The POWHEG BOX user manual: common features
Contents
1. 2 If the above fails if the manyseeds flag is not set or if it is set and a file named pwggrid cj dat already exists and is consistent all files of the form pwggrid 7 dat are loaded and suitably combined and the steps ISG and UBB are skipped go to 6 15 3 If the above fails if a file pwgxgrid dat exist and is consistent this file is loaded and steps ISG is skipped go to 5 4 Step ISG is performed The resulting grid is stored in the file pwgxgrid dat This step whether the manyseeds flag is set or not is performed using the default initial seed value i e not the seed found at the cj line of the pwgseeds dat file In this way all copies of the program being run will use the same importance sampling greed This is mandatory if we want to combine results 5 Step UBB is performed If the manyseeds flag is set this step is performed using the seed found at the cj line of the pwgseeds dat file and the result is stored in the file pwggrid cj dat Otherwise the current seed value is used and the result is stored in a file named pwggrid dat 6 If a file named pwgubound dat exists and is consistent it is loaded The UBR step is skipped goto 10 7 If a file named pwgubound dat exists and is consistent it is loaded The UBR step is skipped goto 10 8 If the above fails if the manyseeds flag is not set or if it is set and a file named pwgubound cJ dat exists and is consistent all files with na2. and the set is chosen according to the tokens lhans1 lhans2 In this case the numbering scheme is that of the LHAGLUE interface leaving the possibility of re evaluate pdf s on the fly using number corresponding to LHpdf file or to interpolate from a previously calculated grid number corresponding to LHgrid file as explained in ref 4 In the example above 10050 corresponds to the central value of From release 1 0 this is the default behaviour of the program as it is distributed the CTEQ6M set in this latter case ebeami 7000 energy of beam 1 in GeV ebeam2 7000 energy of beam 2 in GeV These are the energies of the two beams given in GeV We assume that beam 1 and 2 move along the third axis in the positive and negative direction respectively Parameters to allow or not the use of stored data use old grid 1 use old ubound 1 The meaning of these tokens requires a little knowledge of the operation of POWHEG BOX Before events can be generated the program goes through an initialization stage It first computes the importance sampling grid for the computation of the inclusive cross section The importance sampling data is stored in the file pwgxgrid dat Then the integral is computed and an upper bounding envelope is found for the unweighted generation of the underlying Born configurations This envelope together with the importance sampling data and the cross section are stored in the file pwggrid dat Afterwords a
3. blank lines are ignored 3 Ifa ora appears at any point in a line the part of the line starting from the or symbol up to its end is blanked 4 An entry has the format name value usually followed by a and a comment to clarify the meaning of the variable The name keyword has no more than 20 characters and value is an integer or floating point number 5 A maximum of 100 keywords are allowed The order in which the lines are written is not important although it is a good practice to put the mandatory tokens at the beginning of the file 3During the integration stage and or during the generation of the event file the possibility to perform a NLO analysis or an analysis at the level of the POWHEG output before interfacing to the shower is left to the user In these cases the analysis is executed with the string WHCPRG set to NLO The input parameters are read by the real 8 function powheginput string whose source code is in the file powheginput f The statement rvalue powheginput myparm returns the value of the token myparm stored in powheg input If the token is not found in the input file a message is printed and the program is stopped The file is read only once on the first invocation of the function powheginput and token value pairs are stored in internal arrays so that subsequent calls to powheginput are relatively fast The statement rvalue powheginput myparm also returns
4. With this method events do not concentrate in the low kr region However their weight in the low kr region becomes divergent After shower if one imposes the analysis cut one gets a finite cross section since it is unlikely that events with small transverse momentum at the Born level may pass the analysis cut after shower In fact shower transverse momenta larger than the one present in the initial Born process must be suppressed in the Monte Carlo generator In recent POWHEG BOX revisions both methods can be implemented at the same time We wanted in fact to be able to implement the following three options e Generate events using a transverse momentum generation cut e Generate events using a Born suppression factor and a small transverse momentum cut just enough to avoid unphysical values of the strong coupling constant and of the factorization scale that appears in the parton density functions e Apply a Born suppression factor and set the transverse momentum cut to zero In this case the program cannot be used to generate events It can be used however to produce NLO fixed order distributions provided the renormalization and factor ization scales are set in such a way that they remain large enough even at small ke This feature is only used for the generation of fixed order distributions The generation cut is activated by setting the token bornktmin to the desired value in the powheg input file The Born suppression is activate
5. features however become really useful for processes where the Born contribution itself is singular The simplest examples are the Z jet and the dijet production processes Here we discuss Z jet The dijet case is fully analogous 5 1 1 Generation cut and Born suppression factor The Z 1j process differs substantially from all processes previously implemented in POWHEG in the fact that the Born diagram itself is collinear and infrared divergent In all previous implementations the Born diagram was finite and it was thus possible to generate an unweighted set of underlying Born configurations covering the whole phase space In the present case this is not possible since they would all populate the very low transverse momentum region Of course this problem is also present in standard Shower Monte Carlo programs where it is dealt with by generating the Born configuration with a cut kgen on the transverse momentum of the Z boson After the shower one must discard all events that fail some transverse momentum analysis cut kan in order to get a realistic sample The analysis cut kan may be applied to the transverse momentum of the Z or to the hardest jet We assume here for sake of discussion that the analysis cut is applied to the Z transverse momentum 11 Taking kan Z kgen is not enough to get a realistic sample In fact in an event generated at the Born level with a given kr lt kgen the shower may increase the transverse mome
6. grid for the normalization of the upper bounding function for the generation of radiation is computed The normalization grid is stored in the file pwgubound dat The generation of the grids is time consuming but the time spent in this calculation is negligible in a normal run when hundreds of thousands of events are generated On the other hand sometimes it is useful for example when debugging an analysis program to skip the grid generation stage For this purpose if the use old grid token is set equal to 1 and if pwggrid dat exists and is consistent it is loaded and the old grid and old value of the cross section are used Otherwise first the program checks if a pwgxgrid dat file exists If it does the file is loaded and the generation of the importance sampling grid is skipped Observe that the program does check the file for consistency with the current run but the check is not exhaustive The user should make sure that a consistent grid is used The token use old ubound has the same role as use old grid but it applies to the step where the normalization of the upper bounding function is filled The following parameters are used to control the generation of these grids Parameters that control the grid for Born variables generation ncalli 50000 number of calls for initializing the integration grid itmx1 5 number of iterations for initializing the integration grid ncall2 50000 number of calls for computing integral itmx2
7. the events generated by POWHEG are unweighted i e all come with the same weight At times however this feature has some drawbacks In this subsection we describe how the generation of weighted events can be performed and we specify when this is particularly needed Two flags control the nature of the output in the POWHEG BOX withnegweights and bornsuppfact If neither of these flags is set events are output with weight 1 i e the XWGTUP variable is set to 1 The IDWTUP variable in the Les Houches interface is set to 3 in this case The total cross section is stored in the xsecup variable Negative weighted events are neglected with this choice 5In old versions the flag bornsuppfact was named ptsupp which is now deprecated 10 If withnegweights is set to 1 true in our convention negative weighted events are not discarded The IDWTUP variable in the Les Houches interface is set to 4 and the weight of the event is set to its sign times the sum of the total cross section for positive weighted events plus the absolute value of the cross section for negative weighted events In this way the average value of XWGTUP equals the real cross section as required by the Les Houches convention when the IDWTUP variable is set to 4 The variable xsecup always stores the real cross section If the bornsuppfact token is set a suppression factor that depends upon the under lying Born configuration of each event is supplied with it The cro
8. 5 number of iterations for computing integral foldcsi 1 number of folds on x integration foldy 1 number of folds on y integration foldphi 1 number of folds on phi integration nubound 50000 number of calls to setup upper bounds for radiation iymax 1 lt 100 number of intervals in y grid to compute upper bounds icsimax 1 lt 100 number of intervals in csi grid xupbound 2 increase upper bound for radiation generation by given factor In the proc_dir directories there are example powheg input files with the settings of these variables that have been found to be satisfactory for the specific parameters that have been used The values of some of the tokens may be changed in the following cases e The integration does not seem to converge well and the file pwgbtlgrid top repre senting the importance sampling subdivision of each integration coordinate for each iteration itmx1 does not seem to converge well In this case ncall1 and if that does not work also itmx1 should be increased e The integration results have large errors One may try to increase ncall12 itmx2 e The number of upper bound failure for the generation of the inclusive cross section reported in pwgcounters dat is an important fraction of the total number of gen erated events As a rule of thumb one might expect that fraction to represent the error on the generated distribution In this case increase ncal12 itmx2 e If the fraction of negative weight
9. G the appropriate include files should also be substituted 3 1 The analysis routines The file pwhg_analysis f contains a template analysis that one can take as a starting point for more complex analysis It uses pwhg_bookhist the histogramming package of M L Mangano with minor modifications and it produces topdrawer outputs The routines in pwhg_analysis f are adequate for both fortran HERWIG and PYTHIA since they rely on the standard common blocks of ref 8 If the user wants to use other analysis routines he she can simply modify the pwhg_analysis f file or write his her own 4 Input parameters When generating the hard events the pwhg_main program needs to set some physical and some technical parameters Some of these parameters are mandatory some other are not Moreover there are some parameters which are needed only for some processes In the POWHEG BOX an independent facility to set these input parameters is available All parame ters are stored in a file named powheg input If the file powheg input is not present the program asks the user to enter a prefix and then looks for the file lt prefix gt powheg input In this case all the output files created by POWHEG BOX in the current run will carry the prefix lt prefix gt instead of pwg Examples of powheg input files can be found in the testrun subdirectory of proc_dir The format of these files is as follows 1 Lines are no more than 100 characters long 2 Empty
10. Preprint typeset in JHEP style PAPER VERSION March 16 2011 POWHEG BOX rev PEELE oook kkk release 1 0 The POWHEG BOX user manual common features Simone Alioli Deutsches Elektronen Synchrotron DESY Platanenallee 6 D 15738 Zeuthen Germany E mail simone alioli desy de Paolo Nason INFN Sezione di Milano Bicocca Piazza della Scienza 3 20126 Milan Italy E mail Paolo Nason mib infn it Carlo Oleari Universita di Milano Bicocca and INFN Sezione di Milano Bicocca Piazza della Scienza 8 20126 Milan Italy E mail Carlo Oleari mib infn it Emanuele Re Institute for Particle Physics Phenomenology Department of Physics University of Durham Durham DH1 8LE UK E mail emanuele re durham ac uk ABSTRACT This note documents the use of the package POWHEG BOX In this document the user will find a description of the general features of the package Issues related to specific processes are discussed in dedicated manuals available in the corresponding folders KEYWORDS POWHEG Shower Monte Carlo NLO Contents 1 Introduction 1 2 Organization of the package 2 3 Modes of operation 3 1 The analysis routines 5 4 Input parameters 5 4 1 Mandatory parameters 6 4 2 Optional parameters 9 5 Special modes of operation 10 5 1 Weighted event generation 10 5 1 1 Generation cut and Born suppression factor 11 5 2 The pdfreweight flag 13 5 3 The manyseeds flag 14 6 Counters and statistics 16 7 Random number
11. d by setting the token bornsuppfact 12 to a positive real value The process specific subroutine born_suppression sets the sup pression factor to ka k bornsuppfact If bornsuppfact is negative the suppression factor is set to 1 In the POWHEG approach negative weighted events can only arise if one is approaching a region where the NLO computation is no longer feasible In our studies for the Z 1j process we approach this region at small transverse momentum In order to better see what happens there rather than neglecting negative weights that is the default behavior of the POWHEG BOX we have introduced a new feature in the program that allows one to track also the negative weighted events This feature is activated by setting the token withnegweights to 1 true If withnegweights is set to 1 events with negative weight can thus appear in the Les Houches event file While we normally set the IDWTUP flag in the Les Houches interface to 3 in this case we set it to 4 With this flag the SMC is supposed to simply process the event without taking any other action Furthermore the XWGTUP Les Houches common block variable is set by the POWHEG BOX to the sign of the event times the integral of the absolute value of the cross section in such a way that its average equals the true total cross section Notice that if withnegweights is set and a Born suppression factor is also present the events will have variable XWGTUP of either
12. ded other parameters that may also be useful for developers The normal user is asked not to change them since their invocation is time consuming and or may cause some conflicts with other settings If instead the user is interested in changing them a detailed explana tion of their behavior can be found on ref 3 testsuda 0 default 0 do not test tests the Sudakov FF by numerical integration testplots 0 default 0 do not do NLO and PWHG distributions bornonly 0 default 0 if 1 do Born only smartsig 0 default 1 remember equal amplitudes 0 do not remember withsubtr 0 default 1 subtract real counterterms 0 do not subtract radregion 1 default all regions only generate radiation in the selected singular region iupperisr 1 default 1 choice of ISR upper bounding functional form iupperfsr 2 default 2 choice of FSR upper bounding functional form flgdebug 1 default 0 write extra information on LHEF Other technical parameters have been introduced for more specific debugging purposes for example par_diexp par_2gsupp jacsing were relevant for dijets As already stated we refer to the process manual for details on other specific parameters 5 Special modes of operation In this subsection we describe some special features that were added to our package to comply with experimental needs or to deal with complicated processes namely V j and dijets 5 1 Weighted event generation In normal conditions
13. ed to dynamic libraries according to the corresponding Makefile In the following we denote a generic process folder as proc_dir Once downloaded an executable related to a given process can be built with the fol lowing commands cd POWHEG BOX proc_dir make lt target gt where the choice of the lt target gt depends upon the way one wants to interface the program with a Shower Monte Carlo Specifically lt target gt needs to refer to one of the targets con tained in the Makefile All the process directories contain in their Makefile instructions to build the following programs pwhg_main lhef_analysis main HERWIG lhef main PYTHIA lhef For some processes other executables can be built If this is the case we refer to the specific process manual for more details In the next section we will briefly describe the purpose of the aforementioned programs From release 1 0 the POWHEG BOX program also includes a stand alone PDF package originally due to M L Mangano In order to use this package it is enough to set the Makefile variable PDF equal to native When this is done the program doesn t need to be linked against any external PDF library as LHAPDF and therefore can be run out of the box Some data files needed for this package are in the POWHEG BOX pdfdata directory and a symbolic link of the data file to the directory where the run is performed is needed The name of the link should not include the tb1 or dat ending
14. event file The format of the event file supported by POWHEG BOX is the Standard format for Les Houches event files documented in ref 7 Les Houches Event Files LHEF from now on generated with our program have the suffix lhe The program for the generation of LHEF can be built from a proc_dir with the command make pwhg main The program must be executed in a directory where the file powheg input is present The only libraries needed by this program to work are LHAPDF and in some cases FASTJET The event file is named pwgevents lhe By using an input card with a prefix the user is given the possibility to change the name of this file and of all the other output files as documented in sec 4 2 Perform the parton shower The subsequent step is to read the events and process them with the SMC An ex ample program that reads the event file showers it with HERWIG and analyzes the results can be built as follows make main HERWIG lhef A similar program named main PYTHIA lhef is provided for PYTHIA and can be built with the command make main PYTHIA lhef 2Notice that this structure has been introduced from release 1 0 whereas in previous versions all the driver files were placed in the processes subdirectories and were named main HERWIG lhef f and main PYTHIA lhef f A version of PYTHIA and HERWIG is included in the POWHEG BOX package These can be substituted by the user s favorite version In the case of HERWI
15. eweighting each event with a weight equal to the ratio between the new and the reference PDF s values 13 This is done also when using SMC programs In particular the evaluation of PDF s ratio is performed by saving for each event the momentum fractions of the incoming partons which are involved in the hard collision and the corresponding values of the reference PDF s Strictly speaking this procedure is not valid since the result of a SMC does not depend linearly on PDF s because these enter also in the Sudakov form factors Nevertheless it is common belief that this procedure should capture the dominant part of uncertainties due to PDF s also in SMC simulations In this respect the output of POWHEG is affected by the same problems of SMC s the dependence on PDF s being non linear Moreover in POWHEG the real contributions are evaluated with different structure functions and different Feynman x s so that another reason to doubt about this procedure is present Despite all the aforementioned caveat we decided to make the information needed for the reweighting procedure available also in POWHEG to give the opportunity to perform such studies In particular when the token pdfreweight is set to 1 at the end of each event in the LHEF a line is added The format is the following pdf idi id2 x1 x2 xmufact xf1 xf2 Apart from the pdf tag the other numbers refer to the two id s of the incoming par tons that en
16. generator 17 1 Introduction The POWHEG BOX program is a framework to implement Next to Leading Order NLO cal culations in Shower Monte Carlo SMC programs according to the POWHEG method An explanation of the method and a detailed discussion of how the code is organized can be found in refs 1 2 3 The code is distributed according to the MCNET GUIDELINES for Event Generator Authors and Users and can be found in http powhegbox mib infn it The latest version of the package can be downloaded trough SVN svn checkout username anonymous password anonymous svn powhegbox mib infn it trunk POWHEG BOX Previous revisions are available using the revision n option Periodic releases of the code will be available in svn powhegbox mib infn it tags 2 Organization of the package In the POWHEG BOX directory the main directory from now on there are the common source files an include directory a Docs directory and several other process folders with a name reminiscent of the process to which they correspond the Z folder contains the code to simulate the Drell Yan process at NLO with POWHEG Zj the code to simulate Z in association with 1 jet at NLO and so on Each folder contains a Makefile and some specific source files which are compiled and linked with the common files to build an independent program Therefore an executable built in a process directory is a self contained program it needs only to be link
17. he POWHEG method JHEP 0711 2007 070 arXiv 0709 2092 hep phj 2 3 S Alioli P Nason C Oleari and E Re A general framework for implementing NLO calculations in shower Monte Carlo programs the POWHEG BOX JHEP 1006 043 2010 ar Xiv 1002 2581 hep ph 4 M R Whalley D Bourilkov and R C Group The Les Houches accord PDFs LHAPDF and LHAGLUE arXiv hep ph 0508110 5 M Cacciari and G P Salam Dispelling the N myth for the k jet finder Phys Lett B 641 57 2006 arXiv hep ph 0512210 6 E Boos et al Generic user process interface for event generators arXiv hep ph 0109068 7 J Alwall et al A standard format for Les Houches event files Comput Phys Commun 176 2007 300 arXiv hep ph 0609017 8 T Sj strand et al in Z physics at LEP1 Event generators and software eds G Altarelli R Kleiss and C Verzegnassi Vol 3 pg 327 9 P Nason MINT a Computer Program for Adaptive Monte Carlo Integration and Generation of Unweighted Distributions arXiv 0709 2085 hep ph S47
18. ic processes For some processes other special actions are needed to install the program In those cases we refer again to the specific process manuals 3 Modes of operation The main purpose of our implementations of the POWHEG method is to generate hard events that can then be fed into a SMC program for subsequent showering To this aim POWHEG BOX saves the hard event information according to the conventions of the Les Houches Interface for User Processes LHIUP from now on 6 The SMC should also comply with these conventions as is the case for PYTHIA and HERWIG in order to be used in conjunction with the POWHEG BOX Having this in mind we now briefly explain the purpose of the 4 programs we mentioned above e pwhg_main when this executable is run the program performs several steps which are documented in detail in 3 The final output is the file with extension lhe standing for Les Houches events that contains the generated hard events written according to the format described in ref 7 Several other files are produced where internal information are stored the most important ones are dat files and some top files with plots Since this is the main program a more detailed description on how to run the program write or modify input cards and interpret the results is given later in this document e lhef_analysis when this executable is run the program performs an analysis of the events in the 1lhe file The purpose
19. integer is the line number of the random seed to be used for the current run Assuming a es that the number 17 is given as input to the pwhg main program the run will produce files named pwgxgrid dat and pwggrid 0017 dat The pwhg main program can be run with different integers as input Each run can be sent for example to a different node of a cluster At the end of this step a bunch of pwggrid dat pwgubound dat files and pwgNLO top files will be present in the run directory The pwgNLO 7777 top files are statistically independent topdrawer histograms They can be combined to pro vide a higher statistics NLO analysis of the current analysis routines If the subsequent runs are sent after the file pwgxgrid dat was already produced and if the flag use old grid is set to one the importance sampling grid will be loaded from the pwgxgrid dat Otherwise it will be recreated 4 Now set the nev token to a given number make sure that the flag use old grid is set to one and run a bunch of copies of the pwhg_main program each with the same integers as input The program will now load all the pwggrid dat and pwgubound dat that it can find and combines them adequately assuming that they are all statistically independent and will start to generate events The events will be in files pwgevent lhe and they will all be independent statistically The sequence above is a quite simple two ste
20. ite ups This generator has default initialization If a user wishes to start the program with different seeds he she should add lines similar to Random number generator initializing parameters iseed 6093726 initialize random number sequence randi 1 initialize random number sequence rand2 1 initialize random number sequence to the input card This results in a call to the rm48in iseed rand1 rand2 subroutine that seeds the generator with the integer iseed and skip the first randit trand2 10 8 numbers as documented in the CERNLIB manual This can be useful if one wants to resume a previous run In that case one has simply to use as initializing values those reported in the lt prefix gt events lhe file If instead one just wants to change the seed only he she can comment or skip the rand1 and rand2 lines in the input card We remind the reader that a change in the random number generator initialization affects the POWHEG BOX random number sequence both in the generation of events and in NLO computation or upper bound searching when the corresponding grids are not present If the program is interfaced to a SMC the user should also take care to initialize the seeds of the latter References 1 P Nason A new method for combining NLO QCD with shower Monte Carlo algorithms JHEP 0411 2004 040 arXiv hep ph 0409146 S Frixione P Nason and C Oleari Matching NLO QCD computations with Parton Shower simulations t
21. mes of the form pwgubound dat are loaded and combined The UBR step is skipped go to 10 9 The UBR step is performed If the manyseeds flag is not set the result is stored in a file named pwgubound dat Otherwise it is stored in the file pwgubound cj dat 10 Now nev events are generated If the manyseeds flag is not set the result is stored in a file named pwgevents lhe Otherwise it is stored in the file pwgevents cj dat This logic has the purpose to allow several possible combinations of actions For example one can use grids generated in parallel runs to produce events without using the manyseeds flag Or one can use grids generated without the manyseeds flag for generating events in parallel with the manyseeds flag set 6 Counters and statistics Several results relevant to the interpretation of the output of the run are written into the files pwgstat dat and pwgcounters dat The fraction of negative weights the total cross section the number of upper bound failures in the generation of the inclusive cross section and the generation efficiency together with failures and efficiency in the generation of hard radiation are printed there These are quite self explanatory and we do not comment them any further These numbers are sufficient to take action in case of problems as explained in sec 4 16 7 Random number generator POWHEG BOX uses the RM48 random number generator documented in the CERNLIB wr
22. ntum of the jet so that kr gt kan Thus the generation cut even if it is below the analysis cut may reduce the number of events that pass the analysis cut Of course as we lower kgen keeping kan fixed we will reach a point when very few events below kgey will pass the analysis cut kan In fact generation of radiation with transverse momentum larger than kgen is strongly suppressed in POWHEG and in turn radiation from subsequent shower is required to be not harder than the hardest radiation of POWHEG Thus given the fact that we want to generate a sample with a given kan cut we should choose kgen small enough so that the final sample remains substantially the same if kgen is lowered even further A second option for the implementation of processes with a divergent Born contribution is also available It requires that we generate weighed events rather than unweighted ones This is done by using a suppressed cross section for the generation of the underlying Born configurations Bsupp B x rth 5 1 where B is the inclusive NLO cross section at fixed underlying Born variables and kr is the transverse momentum of the vector boson in the underlying Born configuration In this way Bane is integrable and one can use it to generate underlying Born configurations according to its value The generated event however should be given a weight 1 F kr rather than a constant one in order to compensate for the initial F kr suppression factor
23. of the data file for example cteq6m tbl1 should be linked to cteq6m If instead one needs other PDF sets the user is asked to have the LHAPDF library 4 in stalled on his her system and to take care to insert its correct search path in the Makefile or simply to add the path of the lhapdf config executable to the PATH environmental ln several testrun directories the user will find the link already present variable We remind that in case of linking against dynamic shared library the correct LHAPDF library path should also be added to the LD_LIBRARY_PATH environmental variable otherwise run time errors may occur For some processes the default analysis routine that comes with the package relies on jet algorithms as can be inferred from the Makefile or by direct inspection of the pwhg analysis f file in proc_dir When this is the case in order to build the program the user needs to have the FASTJET library 5 installed It is up to the user to correctly install it and to modify the Makefile accordingly For most systems adding the fastjet config executable to the PATH environmental variable is enough A dummy analysis file not invoking FASTJET is also provided for users that have problems linking the FASTJET library The Makefile is set up to use the compiler gfortran on Linux platforms If one wishes to use g77 one should change the appropriate lines in the Makefile The ifort compiler has also been tested for some specif
24. of this program is to allow a quick analysis of partonic events before showering them Therefore it is useful mainly for devel opers rather than for users The output is the file LHEF_analysis top where NLO distributions with N LL resummation of soft gluon effects are present e main HERWIG lhef and main PYTHIA lhef these programs are used to feed the events stored in the lhe file to the HERWIG or PYTHIA program to obtain a full event simulation The corresponding main sources main HERWIG f and main PYTHIA f are located in the main directory However some of the routines called therein need to be customized differently for different processes for this reason their source code has been placed in the process dependent files setup HERWIG lhef f and setup PYTHIA lhef f These files are also important because they show how to call the template analysis at the end of the event generation Having said this we also recall that these drivers should be considered as templates to write an interface to use POWHEG within the experimental collaboration software since the information in the 1lhe files are already written according to a standard documented format 7 The typical procedure to simulate events with our POWHEG implementation is performed in two stages which we briefly describe in the following 1 Generating the hard events The first step of a POWHEG simulation is to generate and store the hard events in a file which we call the
25. p procedure It is useful however to better clarify the logic that the POWHEG BOX follows in this procedure There are 3 steps in the initialization phase of POWHEG First of all an importance sampling grid is determined Let us call this stage ISG Importance Sampling Grid The second step is the calculation of the integrals and the determination of an upper bounding envelope for the B function to be used for the generation of underlying Born configurations We call this stage the UBB Upper Bounds for underlying Born As a third step the upper bound normalization for radiation is determined We call this stage UBR Upper Bounds for Radiation First of all we remark that if the use old grid flag is not set to 1 no grid file is loaded Similarly if the use old ubound flag is not set to 1 no ubound file is loaded In other words if these flags are set looking for a corresponding file will always yield a negative result The reader should keep this in mind when reading the following procedure If the flag manyseeds is set the pwhg_main program asks for an integer We will call cj this integer between 0 and 9999 We denote with cj the corresponding string of four digits leading digits are set to 0 thus if ic 1 cj 0001 We will denote as 7777 any four digit string The logic of grid loading in POWHEG BOX is as follows 1 Ifa file pwggrid dat exist and is consistent this file is loaded and steps ISG and UBB are skipped go to 6
26. rmthr 1 5 default 1 5 GeV charm threshold for gluon splitting bottomthr 5 0 default 5 0 GeV bottom threshold for gluon splitting charmthrpdf 1 5 default 1 5 GeV pdf charm threshold bottomthrpdf 5 0 default 5 0 GeV pdf bottom threshold The first token can be used to set explicitly the value of Agcp to a given value In a standard run there is no need to perform this operation actually it may produce wrong results because by default the value read from the PDF table is properly used The other parameters are cutoff used in the programs the first three are the cutoff for gener ating emission off light c and b quarks respectively They are also used to set heavy flavor thresholds in the strong coupling running Instead the last two parameters control the threshold values at which heavy flavor PDF s start to be nonzero withdamp 1 default 0 do not use use Born zero damping factor hfact 100 default no dumping factor dump factor for high pt radiation gt 0 dampfac h 2 pt2 h 2 These tokens control the separation of the full real matrix element in a singular and a nonsingular part Their use may be needed in presence of processes where the Born cross section vanishes in some phase space region They are also useful to test the behavior of the program in presence of large K factors Their exact meaning is explained in ref 3 For testing the correct behavior of the program and to obtain NLO distributions we ad
27. s is large one may perform a so called folded integration This procedure has been described in refs 2 3 9 to which we refer the reader more details The procedure is enabled by setting at least one of the parameters foldcsi foldy foldphi to a value different from 1 allowed values are 1 2 5 10 25 50 The speed of the program is inversely proportional to the product of these numbers so that a reasonable compromise should be found For processes where the fraction of negative weights in the B calculation is non negligible the corresponding input card comes already with reasonable folding parameters set e If there are too many upper bound violations in the generation of radiation one may increase nubound and or xupbound e If the efficiency in the generation of radiation is too small one may try to increase iymax icsimax In order to check whether any of these conditions occurs the user should inspect the files pwgstat dat and pwgcounters dat at the end of the run as explained in section 6 In formation present in these files are also printed on the shell during the run The input cards of some processes need more parameters such as the value of masses widths couplings etc Specific information can be found in the corresponding dedicated manual The tokens to control scale variations are not mandatory parameters However we describe them here facscfact 1 factorization scale factor mufact muref facscfact ren
28. scfact 1 renormalization scale factor muren muref renscfact Factorization and renormalization scale factors appearing here have to do with the com putation of the inclusive cross section i e the B function 1 2 3 and can be varied by a factor of order 1 to study scale dependence As usual the value used as central value depends on the process at hand The exact details for the process at hand can be found in the dedicated manual The relevant fortran code can be found in the set_fac_ren_scales routine Born_phsp f file which the experienced user can modify at his her will and risk 4 2 Optional parameters In addiction to the mandatory parameters presented above POWHEG BOX also accepts other parameters For some processes the use of these parameters is actually needed in some cases they are mandatory In these cases proper warnings and instructions are present in the dedicated manual and the input files have the corresponding lines uncommented and with the tokens set to proper values In general however the user should not worry if these parameters are not present or are commented This means that there is no need to use them and the default values are used In the following we describe them since they can be useful for a more advanced use of the program who is not interested can safely skip this section QCDlambdad 0 25 for not equal pdf sets ptsqmin 0 8 default 0 8 GeV minimum pt for generation of radiation cha
29. signs In this case XWGTUP is set to the sign of the event times the absolute value of the cross section divided by the suppression factor bornsuppfact Also in this case the average value of XWGTUP coincides withe the true total cross section We preferred not to use the option 3 in case of signed events with constant absolute value This option is advocated by the Les Houches interface precisely in such cases However the Les Houches interface does not provide a standard way to store the integral of the absolute value of the cross section that would be needed to compute correctly the weight of the event In fact the XSECUP variable is reserved for the true total cross section More specifically if we have N events of either sign they should be weighted with the sum of the positive plus the absolute value of the negative part of the cross section in such a way that N Wi le lel N a lel Noxo 5 2 i 1 where W are the sign of the event 1 because LW _ 04 I 5 3 N a 11 Weighted events are also useful if one wants to generate a homogeneous sample from relatively low up to very high transverse momenta It is convenient in this case to pick a very large bornsuppfact value of the order of the maximum transverse momentum one is interested in The large momentum region will be more populated in this way 5 2 The pdfreweight flag Nowadays it is common practice to estimate uncertainties due to PDF s by r
30. ss section computed by the pwhg main program is in this case not valid It is the integral of the cross section times the suppression factor Events are generated using this fake cross section and thus are weighted with the inverse of the suppression factor The IDWTUP variable in the Les Houches interface is set to 4 The weight of the event is in this case the sign times the total cross section for positive weighted events plus the absolute value of the cross section for negative weighted events times the inverse of the weighting factor The weight factor is returned by the user routine born_suppression that can use the value of the bornsuppfact token as a parameter to compute the suppression factor Also in this case the average value of the weight of the event is equal to the real cross section This option can be active in conjunction with the withnegweights flag These flags have many uses On one side one might like to know where negative weighted events end up Even if they constitute a small fraction we may worry that they could end up in some tiny tail of some important distribution One may also prefer to work with negative weight in cases when getting rid of them requires high folding numbers the foldcsi foldy and foldphi tokens and thus has a high cost in computer time The bornsuppfact feature can be use to enhance a region of phase space like a high kr tail where it would be otherwise difficult to get high statistics These
31. ter the Born process PDG conventions to the two momentum fractions of these partons to the value of the factorization scale at which PDF s were evaluated and finally to the value of the PDF s times the momentum fractions With this information it should be straightforward to perform the reweighting procedure Before concluding this subsection we would like to stress again that we cannot guaran tee that this reweighting procedure gives the same results one would obtain by performing a completely new run using different PDF s from the very beginning 5 3 The manyseeds flag The run time needed to produce a large sample of events may become significant especially for complicated processes To circumvent this problem we introduced a feature that can be used for running the POWHEG BOX on several nodes of a cluster and optimize the efficiency The relevant flag is manyseeds When it is set to 1 the POWHEG BOX program looks for a file named pwgseeds dat and stops if this file is not found The typical sequence of operations to perform parallel runs is then the following one 1 Prepare a powheg input file with the manyseeds flag is set to 1 Set the number of events nev to 0 2 Prepare a file pwgseeds dat containing a sequence one per line of different random number seeds For example first line 1 second line 2 etc but any number will do 3 Run the pwhg_main program It will ask an integer for input Input an integer That
32. the value of the token myparm However in case the token myparm is not present the program does not stop and returns the value 10 This is used for optional keywords that are given a default value when absent The file powheginput f is a stand alone code and can be linked to any program In this way a SMC that is reading an event file may get parameters of the POWHEG BOX run if needed In the rest of this section we describe the typical lines of an input file 4 1 Mandatory parameters numevts 100000 number of events to be generated ihi 1 hadron 1 type 1 proton 1 antiproton ih2 1 hadron 2 type 1 proton 1 antiproton lhans1 10050 pdf set for hadron 1 LHAGLUE number lhans2 10050 pdf set for hadron 2 LHAGLUE number ndnsi 131 pdf for hadron 1 when using the hvqpdf package ndns2 131 pdf for hadron 2 when using the hvqpdf package The first entry is self explanatory The integers ih1 ih2 and lhans1 lhans2 or ndns1 ndns2 characterize instead the hadron type and the PDF set used by POWHEG BOX The hadron type in ih1 and ih2 can be 1 for a proton or 1 for an antiproton When in the Makefile t one of the 2 internal PDF sets is used see sec 2 In this case the set is chosen accordingly to the value of the tokens ndns1 ndns2 where the hvqpdf num one sets PDF native bering is assumed Otherwise if the Makefile variable PDF is set equal to Lhapdf then the program uses the LHAPDF library
Download Pdf Manuals
Related Search