Dell PowerEdge R720 White Paper
Contents
1. Required SPEC disclosure information R720 scores 1 235 223 ssj_ops and 303W 100 target load and 3 352 overall ssj_ops watt vs DL380p Gen8 1 217 238 ssj_ops and 324W 100 and 2 861 overall ssj_ops watt Comparison based on results by Dell Labs June 2012 SPEC and the benchmark name SPECpower_ssj are registered trademarks of the Standard Performance Evaluation Corporation For more information about SPECpower see www spec org power_ssj2008 Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Summary The results of the SPECpower_ssj2008 testing show similar performance between the two systems with the Dell PowerEdge R720 having a 1 advantage in raw performance over the HP ProLiant DL380p Gen8 However the ProLiant DL380p Gen8 has 15 lower power efficiency overall and 28 lower power efficiency at the important 70 load level This lower power efficiency led to the DL380p Gen8 drawing as much as 53 watts more compared to the similarly configured R720 The two socket 2U server form factor has become a staple in many data centers because these servers strike a good balance between expandability and rack density As energy costs rise power efficiency becomes an increasingly important factor in the decision to purchase these important data center components The greater performance lower power draw and higher performance watt ratio of the Dell PowerEdge R720 make it a comp2. Comparing Power Efficiency of the Dell PowerEdge R720 and HP ProLiant DL380p Gen Rack Servers This white paper compares the power efficiency of the Intel Xeon processor E5 2600 based Dell and HP 2U rack servers Brian Bassett Solutions Performance Analysis Dell Enterprise Solutions Group Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers This document is for informational purposes only and may contain typographical errors and technical inaccuracies The content is provided as is without express or implied warranties of any kind 2012 Dell Inc All rights reserved Dell and its affiliates cannot be responsible for errors or omissions in typography or photography Dell the Dell logo and PowerEdge are trademarks of Dell Inc Intel and Xeon are registered trademarks of Intel Corporation in the U S and other countries Microsoft Windows and Windows Server are either trademarks or registered trademarks of Microsoft Corporation in the United States and or other countries Other trademarks and trade names may be used in this document to refer to either the entities claiming the marks and names or their products Dell disclaims proprietary interest in the marks and names of others SPEC and the benchmark name SPECpower_ssj are registered trademarks of the Standard Performance Evaluation Corporation For more information about SPECpower see www spec org power_ssj2008 July 2012 Revi
3. DL380p Gen8 amp Total RAM in system GB 32 32 Vendor Samsung Samsung Type 2Rx8 PC3L 10600R 2Rx8 PC3L 10600R Speed MHz 1333 1333 Timing latency CAS 9 CAS 9 Number of RAM modules 4x8GB 4x8GB Hard disk Number of disks in system 2 2 Size GB 73 73 RPM 15 000 15 000 Type SAS 6 Gbps SAS 6 Gbps RAID type RAID 1 RAID 1 0 Controller Dell PERC H710P 1GB HP Smart Array P420i 1GB Operating system Name Microsoft Windows Server 2008 R2 Microsoft Windows Server Enterprise SP1 2008 R2 Enterprise SP1 Build number 7601 7601 File system NTFS NTFS Language English English Network adapter Vendor and model number Intel 1350 t Quad Port Gigabit Broadcom 331FLR Quad Port Gigabit Type Integrated Integrated On the HP Smart Array P400i RAID 1 0 is the only option available in a two hard drive configuration Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Appendix C Server firmware and drivers Table 3 Server firmware and drivers Driver firmware versions PowerEdge R720 Proliant DL380p Gen8 amp System BIOS 1 1 2 2012 02 25 Network drivers 11 14 44 7 0 6 0 HBA firmware 21 0 2 0001 2 14 HBA drivers 5 1 112 64 6 22 0 64 Video driver 2 4 1 0 6 12 1 1020 Integrated management 1 06 06 1 05 controller firmware Management controller driver N A 3 5 0 0 Power Efficiency Compa
4. M options were used on both servers as they are the best known JVM tunings for SPECpower_ssj2008 for the IBM J9 JVM Xms1024m Xmx1024m Xmn853m Xaggressive Xcompressedrefs XlockReservation Xnoloa XtlhPrefetch Xlp Xgcpolicy gencon The following bindings were used to ensure that each of the sixteen JVMs ran on 2 logical processors start affinity 3 C 30 C0 300 C00 3000 C000 30000 C0000 300000 C00000 3000000 C0000000 30000000 C0000000 Power and temperature configuration We used the Yokogawa WT210 Digital Power Meter for the actual power measurement of the servers as this is the most commonly used analyzer for SPECpower_ssj2008 publications at the time that this study was undertaken The WT210 unit used was calibrated in October of 2011 to ensure accurate power consumption measurements To ensure a fair comparison the systems were mounted in the same rack and inlet temperature was measured at the front of both systems during their runs using a Digi International Watchport H temperature probe As the attached Power _Temperature Details reports show the two systems were run in environments with a temperature difference of less than 1 degree Celsius Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Appendix B Server hardware configuration information Table 2 Server hardware configuration information Server hardware Memory modules PowerEdge R720 Proliant
5. T oie sieaisissns ccedie ei asa gis accra EEEE siante gente acess datas slema sree aie E elena siditia giana diene eva EEES 12 SPECpower ssj2008 configurations sicics sive ercins sivneaieies scewis eons ones einki EEE alenie ging NEKEA ESNEA econ oe 12 Power and temperature Configuration issccieccsevcrseraoeee nea cdveesse ones EE EREE edea nara catece 13 Appendix B Server hardware configuration information ssssssssssssssesssessessssseeesesssseeeeees 14 Appendix C Server firmware and rivers cssccnssenesssaccncsceescevcesasenesdsesdancs sedanesdsneeneacsesesees 15 Appendix D SPECpower_ssj2008 results sccsccssccsescsaccseccseccevsdsecenesdsecdsved aad ses sinceneccasceewes 16 Tables Table 1 Detailed configuration for power efficiency comparison esssssssssssssssssesesesesesesesesesese 5 Table 2 Server hardware configuration information sesesesesesesesesesesesesesesesesesesesesesesesesese 14 Table 3 Server firmware and drivers is cissssenssabersowsdstesoteasnerenaeande naan seus aiedeewewbas seevonsdebwesigs 15 Figures Figure 1 Normalized SPECpower_ssj2008 results for PowerEdge R720 and ProLiant DL380p Gen 6 Figure 2 Performance per watt ratios for all target loads ssesssessssssssssssoosesssssssosssssssooosesee 7 Figure 3 Power utilization at 70 target load eesesesesesesesesesesesesesesesesesesesesesesesesesesesesesese 8 Figure 4 Power utilization at active idle ici sces ceva avvaesaa
6. avee seeaavea bees onda eda axea seed onde EEE EEEE E 9 Figure 5 SPECpower_ssj2008 results for PowerEdge R720 sssssssrsssossssesssoosssessososesssessoosesee 16 Figure 6 SPECpower_ssj2008 results for ProLiant DL380p Gen8 sce eee eee ee eee eee eee eeeeeeeeaees 17 iii Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Executive summary With power and cooling costs accounting for increasingly large portions of IT budgets IT departments looking to minimize total cost of ownership TCO are making power efficiency a priority when choosing server hardware In this white paper we examine the power efficiency of two rack servers in the two processor 2U form factor a form factor that is one of the most popular in many data centers because of the balance it strikes between density and expandability Introduction Dell commissioned its Solutions Performance Analysis team to compare the Dell PowerEdge R720 and the HP ProLiant DL380p Gen8 both of which support the Intel Xeon processor E5 2600 product family Using the industry standard SPECpower_ssj2008 benchmark the two servers were tested for performance power draw and performance watt with the hardware configured as similarly as possible to ensure an apples to apples comparison The results showed the Dell PowerEdge R720 consumed less power in the tested configuration while delivering better raw performance a
7. elling solution compared to the HP ProLiant DL380p Gen8 Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Appendix A Test methodology SPECpower_ssj2008 standard SPECpower_ssj2008 is an industry standard benchmark created by the Standard Performance Evaluation Corporation SPEC to measure a server s power and performance across multiple utilization levels SPECpower_ssj2008 consists of a Server Side Java SSJ workload along with data collection and control services SPECpower_ssj2008 results portray the server s performance in ssj_ops server side Java operations per second divided by the power used in watts ssj_ops watt SPEC created SPEcpower_ssj2008 for those who want to accurately measure the power consumption of their server in relation to the performance that the server is capable of achieving with ssj2008 workload SPECpower_ssj2008 consists of three main software components e Server Side Java SSJ Workload Java database that stresses the processors caches and memory of the system as well as software elements such as OS elements and the Java implementation chosen to run the benchmark e Power and Temperature Daemon PTDaemon Program that controls and reports the power analyzer and temperature sensor data e Control and Collect System CCS Java program that coordinates the collection of all the data For more information on how SPECpower_ssj008 works see http w
8. gher is better 4 214 4 210 3 752 3 787 3 539 3 491 3 319 3 108 93 010 ik 0263 009 2 415 1 947 1 784 1 080 949 E z v Q N Q a a N O o N n n a v z fo jos Q w A n 60 Target Load E Dell PowerEdge R720 m HP ProLiant DL380p Gens Figure 2 also shows that the PowerEdge R720 has a higher power efficiency ratio at all target loads When IT professionals are sizing their data center needs a target CPU utilization of 70 is often recommended to allow extra capacity for load peaks This makes that interval particularly relevant for evaluating power efficiency At the 70 target load interval the R720 s ratio of 3 975 ssj_ops watt gives it a 28 power efficiency advantage over the ProLiant DL380p Gen8 with its ratio of 3 108 Required SPEC disclosure information R720 scores 1 235 223 ssj_ops and 303W 100 target load and 3 352 overall ssj_ops watt vs DL380p Gen8 1 217 238 ssj_ops and 324W 100 and 2 861 overall ssj_ops watt Comparison based on results by Dell Labs June 2012 SPEC and the benchmark name SPECpower_ssj are registered trademarks of the Standard Performance Evaluation Corporation For more information about SPECpower see www spec org power_ssj2008 Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers In Figure 3 the power draw of both servers at the 70 load level is shown Figure 3 Power utilization a
9. n Hewlett Packard Inc ProLiant DL380p SPECpower_ssj2008 2 861 overall Gens ssj_ops watt Dell Inc SPEC License Single Node Dell Inc Test Location io Test Date Jun 7 2012 Jun 2012 Software 45 9011 Publication Unpublished Single Line Supplier powered WARNING PTDaemon 1 4 1 1271fb18 201 10728 is over 6 months old Check http Avww spec org power docs SPECpower Device_List html to determine if a newer version is available see Run Rules Section 1 1 WARNING PTDaemon 1 4 1 1271fb18 201 10728 is over 6 months old Check http Avww spec org power docs SPECpower Device_List html to determine if a newer version is available see Run Rules Section 1 1 856 350 Target Load 246 989 100 150 200 250 300 Average Active Power W file WiZ BrianB Cornpe titrve R720 vs DL320p gen3 data DL330p gen3 J9 runs ssj 0043 DL320p Gen3 J9 2 86 1 ssj 0043 main html 7 17 2012 7 57 18 PJ
10. nagement rather than leaving that function to the operating system For the Dell PowerEdge R720 the following settings were used e Disabled Adjacent Sector Prefetcher in BIOS e Disabled Hardware Prefetch in BIOS e Disabled DCU Streamer Prefetcher in BIOS e System Profile set to Custom Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers e CPU Power Management set to DAPC e Memory Frequency set to Maximum Performance e Intel Turbo Boost Enabled e C1E Enabled e C States Enabled e Memory Patrol Scrub set to Disabled e Memory Refresh Rate set to 1x e Memory Operating Voltage set to Auto As of the date the testing was performed HP had not published any SPECpower results for their Gen8 servers so BIOS settings used for the DL380p Gen8 were based on the settings used in the published SPECpower result for the DL380 G7 For the HP ProLiant DL380p Gen the following settings were used e Disabled HW Prefetch in BIOS e Disabled Adjacent Sector Prefetch in BIOS e Disabled DCU Prefetcher in BIOS e HP Power Regulator set to HP Dynamic Power Savings Mode e Energy Performance Bias set to Balanced Performance e Dynamic Power Savings Mode Response set to Slow in BIOS e Disabled Collaborative Power Control in BIOS e Intel Turbo Boost enabled e Minimum Processor Idle Power Core State set to C6 State e Minimum Processor Idle Power Package State set to Package C6 non retention State OS t
11. nd greater power efficiency Key findings Performance watt The HP ProLiant DL380p Gen8 achieved a 15 lower performance to power ratio overall than in the Dell PowerEdge R720 in the tested configuration At a 70 target load the R720 had a 28 higher performance to power ratio Performance Despite drawing less power the PowerEdge R720 provided slightly higher raw performance than the ProLiant DL380p Gens Power In this apples to apples configuration the ProLiant DL380p Gen8 consumed 7 more power at 100 target load and 10 more power at idle compared to the PowerEdge R720 Test methodology and detailed results are documented in this paper Methodology SPECpower_ssj2008 is an industry standard benchmark created by the Standard Performance Evaluation Corporation SPEC to measure a server s power and performance across multiple utilization levels Appendix A Test methodology details the test methodology used by Dell Appendix B Server hardware configuration information and Appendix C Server firmware and drivers provide detailed configurations for the tests and Appendix D SPECpower_ssj2008 results provides detailed report data that supports the results in this paper Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Apples to apples configuration The two systems were configured as similarly as possible with similar hard drive memory and power supply configurati
12. ons To eliminate any possible power efficiency variations between processors tests on both systems were run using the same pair of Intel Xeon E5 2660 processors Storage controllers were matched as closely as possible between the two systems and both systems were equipped with two platinum rated 750 watt power supply units PSUs The configuration used is summarized in Table 1 Table 1 Detailed configuration for power efficiency comparison Configuration Dell PowerEdge R720 HP Proliant DL380p Gen8 Sockets form factor 2S8 2U 2S 2U Processors 2 x Intel Xeon E5 2660 2 x Intel Xeon E5 2660 Physical logical cores 16 32 16 32 Memory 4 x 8GB Dual Ranked PC3L 10600R LV RDIMMs Dell P N OP9RN2 4 x 8GB Dual Ranked PC3L 10600R LV RDIMMs Part number 647897 B21 Hard drives 2 x 73GB 15k 6Gb RAID 1 Dell P N OXT764 2 x 73GB 15k 6Gb RAID 1 0 HP P N 652597 B21 Storage controller Dell PERC H710P 1GB cache HP Smart Array P420i 1GB cache Power supply quantity rating 2x 750 W DP N OSNF18 2 x 750 W part number 643955 101 Option Kit number 656363 B21 Both systems were equipped with Low Voltage Registered DIMMs LV RDIMMs purchased with the respective systems and running at 1 35 volts The systems used factory default BIOS settings except as noted in Appendix A Test methodology 1 On the HP Smart Array P420i RAID 1 0 is the only option available in a two hard dri
13. rison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Appendix D SPECpower_ssj2008 results Figure 5 SPECpower_ssj2008 results for PowerEdge R720 SPECpower_ssj2008 SPEC power_ssj2008 Copyright 2007 2012 Standard Performance Evaluation Corporation Dell Inc PowerEdge R720 SPECpower_ssj2008 3 352 overall ssj_ops watt Dell Inc SPEC License Single Node Dell Inc Test Location iow Rock TX Test Date Jun 8 2012 Mar 2012 Feb 2011 Publication Unpublished Single Line Supplier powered WARNING PTDaemon 1 4 1 1271 fo18 201 10728 is over 6 months old Check http Avww spec org power docs SPECpower Device_List html to determine if a newer version is available see Run Rules Section 1 1 WARNING PTDaemon 1 4 1 1271 fb18 201 10728 is over 6 months old Check http Avww spec org power docs SPECpower Device_List html to determine if a newer version is available see Run Rules Section 1 1 90 89 6 1 124 925 861 679 Target Load 249 933 i 100 150 200 250 300 Average Active Power W file WiZ BrianB Corpe titrve R720 vs DL320p gen3 data R720 J9 runs ssj 0042 3 352 ssj 0042 main html 7 17 2012 7 09 02 PMJ Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Figure 6 SPECpower_ssj2008 results for ProLiant DL380p Gen8 SPECpower_ssj2008 SPEC power_ssj2008 Copyright 2007 2012 Standard Performance Evaluation Corporatio
14. sion 1 01 Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Contents EX CULIVE SUIMIMANY S dscdaxeccsdaadisanscasead shoantcasimanensanoen sas anoaanocd EEE AA EA EEA EEEE RAAR 4 INMEFOGUCUON carora raai ate smsieaauin KARE San nealetesine wate AAE EEREN seuss 4 KEY HINGINGS A A E ad casesiwawieae A anaes cudandssusacieusscudendestsdeteseledeiavetneetese 4 Performance WAL site iausredonsaitdenis Sanaraned ete caetanacsdadsad NEEE E E EAE ARE EERE ES 4 PENTOFMANGE sonerii niaii a Ei A TA EA E EEEE EE ES 4 POWER visccecns scvsis ririki ti EnEn EEEE EEEE EEEE EEEE EEEE 4 INGE OLOGY te scesarecasstrereveirasssaigrereienatecorsaso cvaracosaescreseise s n aig ote asecsryaierc EAEE EEEE 4 Apples to apples configurations sss ceive csarccee esse cedessaecceeeedesevaseas seveendeaevasueeseveseveesvarsee sean 5 RESULTS eirate sanietaegnaedee cna sneesdiandeae eaenenyeeded one eneeeEee open bee NEEE NEEE EERE 6 SUMMAN sc cccscdiasasicasinads EAEE doscamedanca simaemosasincaswadsinedens damedenesamed snseumeasandnnae noes aneasved aed sae 10 Appendix A Test methodology iscsi coscccedsehacend cadacanicekwcand cagacaneiehacand cadacandcebacandcagacenece hace 11 SPECPOWer SSJ2Z2008 standard vssiuicssseassnessusaseee daces ei eanded etn dedinedn dudes eau dediesiens dues besssunesieeones 11 BIOS SCCUINGS fis ccnscencauasteedansucaassasstaecameueaaeseesetecemsueeees ease eeeubeeeuesesenseeembeetuesmesesieentens 11 OS CUN
15. t 70 target load SPECpower results power utilization at 70 target load 275 Power draw in watts lower is better m Dell PowerEdge R720 m HP ProLiant DL380p Gen8 At this interval the HP ProLiant DL380p Gen8 consumed 53 watts more on average 28 higher than the Dell PowerEdge This additional power utilization of the HP server caused its overall SPECpower_ssj2008 overall ops watt to be worse than the Dell PowerEdge R720 Required SPEC disclosure information R720 scores 1 235 223 ssj_ops and 303W 100 target load and 3 352 overall ssj_ops watt vs DL380p Gen8 1 217 238 ssj_ops and 324W 100 and 2 861 overall ssj_ops watt Comparison based on results by Dell Labs June 2012 SPEC and the benchmark name SPECpower_ssj are registered trademarks of the Standard Performance Evaluation Corporation For more information about SPECpower see www spec org power_ssj2008 Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers SPECpower_ssj2008 also includes a measurement of power while the servers are at 0 target utilization Active Idle The ProLiant DL380p Gen8 drew 17 5W 10 more than the PowerEdge R720 when both systems were idle Figure 4 Power utilization at active idle SPECpower results Power utilization at 0 target load active idle in watts lower is better Power draw in watts lower is better E Dell PowerEdge R720 E HP ProLiant DL380p Gen8
16. uning To improve Java performance large pages were enabled by entering Control Panel gt Administrative Tools gt Local Security Policy gt Local Policies gt User Rights Assignment gt Lock Pages in Memory An option was changed to add Administrator Operating System Power Management mode for both solutions was changed to Power Saver For the ProLiant DL380p Gen8 the Minimum Processor state was changed to 0 and the Maximum Processor State was set to 100 these settings are common tunings for the SPECpower_ssj2008 benchmark when the system is running in Power Saver mode This change was not made on the PowerEdge R720 since the Minimum and Maximum Processor State settings not available for adjustment on the R720 when it is set to Dell Active Power Controller DAPC mode Both solutions were edited to turn off the Hard Drive and display after 1 minute We configured both servers with a separate IP address on the same subnet as our SPECpower_ssj2008 controller system where the Director CCS and PTDaemon components were located and connected both servers directly to the controller system through NIC 1 for their respective runs SPECpower_ssj2008 configuration The IBM J9 Java Virtual Machine JVM was used for both solutions gt JVM build 2 4 J2RE 1 6 0 IBM J9 2 4 Windows Server 2008 amd64 64 jvmwa64 60sr5 20090519_ 35743 Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers The following JV
17. ve configuration Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers Results In this like for like comparison the Dell PowerEdge R720 drew less power than the HP ProLiant DL380p Gen8 at idle at 100 target load and at all points between while maintaining a slight performance advantage at 100 target workload This lead to the 15 lower performance watt ratio for the DL380p Gen8 Figure 1 Normalized SPECpower_ssj2008 results for PowerEdge R720 and ProLiant DL380p Gen8 1 20 1 00 0 80 The DL380p Gen8 consumes Performance Watt 0 60 7 more power at 100 load ratio of the DL380p and 10 more power at idle Gen8 is 15 worse 0 40 0 20 Maximum Power Draw Idle Power Draw Maximum Performance Performance Watt ratio lower is better lower is better higher is better higher is better m Dell PowerEdge R720 m HP ProLiant DL380p Gen8 Power Efficiency Comparison of the Dell PowerEdge R720 and HP ProLiant DL380p Gen8 Rack Servers SPECpower_ssj2008 includes a measurement of power while the servers are at varying levels of target utilization The performance to watt ratio at each target load level is total operations ssj_ops divided by average power consumption of the server at that load level Figure 2 shows power efficiency at each interval Figure 2 Performance per watt ratios for all target loads SPECpower results Performance to power ratio across load levels hi
18. ww spec org power_ssj2008 All results discussed in this whitepaper are from compliant runs in SPEC terminology which means that although they have not been submitted to SPEC for review Dell is allowed to disclose them for the purpose of this study All configuration details required to reproduce these results are listed in Appendices A B and C all result files from the runs compared are included in Appendix D SPECpower_ssj2008 results Both servers were configured by installing a fresh copy of Microsoft Windows Server 2008 Enterprise R2 Service Pack 1 and the operating system install with a two drive RAID 1 choosing the full installation option for each The latest driver and firmware update packages available to both servers were installed at the beginning of this study Refer to Appendix B Server hardware configuration information for details BIOS settings Available BIOS settings differed between the two manufacturers Settings common to the two platforms were set identically and settings not common to the two platforms but known to affect power efficiency were set to the best known setting for each platform For both servers Intel Turbo Boost was enabled by default and memory speed was left at the default value of 1333 MHz Prefetchers built into the Xeon E5 2660 chips were disabled on both systems a common tuning for the SPECpower_ssj2008 benchmark Both systems defaulted to handling their own power ma
Download Pdf Manuals
Related Search