than approximately 100 statements in the source file myprog f and writes the executable code in the default output file a out pgf95 Minline size 100 myprog f Cross reference o M lt pgflag gt Fortran Language Controls This section describes the M lt pgflag gt options that affect Fortran language interpretations by the PGI Fortran compilers These options are only valid to the pgf77 pgf95 and pghpf compiler drivers Syntax Mbackslash the compiler treats the backslash as a normal character and not as an escape character in quoted strings 66 Command Line Options Mnobackslash Mdclchk Mnodclchk Mdefaultunit Mnodefaultunit Mdlines Mnodlines Mdollar char Mextend Mfixed Mfree Miomutex Mnoio
370. ub1 integer 2 numshorl external cfunc call cpfunc booll letterl numinti numint2 numfloatl1 numdoub1 numshor1 write L2 A2 I5 I5 F6 1 F6 1 I5 booll letterl numintl numint2 numfloatl numdoub1 numshorl end Example 10 12 C function cpfunc C define TRUE Oxff define FALSE 0 extern C extern void cpfunc_ Assuming the Fortran program is in a file fmain f and the C function is in a file cpfunc C create char booll letterl int numintl numint2 float numfloati double numdoubl short numshortl1 intlen_letter1 booll TRUE letterl v numintl 11 numint2 44 numfloatl 39 6 numdoubl 39 2 numshortl1 981 an executable using the following command lines pgCC c cpfunc Cc pgf95 cpfunc o fmain f Executing the a out file should produce the following output T v 11 44 39 6 39 2 981 198 Inter language Calling 10 13 Example C Calling Fortran Example 10 13 Fortran Subroutine forts f shows a Fortran subroutine called by the C main program shown in Example 10 14 C main program cpmain C Notice that each call uses the amp operator to pass by reference Also notice that the call to the Fortran subroutine uses all lower case and a trailing _ Example 10 13 Fortran Subroutine forts f subroutine forts booll letterl numintl1 amp numint2 numfloati numdoub1 numshor1 logical 1 booll character le
371. uild an executable file for dhry in which proc7 plus any functions of approximately 10 or fewer statements are inlined one level only Note that the specified functions are inlined only if they are previously placed in the inline library temp il during the extract phase 104 Function Inlining pgf95 dhry f Mextract o temp il pgf95 dhry f Minline 10 Proc7 temp il Assume the program fibo f contains a single function fibo that calls itself recursively The following command line creates the file fibo o in which fibo is inlined into itself pgf95 fibo f c Mrecursive Minline fibo Because this version of fibo recurses only half as deeply it executes noticeably faster Using the same source file dhry f the following example builds an executable for dhry in which all functions of roughly ten or fewer statements are inlined Two levels of inlining are performed This means that if function A calls function B and B calls C and both B and C are inlinable then the version of B which is inlined into A will have had C inlined into it pgf95 dhry f Minline size 10 levels 2 4 5 Restrictions on Inlining The following Fortran subprograms cannot be extracted e Main or BLOCK DATA programs e Subprograms containing alternate return assigned GO TO DATA SAVE or EQUIVALENCE statements e Subprograms containing FORMAT statements Subprograms containing multiple entries A Fortran subprogram is not inlined if any of the following
372. ument or in certain contexts within I O statements must not consist of a concatenation involving a passed length character variable W093 Type conversion of expression performed 258 Messages An expression of some data type appears in a context which requires an expression of some other data type The compiler generates code to convert the expression into the required type 094 Variable is of wrong data type The indicated variable is used in a context which requires a variable of some other data type S095 Expression has wrong data type An expression of some data type appears in a context which requires an expression of some other data type 096 Illegal complex comparison The relations LT GT GE and LE are not allowed for complex values S097 Statement label has been defined more than once More than one statement with the indicated statement number occurs in the subprogram S098 Divide by zero 099 Illegal use of Aggregate record references may only appear in aggregate assignment statements unformatted I O statements and as parameters to subprograms They may not appear for example in expressions Also records with differing structure types may not be assigned to one another 100 Expression cannot be promoted to a vector An expression was used that required a scalar quantity to be promoted to a vector illegally For example the assignment of a character constant string to a character array Record
373. used Fortran uses column major order and C C use row major order For one dimensional arrays this poses no problems For two dimensional arrays where there are an equal number of rows and columns row and column indexes can simply be reversed For arrays other than single dimensional arrays and square two dimensional arrays inter language function mixing is not recommended 10 8 Example Fortran Calling C Example 10 4 C function cfunc_ shows a C function that is called by the Fortran main program shown in Example 10 3 Fortran Main Program fmain f Notice that each argument is defined as a pointer since Fortran passes by reference Also notice that the C function name uses all lower case n and a trailing Inter language Calling 193 Example 10 3 Fortran Main Program fmain f logical 1 booll character letterl integer 4 numintl numint2 real numfloatl double precision numdoub1 integer 2 numshorl external cfunc call cfunc booll letterl numinti numint2 amp numfloat1 numdoub1 numshor1 write L2 A2 I5 I5 F6 1 F6 1 I5 amp booll letterl numintl numint2 numfloatl amp numdoub1 numshorl1 end Example 10 4 C function cfunc_ define TRUE Oxff define FALSE 0 void cfunc_ booll letterl numinti numint2 numfloat1 numdoub1 numshorl len letter1 char booll letterl1 int numintl numint2 float numfloatl double numdoub1 short numshorl1 int len_letter1
374. vel Internal compiler data structure overflow working storage exhausted or some other non recoverable problem related to the size of the subprogram If this error occurs at opt 2 reducing the opt level to 1 may work around the problem Moving the subprogram being compiled to its own source file may eliminate the problem If this error occurs while compiling a subprogram of fewer than 2000 statements it should be reported to the compiler maintenance group as a possible compiler problem F008 Error limit exceeded The compiler gives up because too many severe errors were issued the error limit can be reset on the command line F009 Unable to open assembly file Probably user does not have write permission for the current working directory F010 File write error occurred Probably file system is full S011 Unrecognized command line switch Refer to PDS reference document for list of allowed compiler switches 012 Value required for command line switch Certain switches require an immediately following value such as opt 2 Messages 249 v000 Internal compiler error This message indicates an error in the compiler rather than a user error although it may be possi ble for a user error to cause an internal error The severity may vary if it is informative or warning correct object code was probably generated but it is not safe to rely on this Regardless of the severity or cause internal errors should be
375. ver must be available on the system where the executable is run The Mdll flag must be used when an executable is linked against a DLL built by the PGI compilers The following examples outline how to use Mmakedll and Mmakeimplib to build and use DLLs with the PGI compilers Example 1 Build a DLL out of a single source file object1 f which exports data and a subroutine using DLLEXPORT Build the main source file prog f which uses DLLIMPORT to import the data and subroutine from the DLL object1 f subroutine subi i DECS ATTRIBUTES DLLEXPORT subl integer i common acommon adata integer adata DECS ATTRIBUTES DLLEXPORT acommon print subl adata adata print lt subl a p i adata i end progl f program progl common acommon adata integer adata external subi 166 Libraries and Environment Variables DECS ATTRIBUTES DLLIMPORT subl acommon adata 11 call subi 12 print main adata adata end Step 1 Create the DLL obj1 dll and its import library obj1 lib using the following series of commands pgf95 c objectl f pgf95 Mmakedll objectl obj o obji1 dll Step 2 Compile the main program pgf95 Mdll o progl progl f defaultlib objl The Mdll switch causes the compiler to link against the PGI runtime DLLs instead of the PGI runtime static libraries The Mdll switch is required when linking against any PGI compiled DLL such as obj1 dll The defaultlib switch is use
376. vere errors were issued the error limit can be reset on the command line F009 Unable to open assembly file 250 Messages Probably user does not have write permission for the current working directory F010 File write error occurred Probably file system is full 011 Unrecognized command line switch Refer to PDS reference document for list of allowed compiler switches 012 Value required for command line switch Certain switches require an immediately following value such as opt 2 013 Unrecognized value specified for command line switch 014 Ambiguous command line switch Too short an abbreviation was used for one of the switches W015 Hexadecimal or octal constant truncated to fit data type I016 Identifier truncated to 31 chars An identifier may be at most 31 characters in length characters after the 31st are ignored 017 Unable to open include file File is missing read protected or maximum include depth 10 exceeded Remember that the file name should be enclosed in quotes 018 Illegal label Used for label field errors or illegal values E g in fixed source form the label field first five characters of the indicated line contains a non numeric character 019 Illegally placed continuation line A continuation line does not follow an initial line or more than 99 continuation lines were specified 020 Unrecognized compiler directive Refer to user s ma
377. wever performance of programs executed in this manner can be unpredictable and oftentimes will be inefficient OMP_SCHEDULE specifies the type of iteration scheduling to use for DO and PARALLEL DO loops which include the SCHEDULE RUNTIME clause The default value for this variable is STATIC If the optional chunk size is not set a chunk size of is assumed except in the case of a STATIC schedule For a STATIC schedule the default is as defined in Section 5 6 DO END DO Examples of the use of OMP_SCHEDULE are as follows setenv OMP_ SCHEDULE STATIC 5 setenv OMP_ SCHEDULE GUIDED 8 setenv OMP_ SCHEDULE DYNAMIC OMP_DYNAMIC currently has no effect OMP_NESTED currently has no effect MPSTKZ increase the size of the stacks used by threads executing in parallel regions It is for use with programs that utilize large amounts of thread local storage in the form of private variables or local variables in functions or subroutines called within parallel regions The value should be an integer lt n gt concatenated with M or m to specify stack sizes of n megabytes For example setenv MPSTKZ 8M 124 OpenMP Directives for Fortran Chapter6 OpenMP Pragmas for C and C The PGCC ANSI C and C compilers support the OpenMP C C Application Program Interface The OpenMP shared memory parallel programming model is defined by a collection of compiler directives or pragmas library routines and environment variab
378. wing example the object file a out contains symbolic debugging information pgf95 g myprog f gopt Use of g alters how optimized code is generated in ways that are intended to enable or improve debugging of optimized code The gopt option instructs the compiler to include symbolic debugging information in the object file and to generate optimized code identical to that generated when g is not specified Default The compiler does not put debugging information into the object module Usage In the following example the object file a out contains symbolic debugging information pgf95 gopt myprog f g77libs Linux only Use the g77libs option on the link line if you are linking g77 compiled program units into a pgf95 compiled main program using the pgf95 driver When this option is present the pgf95 driver will search the necessary g77 support libraries to resolve references specific to g77 compiled program units The g77 compiler must be installed on the system on which linking occurs in order for this option to function correctly Default The compiler does not search g77 support libraries to resolve references at link time 50 Command Line Options Usage The following command line requests that g77 support libraries be searched at link time pgf95 g77libs myprog f g77_object o help Used with no other options help displays options recognized by the driver on the standard output When used in combinatio
379. with Inline Libraries An inline library is implemented as a directory with each inline function in the library stored as a file using an encoded form of the inlinable function A special file named TOC in the inline library directory serves as a table of contents for the inline library This is a printable ASCII file which can be examined to find out information about the library contents such as names and sizes of functions the source file from which they were extracted the version number of the extractor which created the entry etc Libraries and their elements can be manipulated using ordinary system commands e Inline libraries can be copied or renamed e Elements of libraries can be deleted or copied from one library to another The Is command can be used to determine the last change date of a library entry Dependencies in Makefiles When a library is created or updated using one of the PGI compilers the last change date of the library directory is updated This allows a library to be listed as a dependence in a makefile and ensures that the necessary compilations will be performed when a library is changed 4 2 2 Updating Inline Libraries Makefiles If you use inline libraries you need to be certain that they remain up to date with the source files into which they are inlined One way to assure inline libraries are updated is to include them in a makefile The makefile fragment in the following example assumes the file utils
380. y Kernighan and Ritchie Prentice Hall C A Reference Manual by Samuel P Harbison and Guy L Steele Jr Prentice Hall 1987 The Annotated C Reference Manual by Margaret Ellis and Bjarne Stroustrup AT amp T Bell Laboratories Inc Addison Wesley Publishing Co 1990 OpenMP Application Program Interface Version 2 5 May 2005 OpenMP Architecture Review Board 1997 2005 Preface xvii xviii Preface Chapter 1 Getting Started This chapter describes how to use the PGI compilers The command used to invoke a compiler for example the pgf95 command is called a compiler driver The compiler driver controls the following phases of compilation preprocessing compiling assembling and linking Once a file is compiled and an executable file is produced you can execute debug or profile the program on your system Executables produced by the PGI compilers are unconstrained meaning they can be executed on any compatible x86 or x64 processor based system regardless of whether the PGI compilers are installed on that system 1 1 Overview In general using a PGI compiler involves three steps 1 Produce a program in a file containing a f extension or another appropriate extension see Section 1 3 1 Input Files This may be a program that you have written or a program that you are modifying 2 Compile the program using the appropriate compiler command 3 Execute debug or profile the executable file on your system The PGI
381. ymbols that you expect it to contain You can also create your own def file containing the symbols you want to export to the DLL To use your def file add it to the link line and omit output def Passed to linker Generate an import library named lt file gt for the DLL A DLL s import library is the interface used when linking an executable that depends on routines in a DLL Passed to linker Use this flag to export all global and weak defined symbols to the DLL Even with this flag some symbols are not exported see no default excludes Passed to linker When export all symbols is used there are still some special symbols i e DI Main 12 that are not exported Use no default excludes to export these symbols to the DLL To use the PGI compilers to create an executable that links to the DLL form of the runtime use the compiler flag Mdll The executable built will be smaller than one built without Mdll the PGI runtime DLLs however must be available when the executable is run The Mdll flag must be used when an executable is linked against a DLL built by the PGI compilers Each PGI compiler can also create DLLs for Windows The following examples outline how to use Mmakedll to do so Example 8 1 Build a DLL out of two source files object1 f and object2 f and use it to build the main source file prog1 f 1 object1 f subroutine subfl1 n integer n n 1 print n n return end
382. zed loops when appropriate For each vectorized loop the compiler decides whether to generate altcode and what type or types to generate which may be any or all of altcode without iteration peeling altcode with non temporal stores and other data cache optimizations and altcode based on array alignments calculated dynamically at runtime The compiler also determines suitable loop count and array alignment conditions for executing the alcode This option is enabled by default This disables alternate code generation for vectorized loops Instructs the vectorizer to enable certain associativity conversions that can change the results of a computation due to roundoff error A typical optimization is to change an arithmetic operation to an arithmetic operation that is mathematically correct but can be computationally different due to round off error Instructs the vectorizer to disable associativity conversions Instructs the vectorizer when performing cache tiling optimizations to assume a cache size of n The default is set using per processor type either using the tp switch or auto detected from the host computer Generate vector code for all loops where possible regardless of the number of statements in the loop This overrides a heuristic in the vectorizer 77 that ordinarily prevents vectorization of loops with a number of statements that exceeds a certain threshold The default is nosizelimit smallvect n Instructs

PGI User's Guide

Contents

Download Pdf Manuals

Related Search

Related Contents