Ib Msystemsperformancebenchmarks
- 1. IBM Power Systems Performance Report POWER7, POWER6 and POWER5 results February 9, 2010
- 2. Table of Contents Performance of IBM UNIX, IBM i and Linux Operating System Servers ................................... 3 Section 1 - AIX SPEC CPU2006 HPC Performance ......................................................................... 4 Section 1a - AIX SPEC CPU2000 Performance ................................................................................ 4 Section 2 - AIX Multiuser Performance (rPerf, SPEC CPU2006).................................................. 5 Section 2a - AIX Multiuser Performance (rPerf, SPEC CPU2000, SPECweb99) ...................... 7 Section 2b - Power Sytems Multiuser Performance using AIX 5L V5.2 .................................. 10 Section 2c - AIX Capacity Upgrade on Demand Relative Performance Guidelines ........... 122 Section 2d - CPW Published Results………………………………………………………………18 Section 3- TPC-C Version 5 Published Results ............................................................................. 19 Section 4 - TPC-H Published Results ............................................................................................... 20 Section 5 - AIX SPECsfs97_R1 Benchmark Results..................................................................... 20 Section 5a - AIX NotesBench Published Results .......................................................................... 20 Section 6 - AIX Java Benchmarks (SPECjvm98, SPECjbb2000, SPECjbb2005) Published Results .................................................................................................................................................................... 21 Section 6a - IBM i Java Benchmarks (SPECjbb2005) Published Results………..…………..21 Section 6b - AIX SPECjAppServer2004 Published Results…………………………………….21 Section 7 - SAP Standard Application Benchmarks Published Results ............................. 2222 Section 8 - AIX PeopleSoft Benchmarks Published Results...................................................... 23 Section 9 - AIX Oracle e-Business Suite (eBS) Benchmarks Published Results ................. 24 Section 10 - AIX Siebel Benchmarks Published Results............................................................. 24 Section 11 - AIX Sybase Benchmarks Published Results .......................................................... 25 Section 12 - AIX Manugistics Benchmarks Published Results ................................................. 25 Section 13 - Linux Published Benchmark Results........................................................................ 25 Section 14 - Historical Multiuser Performance .............................................................................. 29 Notes on Performance Benchmarks and Values........................................................................... 34 Page 2
- 3. Performance of IBM UNIX, IBM i and Linux Operating System Servers February 9, 2010 This document contains performance and benchmark results for IBM servers and workstations running the UNIX® (AIX®), IBM i and Linux® operating systems. This includes the IBM Power™ Systems servers (System p™, System p5™, eServer™ p5, pSeries®, OpenPower® and IBM RS/6000®; BladeCenter® Power Architecture® technology-based blades) and IntelliStation® POWER™ workstations. This document contains the results for the new POWER7 systems. Section One of this report includes the SPEC CPU2006 results. SPEC CPU2000 results are presented in Section 1a. Section Two is multiuser performance. The rPerf, CPW and SPEC_rate2006 are presented in this section. Multiuser Performance of SPEC CPU2000 along with rPerf and SPECweb99 are presented in Section 2a. Multiuser performance using AIX® V5.2 is presented in Section 2b. Capacity Upgrade on Demand relative performance guidelines are presented in Section 2c. Section 2d of this report includes CPW benchmark information provided for new POWER6 processor-based servers running the IBM i operating system. Section Three presents the TPC-C version 5 results. Section Four provides published TPC-H results. Section Five reflects the published SPECsfs97 benchmark results. The NotesBench results to date are presented in Section 5a. Section Six reflects the published SPECjvm98, SPECjbb2000 and SPECjbb2005 Java™ benchmarks. Section Seven through Twelve include published application performance benchmarks for SAP, PeopleSoft, Oracle Applications, Seibel, Sybase and Manugistics. Section Thirteen contains Linux operating system performance results. Section Fourteen is a historical list of commercial performance estimates for IBM System p™, RS/6000 models and RS/6000 SP™ nodes that have been withdrawn from marketing. IBM has discontinued Relative OLTP results. All performance measurements for the IBM System p, IBM System p5, IBM eServer p5, IBM eServer pSeries, IBM RS/6000 servers, IntelliStation POWER workstations and BladeCenter blades were made with systems running the AIX operating system unless otherwise indicated to have used Linux. For new and upgraded systems, AIX Version 4.3, AIX 5 or AIX 6 were used. All other systems used previous versions of AIX. Footnotes used in following tables: # - System has been announced as withdrawn from marketing; * - Submitted to SPEC, waiting review; e – estimate; n – new; u – upgrade; N/A - not available; P5 – IBM POWER5™; P5+ – IBM POWER5+™; P6 – IBM POWER6™; P6+ – IBM POWER6+™; P7 – POWER7™, IS – IntelliStation, OP – OpenPower, PPC – PowerPC® Changes from last version are highlighted in yellow. Page 3
- 4. Section 1 - AIX SPEC CPU2006 Performance SPEC SPEC L1 L2/L3 SPEC int_ SPEC fp_ Processor Cache Cache int base fp base Model / # Cores GHz (KB) (MB) 2006 2006 2006 2006 520 P6/1 4.2 64/64 8/- -- -- -- -- 520 P6/4 4.2 64/64 16/- -- -- -- -- 550 P6/1 4.2 64/64 8/32 -- -- -- -- 550 P6/8 4.2 64/64 32/128 -- -- -- -- 570 P6/1 4.7 64/64 8/32 21.6 17.8 22.3 18.7 570 P6/4 4.7 64/64 16/64 -- -- -- -- 570 P6/8 4.7 64/64 32/128 -- -- -- -- 570 P6/16 4.7 64/64 64/256 -- -- -- -- 575 P6/1 4.7 64/64 32/128 -- -- -- -- 575 P6/32 4.7 64/64 128/512 -- -- -- -- 595 P6/1 5.0 64/64 32/128 -- -- 24.9 20.1 595 P6/64 5.0 64/64 256/1024 -- -- -- -- Section 1a - AIX SPEC CPU2000 Performance SPEC SPEC L1 L2/L3 SPEC int_ SPEC fp_ Processor Cache Cache int base Fp base Model / # Cores MHz (KB) (MB) 2000 2000 2000 2000 JS20 PPC970/1 2200 64/32 0.5/- 1,040 986 1,241 1,178 JS20 PPC970/2 2200 64/32 1.0/- -- -- -- -- JS21 PPC970+/1 2500 64/32 1.0/- 1,587 1,509 2,119 1,936 JS21 PPC970+/4 2500 64/32 4.0/- -- -- -- -- JS21 PPC970+/1 2700 64/32 1.0/- 1,706 1,623 2,259 2,060 JS21 PPC970+/2 2700 64/32 2.0/- -- -- -- -- #IS-285 P5+/1 1900 64/32 1.9/36 1,512 1,469 3,027 2,838 #IS-285 P5+/2 1900 64/32 1.9/36 -- -- -- -- IS-285 P5+/1 2100 64/32 1.9/36 1,747 1,670 3,324 3,100 IS-285 P5+/2 2100 64/32 1.9/36 -- -- -- -- #p5-505 P5/1 1650 64/32 1.9/0 1,297 1,259 2,528 2,390 #p5-505 P5/2 1650 64/32 1.9/36 -- -- -- -- p5-505 P5+/1 2100 64/32 1.9/36 1,704 1,617 3,301 3,057 p5-505Q P5+/1 1650 64/32 1.9/72 1,371 1,311 2,610 2,442 #p5-510 P5/1 1650 64/32 1.9/36 1,260 1,203 2,236 2,071 #p5-510 P5/2 1650 64/32 1.9/36 -- -- -- -- #p5-510 P5+/1 1900 64/32 1.9/36 1,536 1,479 3,048 2,850 p5-510 P5+/1 2100 64/32 1.9/36 1,704 1,617 3,301 3,057 #p5-510Q P5+/1 1500 64/32 1.9/72 1,231 1,164 2,377 2,217 p5-510Q P5+/1 1650 64/32 1.9/72 1,371 1,311 2,610 2,442 #p5-520 P5/1 1500 64/32 1.9/0 -- -- 2,041 1,909 #p5-520 P5/2 1500 64/32 1.9/36 -- -- -- -- #p5-520 P5/1 1650 64/32 1.9/36 1,248 1,201 2,138 2,034 #p5-520 P5/2 1650 64/32 1.9/36 -- -- -- -- #p5-520 P5+/1 1650 64/32 1.9/0 1,337 1,288 2,676 2,502 #p5-520 P5+/1 1900 64/32 1.9/36 1,513 1,470 3,030 2,839 #p5-520 P5+/2 1900 64/32 1.9/36 -- -- -- -- p5-520 P5+/1 2100 64/32 1.9/36 1,704 1,617 3,301 3,057 p5-520Q P5+/1 1650 64/32 1.9/72 1,371 1,311 2,610 2,442 #p5-550 P5/1 1500 64/32 3.8/72 -- -- 2,072 1,914 #p5-550 P5/4 1500 64/32 3.8/72 -- -- -- -- Page 4
- 5. SPEC SPEC L1 L2/L3 SPEC int_ SPEC fp_ Processor Cache Cache int base Fp base Model / # Cores MHz (KB) (MB) 2000 2000 2000 2000 #p5-550 P5/1 1650 64/32 1.9/36 1,248 1,200 2,221 2,121 #p5-550 P5/4 1650 64/32 3.8/72 -- -- -- -- #p5-550 P5+/1 1650 64/32 1.9/36 1,336 1,288 2,657 2,483 p5-550 P5+/1 1900 64/32 1.9/36 1,510 1,467 3,007 2,815 p5-550 P5+/4 1900 64/32 3.8/72 -- -- -- -- p5-550 P5+/1 2100 64/32 1.9/36 1,743 1,669 3,321 3,125 p5-550 P5+/4 2100 64/32 3.8/72 -- -- 4,051 3,210 #p5-550Q P5+/1 1500 64/32 1.9/72 1,187 1,156 2,263 2,179 #p5-550Q P5+/8 1500 64/32 7.6/144 -- -- -- -- p5-550Q P5+/1 1650 64/32 1.9/72 1,367 1,307 2,612 2,458 p5-550Q P5+/8 1650 64/32 7.6/144 -- -- -- -- p5-560Q P5+/1 1500 64/32 1.9/72 1,204 1,160 2,360 2,197 p5-560Q P5+/16 1500 64/32 15.2/288 -- -- -- -- p5-560Q P5+/1 1800 64/32 1.9/72 -- -- -- -- p5-560Q P5+/16 1800 64/32 15.2/288 -- -- -- -- #p5-570 P5/1 1900 64/32 1.9/36 1,452 1,398 2,702 2,576 #p5-570 P5/4 1900 64/32 3.8/72 -- -- -- -- #p5-570 P5/8 1900 64/32 7.6/144 -- -- -- -- #p5-570 P5/16 1900 64/32 15.2/288 -- -- -- -- #p5-575 P5/1 1500 64/32 1.9/36 1,143 1,087 2,185 2,050 #p5-575 P5/16 1500 64/32 15.2/288 -- -- -- -- #p5-575 P5/1 1900 64/32 1.9/36 1,456 1,385 2,600 2,413 #p5-575 P5/8 1900 64/32 15.2/288 -- -- -- -- p5-575 P5+/1 1900 64/32 1.9/36 1,526 1,473 3,042 2,830 p5-575 P5+/16 1900 64/32 15.2/288 -- -- -- -- p5-575 P5+/1 2200 64/32 1.9/36 1,765 1,705 3,513 3,271 p5-575 P5+/8 2200 64/32 15.2/288 -- -- -- -- #p5-590 P5/1 1650 64/32 1.9/144 1,259 1,200 2,450 2,276 #p5-590 P5/32 1650 64/32 30.4/576 -- -- -- -- #p5-595 P5/1 1900 64/32 1.9/144 1,452 1,392 2,796 2,585 #p5-595 P5/64 1900 64/32 60.8/1152 -- -- -- -- p5-595 P5+/1 2300 64/32 1.9/144 1,900 1,820 3,642 3,369 Section 2 - AIX Multiuser Performance (rPerf, SPEC CPU2006) Note : upto 5% additional rPerf with Intelligent Energy Optimzation enabled depending on the system SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor Cache Cache rate base rate base Model / # Cores GHz (KB) (MB) rPerf 2006 2006 2006 2006 JS12 P6/2 3.8 64/64 8/- 14.71 -- -- -- -- JS22 P6/4 4.0 64/64 8/- 30.26 84.7 77.8 75.6 67.8 JS23 P6+/4 4.2 64/64 16/64 36.28 -- -- -- -- JS43 P6+/8 4.2 64/64 32/128 68.20 -- -- -- -- 520 P6/1 4.2 64/64 8/- 8.39 -- -- -- -- 520 P6/2 4.2 64/64 8/- 15.95 -- -- -- -- 520 P6/4 4.2 64/64 16/- 31.48 90.6 82.3 80.8 71.4 520 P6+/2 4.7 64/64 8/32 20.13 -- -- -- -- 520 P6+/4 4.7 64/64 16/64 39.73 124 101 105 88.7 550 P6/2 3.5 64/64 8/32 15.85 -- -- -- -- 550 P6/4 3.5 64/64 16/64 31.27 -- -- -- -- 550 P6/6 3.5 64/64 24/96 45.04 -- -- -- -- 550 P6/8 3.5 64/64 32/128 58.80 179 152 154 135 Page 5
- 6. SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor Cache Cache rate base rate base Model / # Cores GHz (KB) (MB) rPerf 2006 2006 2006 2006 550 P6/2 4.2 64/64 8/32 18.38 -- -- -- -- 550 P6/4 4.2 64/64 16/64 36.28 -- -- -- -- 550 P6/6 4.2 64/64 24/96 52.24 -- -- -- -- 550 P6/8 4.2 64/64 32/128 68.20 212 179 178 156 550 P6+/2 5.0 64/64 8/32 21.18 -- -- -- -- 550 P6+/4 5.0 64/64 16/64 41.81 -- -- -- -- 550 P6+/6 5.0 64/64 24/96 60.20 -- -- -- -- 550 P6+/8 5.0 64/64 32/128 78.60 263 215 222 188 560 P6+/4 3.6 64/64 16/64 31.32 -- -- -- -- 560 P6+/8 3.6 64/64 32/128 57.32 -- -- -- -- 560 P6+/16 3.6 64/64 64/256 100.30 363 289 263 226 750 P7/6 3.30 32/32 1.5/24 70.07 -- -- -- -- 750 P7/12 3.30 32/32 3/48 134.54 -- -- -- -- 750 P7/24 3.30 32/32 6/96 252.26 -- -- -- -- 750 P7/8 3.00 32/32 2/32 81.24 -- -- -- -- 750 P7/16 3.00 32/32 4/64 155.99 -- -- -- -- 750 P7/32 3.00 32/32 8/128 292.47 -- -- -- -- 750 P7/8 3.30 32/32 2/32 86.99 -- -- -- -- 750 P7/16 3.30 32/32 4/64 167.01 -- -- -- -- 750 P7/32 3.30 32/32 8/128 313.15 1010 911 825 750 750 P7/32 3.55 32/32 8/128 331.06 1060 949 851 776 755 P7/32 3.30 32/32 8/128 - 1010 911 825 750 570 P6/2 3.5 64/64 8/32 15.85 -- -- -- -- 570 P6/4 3.5 64/64 16/64 31.69 -- -- -- -- 570 P6/8 3.5 64/64 32/128 58.95 -- -- -- -- 570 P6/12 3.5 64/64 48/192 83.35 -- -- -- -- 570 P6/16 3.5 64/64 64/256 105.75 -- -- -- -- 570 P6/2 4.2 64/64 8/32 18.38 -- -- -- -- 570 P6/4 4.2 64/64 16/64 36.76 -- -- -- -- 570 P6/8 4.2 64/64 32/128 68.38 -- -- -- -- 570 P6/12 4.2 64/64 48/192 96.68 -- -- -- -- 570 P6/16 4.2 64/64 64/256 122.67 -- -- -- -- 570/32 P6+/4 4.2 64/64 16/64 35.50 -- -- -- -- 570/32 P6+/8 4.2 64/64 32/128 64.96 -- -- -- -- 570/32 P6+/16 4.2 64/64 64/256 113.68 -- -- -- -- 570/32 P6+/24 4.2 64/64 96/384 153.46 -- -- -- -- 570/32 P6+/32 4.2 64/64 128/512 193.25 832 661 602 517 570 P6+/2 4.4 64/64 8/32 19.08 -- -- -- -- 570 P6+/4 4.4 64/64 16/64 38.16 -- -- -- -- 570 P6+/8 4.4 64/64 32/128 70.97 -- -- -- -- 570 P6+/12 4.4 64/64 48/192 100.35 -- -- -- -- 570 P6+/16 4.4 64/64 64/256 127.32 -- -- -- -- 570 P6/2 4.7 64/64 8/32 20.13 60.9 53.2 58.0 51.5 570 P6/4 4.7 64/64 16/64 40.26 122 106 115 102 570 P6/8 4.7 64/64 32/128 74.89 240 206 213 189 570 P6/12 4.7 64/64 48/192 105.89 -- -- -- -- 570 P6/16 4.7 64/64 64/256 134.35 478 410 426 379 570 P6+/2 5.0 64/64 8/32 21.16 -- -- -- -- 570 P6+/4 5.0 64/64 16/64 42.32 -- -- -- -- 570 P6+/8 5.0 64/64 32/128 78.71 -- -- -- -- 570 P6+/12 5.0 64/64 48/192 111.3 -- -- -- -- 570 P6+/16 5.0 64/64 64/256 141.21 542 466 544 465 575 P6/32 4.7 64/64 128/512 -- 934 812 839 730 Page 6
- 7. SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor Cache Cache rate base rate base Model / # Cores GHz (KB) (MB) rPerf 2006 2006 2006 2006 770 P7/6 3.50 32/32 1.5/24 72.55 -- -- -- -- 770 P7/12 3.50 32/32 3/48 140.75 -- -- -- -- 770 P7/24 3.50 32/32 6/96 261.19 -- -- -- -- 770 P7/48 3.50 32/32 12/192 493.37 1,930 1,740 1,760 1,560 770 P7/8 3.10 32/32 2/32 85.20 -- -- -- -- 770 P7/16 3.10 32/32 4/64 165.30 -- -- -- -- 770 P7/32 3.10 32/32 8/128 306.74 -- -- -- -- 770 P7/64 3.10 32/32 16/256 579.39 2,140 1,930 1,900 1,710 780 P7/8 3.86 32/32 2/32 100.75 -- -- -- -- 780 P7/16 3.86 32/32 4/64 195.45 652 586 586 531 780 P7/32 3.86 32/32 8/128 362.70 -- -- -- -- 780 P7/64 3.86 32/32 16/256 685.09 2,530 2,300 2,240 2,030 780 P7/8 4.14 32/32 4/64 115.86 -- -- -- -- 780 P7/16 4.14 32/32 8/128 226.97 -- -- -- -- 780 P7/32 4.14 32/32 16/256 425.50 1,460 1,300 1,300 1,190 595 P6/8 4.2 64/64 32/128 75.58 -- -- -- -- 595 P6/16 4.2 64/64 64/256 142.90 -- -- -- -- 595 P6/32 4.2 64/64 128/512 266.51 -- -- -- -- 595 P6/48 4.2 64/64 192/768 373.60 -- -- -- -- 595 P6/64 4.2 64/64 256/1024 479.89 1650 1420 -- -- 595 P6/8 5.0 64/64 32/128 87.10 -- -- -- -- 595 P6/16 5.0 64/64 64/256 164.67 -- -- -- -- 595 P6/32 5.0 64/64 128/512 307.12 -- -- -- -- 595 P6/48 5.0 64/64 192/768 430.53 -- -- -- -- 595 P6/64 5.0 64/64 256/1024 553.01 2083 1822 2108 1822 Note : upto 5% additional rPerf with Intelligent Energy Optimzation enabled depending on the system Section 2a - AIX Multiuser Performance (rPerf, SPEC CPU2000, SPECweb99) SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor Cache Cache rate base rate base SPEC Model / # Cores MHz (KB) (MB) rPerf 2000 2000 2000 2000 web99 JS20 PPC970/1 1600 64/32 0.5/- 1.53 -- -- -- -- -- JS20 PPC970/2 1600 64/32 1.0/- 2.65 -- -- -- -- -- JS20 PPC970/1 2200 64/32 0.5/- 1.95 -- -- -- -- -- JS20 PPC970/2 2200 64/32 1.0/- 3.40 21.5 20.2 20.0 19.2 -- JS21 PPC970+/2 2700 64/32 2.0/- 5.31 38.5 36.6 43.9 40.9 -- JS21 PPC970+/4 2500 64/32 4.0/- 8.72 67.5 64.2 58.8 56.1 -- JS21 PPC970+/4 2300 64/32 4.0/- 8.15 -- -- -- -- -- #IS-285 P5+/2 1900 64/32 1.9/36 -- 39.6 38.8 67.6 65.4 -- IS-285 P5+/2 2100 64/32 1.9/36 -- 45.0 43.9 72.9 70.0 -- #p5-505 P5/2 1500 64/32 1.9/36 9.13 -- -- -- -- -- #p5-505 P5/1 1650 64/32 1.9/0 3.51 -- -- -- -- -- #p5-505 P5/2 1650 64/32 1.9/36 9.86 34.1 33.5 59.4 57.0 -- p5-505 P5+/1 1900 64/32 1.9/0 4.10 -- -- -- -- -- p5-505 P5+/2 1900 64/32 1.9/36 11.49 -- -- -- -- -- p5-505 P5+/2 2100 64/32 1.9/36 12.46 44.6 43.4 73.4 71.6 -- p5-505Q P5+/4 1650 64/32 1.9/72 20.25 70.0 68.6 100 97.2 -- #p5-510 P5/1 1500 64/32 1.9/0 3.25 -- -- -- -- -- #p5-510 P5/2 1500 64/32 1.9/36 9.13 -- -- -- -- -- Page 7
- 8. SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor Cache Cache rate base rate base SPEC Model / # Cores MHz (KB) (MB) rPerf 2000 2000 2000 2000 web99 #p5-510 P5/1 1650 64/32 1.9/36 5.24 -- -- -- -- -- #p5-510 P5/2 1650 64/32 1.9/36 9.86 33.0 31.6 43.2 41.5 -- #p5-510 P5+/1 1900 64/32 1.9/36 6.11 -- -- -- -- -- #p5-510 P5+/2 1900 64/32 1.9/36 11.49 39.9 39.4 67.1 65.9 -- p5-510 P5+/1 2100 64/32 1.9/36 6.63 -- -- -- -- -- p5-510 P5+/2 2100 64/32 1.9/36 12.46 44.6 43.4 73.4 71.6 -- #p5-510Q P5+/4 1500 64/32 1.9/72 18.75 63.0 61.4 95.5 89.8 -- p5-510Q P5+/4 1650 64/32 1.9/72 20.25 70.0 68.6 100 97.2 -- #p5-520 P5/1 1500 64/32 1.9/0 3.25 -- -- -- -- -- #p5-520 P5/2 1500 64/32 1.9/36 9.13 -- -- 40.0 38.7 -- #p5-520 P5/2 1650 64/32 1.9/36 9.86 32.9 30.3 43.0 41.5 -- #p5-520 P5+/1 1650 64/32 1.9/0 3.62 -- -- -- -- -- #p5-520 P5+/2 1650 64/32 1.9/36 10.15 34.8 34.4 61.5 59.0 -- #p5-520 P5+/2 1900 64/32 1.9/36 11.16 39.6 38.9 67.6 65.4 -- p5-520 P5+/1 2100 64/32 1.9/36 6.63 -- -- -- -- -- p5-520 P5+/2 2100 64/32 1.9/36 12.46 44.6 43.4 73.4 71.6 -- #p5-520Q P5+/4 1500 64/32 1.9/72 18.75 63.0 61.4 95.5 89.8 -- p5-520Q P5+/4 1650 64/32 1.9/72 20.25 70.0 68.6 100 97.2 -- #p5-550 P5/1 1500 64/32 1.9/0 3.25 -- -- -- -- -- #p5-550 P5/2 1500 64/32 1.9/36 9.13 -- -- -- -- -- #p5-550 P5/4 1500 64/32 3.8/72 18.20 -- -- 80.6 77.4 -- #p5-550 P5/2 1650 64/32 1.9/36 9.86 -- -- -- -- -- #p5-550 P5/4 1650 64/32 3.8/72 19.66 65.5 60.4 84.8 82.1 -- #p5-550 P5+/2 1650 64/32 1.9/36 10.15 -- -- -- -- -- #p5-550 P5+/4 1650 64/32 3.8/72 20.25 69.0 68.1 119 117 -- p5-550 P5+/2 1900 64/32 1.9/36 11.16 -- -- -- -- -- p5-550 P5+/4 1900 64/32 3.8/72 22.26 78.5 77.1 133 129 -- p5-550 P5+/2 2100 64/32 1.9/36 12.46 -- -- -- -- -- p5-550 P5+/4 2100 64/32 3.8/72 24.86 90.0 87.5 149 139 -- #p5-550Q P5+/4 1500 64/32 3.8/72 18.20 -- -- -- -- -- #p5-550Q P5+/8 1500 64/32 7.6/144 34.46 124 122 178 174 -- p5-550Q P5+/4 1650 64/32 3.8/72 20.25 -- -- -- -- -- p5-550Q P5+/8 1650 64/32 7.6/144 38.34 140 137 202 189 -- #p5-560Q P5+/4 1500 64/32 3.8/72 18.75 -- -- -- -- -- #p5-560Q P5+/8 1500 64/32 7.6/144 35.50 -- -- -- -- -- #p5-560Q P5+/16 1500 64/32 7.6/288 65.24 248 243 368 360 -- p5-560Q P5+/4 1800 64/32 3.8/72 21.72 -- -- -- -- -- p5-560Q P5+/8 1800 64/32 7.6/144 41.12 -- -- -- -- -- p5-560Q P5+/16 1800 64/32 7.6/288 75.58 -- ;; -- -- -- #p5-570 P5/2 1500 64/32 1.9/36 9.13 -- -- -- -- -- #p5-570 P5/4 1500 64/32 3.8/72 18.20 -- -- -- -- -- #p5-570 P5/8 1500 64/32 7.6/144 34.46 -- -- -- -- -- #p5-570 P5/2 1650 64/32 1.9/36 9.86 -- -- -- -- -- #p5-570 P5/4 1650 64/32 3.8/72 19.66 -- -- -- -- -- #p5-570 P5/8 1650 64/32 7.6/144 37.22 -- -- -- -- -- #p5-570 P5/12 1650 64/32 11.4/216 53.43 -- -- -- -- -- #p5-570 P5/16 1650 64/32 15.2/288 68.40 -- -- -- -- -- #p5-570 P5/2 1900 64/32 1.9/36 11.16 -- -- -- -- -- #p5-570 P5/4 1900 64/32 3.8/72 22.26 76.3 74.4 130 125 -- #p5-570 P5/8 1900 64/32 7.6/144 42.14 147 141 249 241 -- #p5-570 P5/12 1900 64/32 11.4/216 60.50 -- -- -- -- -- #p5-570 P5/16 1900 64/32 15.2/288 77.45 294 273 460 438 -- p5-570 P5+/2 1900 64/32 1.9/36 12.27 -- -- -- -- -- Page 8
- 9. SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor Cache Cache rate base rate base SPEC Model / # Cores MHz (KB) (MB) rPerf 2000 2000 2000 2000 web99 p5-570 P5+/4 1900 64/32 3.8/72 24.48 -- -- -- -- -- p5-570 P5+/8 1900 64/32 7.6/144 46.36 -- -- -- -- -- p5-570 P5+/12 1900 64/32 11.4/216 66.55 -- -- -- -- -- p5-570 P5+/16 1900 64/32 15.2/288 85.20 -- -- -- -- -- p5-570 P5+2 2200 64/32 1.9/36 13.83 -- -- -- -- -- p5-570 P5+/4 2200 64/32 3.8/72 27.58 -- -- -- -- -- p5-570 P5+/8 2200 64/32 7.6/144 52.21 -- -- -- -- -- p5-570 P5+/12 2200 64/32 11.4/216 74.95 -- -- -- -- -- p5-570 P5+/16 2200 64/32 15.2/288 95.96 -- -- -- -- -- #p5-575 P5/16 1500 64/32 15.2/288 -- 238 230 385 359 -- #p5-575 P5/8 1900 64/32 15.2/288 -- 167 159 282 266 -- p5-575 P5+/16 1900 64/32 15.2/288 -- 314 310 571 541 -- p5-575 P5+/8 2200 64/32 15.2/288 -- 200 196 382 355 -- #p5-590 P5/8 1650 64/32 7.6/144 41.68 -- -- -- -- -- #p5-590 P5/16 1650 64/32 15.2/288 80.86 -- -- -- -- -- #p5-590 P5/24 1650 64/32 22.8/432 116.29 -- -- -- -- -- #p5-590 P5/32 1650 64/32 30.4/576 151.72 529 503 870 824 -- p5-590 P5+/8 2100 64/32 7.6/144 55.74 -- -- -- -- -- p5-590 P5+/16 2100 64/32 15.2/288 108.13 -- -- -- -- -- p5-590 P5+/24 2100 64/32 22.8/432 155.51 -- -- -- -- -- p5-590 P5+/32 2100 64/32 30.4/576 202.88 -- -- -- -- -- #p5-595 P5/16 1650 64/32 15.2/288 80.86 -- -- -- -- -- #p5-595 P5/24 1650 64/32 22.8/432 116.29 -- -- -- -- -- #p5-595 P5/32 1650 64/32 30.4/576 151.72 -- -- -- -- -- #p5-595 P5/40 1650 64/32 38.0/720 182.07 -- -- -- -- -- #p5-595 P5/48 1650 64/32 45.6/864 212.41 -- -- -- -- -- #p5-595 P5/56 1650 64/32 53.2/1008 242.76 -- -- -- -- -- #p5-595 P5/64 1650 64/32 60.8/1152 273.10 -- -- -- -- -- #p5-595 P5/16 1900 64/32 15.2/288 90.67 -- -- -- -- -- #p5-595 P5/24 1900 64/32 22.8/432 130.39 -- -- -- -- -- #p5-595 P5/32 1900 64/32 30.4/576 170.11 -- -- -- -- -- #p5-595 P5/40 1900 64/32 38.0/720 204.14 -- -- -- -- -- #p5-595 P5/48 1900 64/32 45.6/864 238.16 -- -- -- -- -- #p5-595 P5/56 1900 64/32 53.2/1008 272.18 -- -- -- -- -- #p5-595 P5/64 1900 64/32 60.8/1152 306.21 1,147 1,063 1,752 1,684 -- p5-595 P5+/16 2100 64/32 15.2/288 108.13 -- -- -- -- -- p5-595 P5+/24 2100 64/32 22.8/432 155.51 -- -- -- -- -- p5-595 P5+/32 2100 64/32 30.4/576 202.88 -- -- -- -- -- p5-595 P5+/40 2100 64/32 38.0/720 243.46 -- -- -- -- -- p5-595 P5+/48 2100 64/32 45.6/864 284.04 -- -- -- -- -- p5-595 P5+/56 2100 64/32 53.2/1008 324.61 -- -- -- -- -- p5-595 P5+/64 2100 64/32 60.8/1152 365.19 -- -- -- -- -- p5-595 P5+/16 2300 64/32 15.2/288 116.53 -- -- -- -- -- p5-595 P5+/24 2300 64/32 22.8/432 167.58 -- -- -- -- -- p5-595 P5+/32 2300 64/32 30.4/576 218.64 -- -- -- -- -- p5-595 P5+/40 2300 64/32 38.0/720 262.37 -- -- -- -- -- p5-595 P5+/48 2300 64/32 45.6/864 306.10 -- -- -- -- -- p5-595 P5+/56 2300 64/32 53.2/1008 349.83 -- -- -- -- -- p5-595 P5+/64 2300 64/32 60.8/1152 393.55 1,513 1,488 2,406 2,215 -- Page 9
- 10. Section 2b – Power Systems Multiuser Performance using AIX 5L V5.2 Processor L1 Cache L2/L3 Cache Model / # Cores MHz (KB) (MB) rPerf #p5-505 P5/2 1500 64/32 1.9/36 7.02 #p5-505 P5/1 1650 64/32 1.9/0 2.70 #p5-505 P5/2 1650 64/32 1.9/36 7.58 p5-505 P5+/1 1900 64/32 1.9/0 3.15 p5-505 P5+/2 1900 64/32 1.9/36 8.84 p5-505 P5+/2 2100 64/32 1.9/36 9.59 p5-505Q P5+/4 1650 64/32 1.9/72 15.57 #p5-510 P5/1 1500 64/32 1.9/0 2.50 #p5-510 P5/2 1500 64/32 1.9/36 7.02 #p5-510 P5/1 1650 64/32 1.9/36 4.03 #p5-510 P5/2 1650 64/32 1.9/36 7.58 #p5-510 P5+/1 1900 64/32 1.9/36 4.70 #p5-510 P5+/2 1900 64/32 1.9/36 8.83 p5-510 P5+/1 2100 64/32 1.9/0 5.10 p5-510 P5+/2 2100 64/32 1.9/36 9.59 #p5-510Q P5+/4 1500 64/32 1.9/72 14.42 p5-510Q P5+/4 1650 64/32 1.9/72 15.57 #p5-520 P5/1 1500 64/32 1.9/0 2.50 #p5-520 P5/2 1500 64/32 1.9/36 7.02 #p5-520 P5/2 1650 64/32 1.9/36 7.58 #p5-520 P5+/1 1650 64/32 1.9/0 2.78 #p5-520 P5+/2 1650 64/32 1.9/36 7.80 #p5-520 P5+/2 1900 64/32 1.9/36 8.58 p5-520 P5+/1 2100 64/32 1.9/36 5.10 p5-520 P5+/2 2100 64/32 1.9/36 9.59 #p5-520Q P5+/4 1500 64/32 1.9/72 14.42 p5-520Q P5+/4 1650 64/32 1.9/72 15.57 #p5-550 P5/1 1500 64/32 1.9/0 2.50 #p5-550 P5/2 1500 64/32 1.9/36 7.02 #p5-550 P5/4 1500 64/32 3.8/72 14.00 #p5-550 P5/2 1650 64/32 1.9/36 7.58 #p5-550 P5/4 1650 64/32 3.8/72 15.12 #p5-550 P5+/2 1650 64/32 1.9/36 7.80 #p5-550 P5+/4 1650 64/32 3.8/72 15.57 p5-550 P5+/2 1900 64/32 1.9/36 8.58 p5-550 P5+/4 1900 64/32 3.8/72 17.12 p5-550 P5+/2 2100 64/32 1.9/36 9.59 p5-550 P5+/4 2100 64/32 3.8/72 19.12 #p5-550Q P5+/4 1500 64/32 3.8/72 14.00 #p5-550Q P5+/8 1500 64/32 7.6/144 26.50 p5-550Q P5+/4 1650 64/32 3.8/72 15.57 p5-550Q P5+/8 1650 64/32 7.6/144 29.49 p5-560Q P5+/4 1500 64/32 3.8/72 14.42 p5-560Q P5+/8 1500 64/32 7.6/144 27.30 p5-560Q P5+/16 1500 64/32 15.2/288 50.18 p5-560Q P5+/4 1800 64/32 3.8/72 16.71 p5-560Q P5+/8 1800 64/32 7.6/144 31.63 p5-560Q P5+/16 1800 64/32 15.2/288 58.14 #p5-570 P5/2 1500 64/32 1.9/36 7.02 #p5-570 P5/4 1500 64/32 3.8/72 14.00 #p5-570 P5/8 1500 64/32 7.6/144 26.50 #p5-570 P5/2 1650 64/32 1.9/36 7.58 #p5-570 P5/4 1650 64/32 3.8/72 15.12 #p5-570 P5/8 1650 64/32 7.6/144 28.63 Page 10
- 11. Processor L1 Cache L2/L3 Cache Model / # Cores MHz (KB) (MB) rPerf #p5-570 P5/12 1650 64/32 11.4/216 41.10 #p5-570 P5/16 1650 64/32 15.2/288 52.61 #p5-570 P5/2 1900 64/32 1.9/36 8.58 #p5-570 P5/4 1900 64/32 3.8/72 17.12 #p5-570 P5/8 1900 64/32 7.6/144 32.41 #p5-570 P5/12 1900 64/32 11.4/216 46.53 #p5-570 P5/16 1900 64/32 15.2/288 59.57 p5-570 P5+/2 1900 64/32 1.9/36 9.43 p5-570 P5+/4 1900 64/32 3.8/72 18.83 p5-570 P5+/8 1900 64/32 7.6/144 35.66 p5-570 P5+/12 1900 64/32 11.4/216 51.19 p5-570 P5+/16 1900 64/32 15.2/288 65.53 p5-570 P5+/2 2200 64/32 1.9/36 10.63 p5-570 P5+/4 2200 64/32 3.8/72 21.21 p5-570 P5+/8 2200 64/32 7.6/144 40.16 p5-570 P5+/12 2200 64/32 11.4/216 57.65 p5-570 P5+/16 2200 64/32 15.2/288 73.81 #p5-590 P5/8 1650 64/32 7.6/144 32.06 #p5-590 P5/16 1650 64/32 15.2/288 62.20 #p5-590 P5/24 1650 64/32 22.8/432 89.46 #p5-590 P5/32 1650 64/32 30.4/576 116.71 p5-590 P5+/8 2100 64/32 7.6/144 42.87 p5-590 P5+/16 2100 64/32 15.2/288 83.18 p5-590 P5+/24 2100 64/32 22.8/432 119.62 p5-590 P5+/32 2100 64/32 30.4/576 156.06 #p5-595 P5/16 1650 64/32 15.2/288 62.20 #p5-595 P5/24 1650 64/32 22.8/432 89.46 #p5-595 P5/32 1650 64/32 30.4/576 116.71 #p5-595 P5/40 1650 64/32 38.0/720 140.05 #p5-595 P5/48 1650 64/32 45.6/864 163.39 #p5-595 P5/56 1650 64/32 53.2/1008 186.74 #p5-595 P5/64 1650 64/32 60.8/1152 210.08 #p5-595 P5/16 1900 64/32 15.2/288 69.74 #p5-595 P5/24 1900 64/32 22.8/432 100.30 #p5-595 P5/32 1900 64/32 30.4/576 130.86 #p5-595 P5/40 1900 64/32 38.0/720 157.03 #p5-595 P5/48 1900 64/32 45.6/864 183.20 #p5-595 P5/56 1900 64/32 53.2/1008 209.37 #p5-595 P5/64 1900 64/32 60.8/1152 235.54 p5-595 p5+/16 2100 64/32 15.2/288 83.18 p5-595 P5+/24 2100 64/32 22.8/432 119.62 p5-595 P5+/32 2100 64/32 30.4/576 156.06 p5-595 P5+/40 2100 64/32 38.0/720 187.28 p5-595 P5+/48 2100 64/32 45.6/864 218.49 p5-595 P5+/56 2100 64/32 53.2/1008 249.70 p5-595 P5+/64 2100 64/32 60.8/1152 280.92 p5-595 P5+/16 2300 64/32 15.2/288 89.64 p5-595 P5+/24 2300 64/32 22.8/432 128.91 p5-595 P5+/32 2300 64/32 30.4/576 168.19 p5-595 P5+/40 2300 64/32 38.0/720 201.82 p5-595 P5+/48 2300 64/32 45.6/864 235.46 p5-595 P5+/56 2300 64/32 53.2/1008 269.10 p5-595 P5+/64 2300 64/32 60.8/1152 302.73 Page 11
- 12. Section 2c - AIX Capacity Upgrade on Demand Relative Performance Guidelines Model Processor / # Cores GHz rPerf #p5-570 P5/4 1.65 19.66 #p5-570 P5/6 1.65 28.44 #p5-570 P5/8 1.65 37.22 #p5-570 P5/10 1.65 45.33 #p5-570 P5/12 1.65 53.43 #p5-570 P5/14 1.65 60.92 #p5-570 P5/16 1.65 68.40 #p5-570 P5/4 1.9 22.26 #p5-570 P5/6 1.9 32.20 #p5-570 P5/8 1.9 42.14 #p5-570 P5/10 1.9 51.32 #p5-570 P5/12 1.9 60.50 #p5-570 P5/14 1.9 68.98 #p5-570 P5/16 1.9 77.45 p5-570 P5+/4 1.9 24.48 p5-570 P5+/6 1.9 35.42 p5-570 P5+/8 1.9 46.36 p5-570 P5+/10 1.9 56.45 p5-570 P5+/12 1.9 66.55 p5-570 P5+/14 1.9 75.87 p5-570 P5+/16 1.9 85.20 p5-570 P5+/4 2.2 27.58 p5-570 P5+/6 2.2 39.90 p5-570 P5+/8 2.2 52.21 p5-570 P5+/10 2.2 63.58 p5-570 P5+/12 2.2 74.95 p5-570 P5+/14 2.2 85.46 p5-570 P5+/16 2.2 95.96 570 P6/4 3.5 31.69 570 P6/6 3.5 45.32 570 P6/8 3.5 58.95 570 P6/10 3.5 71.15 570 P6/12 3.5 83.35 570 P6/14 3.5 94.55 570 P6/16 3.5 105.75 570 P6/4 4.2 36.76 570 P6/6 4.2 52.57 570 P6/8 4.2 68.38 570 P6/10 4.2 82.53 570 P6/12 4.2 96.68 570 P6/14 4.2 109.67 570 P6/16 4.2 122.67 570/32 P6+/4 4.2 35.50 570/32 P6+/6 4.2 50.23 570/32 p6+/8 4.2 64.96 570/32 p6+/10 4.2 77.14 570/32 p6+/12 4.2 89.32 570/32 p6+/14 4.2 101.50 570/32 p6+/16 4.2 113.68 570/32 p6+/18 4.2 123.62 570/32 p6+/20 4.2 133.57 570/32 p6+/22 4.2 143.52 Page 12
- 13. Model Processor / # Cores GHz rPerf 570/32 p6+/24 4.2 153.46 570/32 p6+/26 4.2 163.41 570/32 p6+/28 4.2 173.36 570/32 p6+/30 4.2 183.30 570/32 p6+/32 4.2 193.25 570 P6+/4 4.4 38.16 570 P6+/6 4.4 54.56 570 p6+/8 4.4 70.97 570 p6+/10 4.4 85.66 570 p6+/12 4.4 100.35 570 p6+/14 4.4 113.84 570 p6+/16 4.4 127.32 570 P6/4 4.7 40.26 570 P6/6 4.7 57.58 570 P6/8 4.7 74.89 570 P6/10 4.7 90.39 570 P6/12 4.7 105.89 570 P6/14 4.7 120.12 570 P6/16 4.7 134.35 570 P6+/4 5.0 42.32 570 P6+/6 5.0 60.52 570 p6+/8 5.0 78.71 570 p6+/10 5.0 95.01 570 p6+/12 5.0 111.30 570 p6+/14 5.0 126.25 570 p6+/16 5.0 141.21 770 P7/4 3.50 49.19 770 P7/6 3.50 72.55 770 P7/8 3.50 95.29 770 P7/12 3.50 140.75 770 P7/16 3.50 180.90 770 P7/20 3.50 221.05 770 P7/24 3.50 261.19 770 P7/28 3.50 299.89 770 P7/32 3.50 338.58 770 P7/36 3.50 377.28 770 P7/40 3.50 415.97 770 P7/44 3.50 454.67 770 P7/48 3.50 493.37 770 P7/4 3.10 44.84 770 P7/8 3.10 85.20 770 P7/12 3.10 125.25 770 P7/16 3.10 165.30 770 P7/20 3.10 200.66 770 P7/24 3.10 236.02 770 P7/28 3.10 271.38 770 P7/32 3.10 306.74 770 P7/36 3.10 340.82 770 P7/40 3.10 374.90 770 P7/44 3.10 408.98 770 P7/48 3.10 443.06 770 P7/52 3.10 477.15 Page 13
- 14. Model Processor / # Cores GHz rPerf 770 P7/56 3.10 511.23 770 P7/60 3.10 545.31 770 P7/64 3.10 579.39 780 P7/4 3.86 53.03 780 P7/8 3.86 100.75 780 P7/12 3.86 148.10 780 P7/16 3.86 195.45 780 P7/20 3.86 237.26 780 P7/24 3.86 279.07 780 P7/28 3.86 320.88 780 P7/32 3.86 362.70 780 P7/36 3.86 403.00 780 P7/40 3.86 443.29 780 P7/44 3.86 483.59 780 P7/48 3.86 523.89 780 P7/52 3.86 564.19 780 P7/56 3.86 604.49 780 P7/60 3.86 644.79 780 P7/64 3.86 685.09 780 P7/4 4.14 59.26 780 P7/8 4.14 115.86 780 P7/12 4.14 171.42 780 P7/16 4.14 226.97 780 P7/20 4.14 276.61 780 P7/24 4.14 326.24 780 P7/28 4.14 375.87 780 P7/32 4.14 425.50 #p5-590 P5/8 1.65 41.68 #p5-590 P5/10 1.65 51.48 #p5-590 P5/12 1.65 61.27 #p5-590 P5/14 1.65 71.07 #p5-590 P5/16 1.65 80.86 #p5-590 P5/18 1.65 89.72 #p5-590 P5/20 1.65 98.58 #p5-590 P5/22 1.65 107.44 #p5-590 P5/24 1.65 116.29 #p5-590 P5/26 1.65 125.15 #p5-590 P5/28 1.65 134.01 #p5-590 P5/30 1.65 142.87 #p5-590 P5/32 1.65 151.72 p5-590 P5+/8 2.1 55.74 p5-590 P5+/10 2.1 68.84 p5-590 P5+/12 2.1 81.93 p5-590 P5+/14 2.1 95.03 p5-590 P5+/16 2.1 108.13 p5-590 P5+/18 2.1 119.98 p5-590 P5+/20 2.1 131.82 p5-590 P5+/22 2.1 143.67 p5-590 P5+/24 2.1 155.51 p5-590 P5+/26 2.1 167.35 p5-590 P5+/28 2.1 179.20 p5-590 P5+/30 2.1 191.04 p5-590 P5+/32 2.1 202.88 Page 14
- 15. Model Processor / # Cores GHz rPerf #p5-595 P5/16 1.65 80.86 #p5-595 P5/18 1.65 89.72 #p5-595 P5/20 1.65 98.58 #p5-595 P5/22 1.65 107.44 #p5-595 P5/24 1.65 116.29 #p5-595 P5/26 1.65 125.15 #p5-595 P5/28 1.65 134.01 #p5-595 P5/30 1.65 142.87 #p5-595 P5/32 1.65 151.72 #p5-595 P5/34 1.65 159.31 #p5-595 P5/36 1.65 166.90 #p5-595 P5/38 1.65 174.48 #p5-595 P5/40 1.65 182.07 #p5-595 P5/42 1.65 189.65 #p5-595 P5/44 1.65 197.24 #p5-595 P5/46 1.65 204.83 #p5-595 P5/48 1.65 212.41 #p5-595 P5/50 1.65 220.00 #p5-595 P5/52 1.65 227.58 #p5-595 P5/54 1.65 235.17 #p5-595 P5/56 1.65 242.76 #p5-595 P5/58 1.65 250.34 #p5-595 P5/60 1.65 257.93 #p5-595 P5/62 1.65 265.52 #p5-595 P5/64 1.65 273.10 #p5-595 P5/16 1.9 90.67 #p5-595 P5/18 1.9 100.60 #p5-595 P5/20 1.9 110.53 #p5-595 P5/22 1.9 120.46 #p5-595 P5/24 1.9 130.39 #p5-595 P5/26 1.9 140.32 #p5-595 P5/28 1.9 150.25 #p5-595 P5/30 1.9 160.18 #p5-595 P5/32 1.9 170.11 #p5-595 P5/34 1.9 178.62 #p5-595 P5/36 1.9 187.13 #p5-595 P5/38 1.9 195.63 #p5-595 P5/40 1.9 204.14 #p5-595 P5/42 1.9 212.64 #p5-595 P5/44 1.9 221.15 #p5-595 P5/46 1.9 229.65 #p5-595 P5/48 1.9 238.16 #p5-595 P5/50 1.9 246.67 #p5-595 P5/52 1.9 255.17 #p5-595 P5/54 1.9 263.68 #p5-595 P5/56 1.9 272.18 #p5-595 P5/58 1.9 280.69 #p5-595 P5/60 1.9 289.19 #p5-595 P5/62 1.9 297.70 #p5-595 P5/64 1.9 306.21 p5-595 P5+/16 2.1 108.13 p5-595 P5+/18 2.1 119.98 p5-595 P5+/20 2.1 131.82 p5-595 P5+/22 2.1 143.67 p5-595 P5+/24 2.1 155.51 Page 15
- 16. Model Processor / # Cores GHz rPerf p5-595 P5+/26 2.1 167.35 p5-595 P5+/28 2.1 179.20 p5-595 P5+/30 2.1 191.04 p5-595 P5+/32 2.1 202.88 p5-595 P5+/34 2.1 213.03 p5-595 P5+/36 2.1 223.17 p5-595 P5+/38 2.1 233.32 p5-595 P5+/40 2.1 243.46 p5-595 P5+/42 2.1 253.61 p5-595 P5+/44 2.1 263.75 p5-595 P5+/46 2.1 273.90 p5-595 P5+/48 2.1 284.04 p5-595 P5+/50 2.1 294.18 p5-595 P5+/52 2.1 304.33 p5-595 P5+/54 2.1 314.47 p5-595 P5+/56 2.1 324.61 p5-595 P5+/58 2.1 334.76 p5-595 P5+/60 2.1 344.90 p5-595 P5+/62 2.1 355.05 p5-595 P5+/64 2.1 365.19 p5-595 P5+/16 2.3 116.53 p5-595 P5+/18 2.3 129.29 p5-595 P5+/20 2.3 142.06 p5-595 P5+/22 2.3 154.82 p5-595 P5+/24 2.3 167.58 p5-595 P5+/26 2.3 180.35 p5-595 P5+/28 2.3 193.11 p5-595 P5+/30 2.3 205.88 p5-595 P5+/32 2.3 218.64 p5-595 P5+/34 2.3 229.57 p5-595 P5+/36 2.3 240.51 p5-595 P5+/38 2.3 251.44 p5-595 P5+/40 2.3 262.37 p5-595 P5+/42 2.3 273.30 p5-595 P5+/44 2.3 284.24 p5-595 P5+/46 2.3 295.17 p5-595 P5+/48 2.3 306.10 p5-595 P5+/50 2.3 317.03 p5-595 P5+/52 2.3 327.97 p5-595 P5+/54 2.3 338.90 p5-595 P5+/56 2.3 349.83 p5-595 P5+/58 2.3 360.76 p5-595 P5+/60 2.3 371.69 p5-595 P5+/62 2.3 382.62 p5-595 P5+/64 2.3 393.55 595 P6/8 4.2 75.58 595 P6/10 4.2 92.41 595 P6/12 4.2 109.24 595 P6/14 4.2 126.07 595 P6/16 4.2 142.90 595 P6/18 4.2 158.35 595 P6/20 4.2 173.80 595 P6/22 4.2 189.25 595 P6/24 4.2 204.70 595 P6/26 4.2 220.15 595 P6/28 4.2 235.60 595 P6/30 4.2 251.06 Page 16
- 17. Model Processor / # Cores GHz rPerf 595 P6/32 4.2 266.51 595 P6/34 4.2 279.89 595 P6/36 4.2 293.28 595 P6/38 4.2 306.67 595 P6/40 4.2 320.05 595 P6/42 4.2 333.44 595 P6/44 4.2 346.83 595 P6/46 4.2 360.21 595 P6/48 4.2 373.60 595 P6/50 4.2 386.89 595 P6/52 4.2 400.17 595 P6/54 4.2 413.46 595 P6/56 4.2 426.74 595 P6/58 4.2 440.03 595 P6/60 4.2 453.32 595 P6/62 4.2 466.60 595 P6/64 4.2 479.89 595 P6/8 5.0 87.10 595 P6/10 5.0 106.49 595 P6/12 5.0 125.88 595 P6/14 5.0 145.28 595 P6/16 5.0 164.67 595 P6/18 5.0 182.48 595 P6/20 5.0 200.28 595 P6/22 5.0 218.09 595 P6/24 5.0 235.90 595 P6/26 5.0 253.70 595 P6/28 5.0 271.51 595 P6/30 5.0 289.31 595 P6/32 5.0 307.12 595 P6/34 5.0 322.54 595 P6/36 5.0 337.97 595 P6/38 5.0 353.40 595 P6/40 5.0 368.82 595 P6/42 5.0 384.25 595 P6/44 5.0 399.68 595 P6/46 5.0 415.10 595 P6/48 5.0 430.53 595 P6/50 5.0 445.84 595 P6/52 5.0 461.15 595 P6/54 5.0 476.46 595 P6/56 5.0 491.77 595 P6/58 5.0 507.08 595 P6/60 5.0 522.39 595 P6/62 5.0 537.70 595 P6/64 5.0 553.01 Page 17
- 18. Section 2d – POWER6 CPW Published Results # Model Cores GHz CPW JS12 2 3.8 7100 JS22 4 4.0 13800 JS23 4 4.2 14400 JS43 8 4.2 24050 520 1 4.2 4300 520 2 4.2 8300 520 4 4.2 15600 520 2 4.7 9500 520 4 4.7 18300 550 2 3.5 7750 550 4 3.5 15000 550 6 3.5 20300 550 8 3.5 27600 550 2 4.2 9200 550 4 4.2 18000 550 6 4.2 23850 550 8 4.2 32650 550 2 5.0 10600 550 4 5.0 20550 550 6 5.0 28800 550 8 5.0 37950 560 4 3.6 14100 560 8 3.6 27600 560 16 3.6 48500 570 2 3.5 8150 570 4 3.5 16100 570 8 3.5 30100 570 12 3.5 43100 570 16 3.5 57600 570 2 4.2 9650 570 4 4.2 19200 570 8 4.2 35500 570 16 4.2 68600 570 2 4.4 9850 570 4 4.4 19400 570 8 4.4 36200 570 12 4.4 51500 570 16 4.4 70000 570 2 4.7 10800 570 4 4.7 21200 570 8 4.7 40100 570 16 4.7 76900 570 2 5.0 11000 570 4 5.0 21600 570 8 5.0 40300 570 12 5.0 56800 570 16 5.0 77600 570 4 (32) 4.2 16200 570 8 (32) 4.2 31900 570 16(32) 4.2 56400 570 24(32) 4.2 81600 570 32(32) 4.2 104800 595 8 4.2 35,500 595 16 4.2 66,400 Page 18
- 19. 595 24 4.2 93,800 595 32 4.2 128,000 595 64 4.2 256,200 595 8 5.0 41,000 595 16 5.0 77,000 595 24 5.0 108,100 595 32 5.0 147,900 595 64 5.0 294,700 750 8 3.0 44600 750 16 3.0 82600 750 24 3.0 122500 750 32 3.0 158300 750 6 3.3 37200 750 12 3.3 69200 750 18 3.3 94900 750 24 3.3 135300 750 8 3.3 47800 750 16 3.3 88700 750 24 3.3 129700 750 32 3.3 168800 750 32 3.55 181000 770 4 3.1 22750 770 8 3.1 45000 770 16 3.1 88800 770 32 3.1 155850 770 48 3.1 229800 770 64 3.1 292700 770 4 3.5 24900 770 6 3.5 37400 770 12 3.5 73100 770 18 3.5 99000 770 24 3.5 131050 770 48 3.5 248550 780 4 3.8 26600 780 8 3.8 54400 780 16 3.8 105200 780 32 3.8 177400 780 48 3.8 265200 780 64 3.8 343050 CPW values for non-POWER6 System i™ models may be obtained from http://www.ibm.com/systems/i/solutions/perfmgmt/resource.html . Section 3 - TPC-C Version 5 Published Results L2/L3 Processor # Cache $/tmp Avail. Model / # Cores Nodes GHz (MB) tpmC C Database AIX Date 550 P6/8 1 4.2 32/128 629,159.00 2.49 DB2 9.5 5.3.0 04/20/08 #p5-570 P5/4 1 1.9 3.8/72 194,391.43 5.62 Oracle DB 10g 5.3.0 09/30/04 #p5-570 P5/4 1 1.9 3.8/72 203,439.87 3.93 Oracle DB 10g 5.3.0 10/17/05 #p5-570 P5/8 1 1.9 7.6/144 371,044.22 5.26 Oracle DB 10g 5.3.0 09/30/04 #p5-570 P5/8 1 1.9 7.6/144 429,899.7 4.99 DB2® UDB V8.1 5.3.0 09/30/04 #p5-570 P5/16 1 1.9 15.2/288 809,144.09 4.95 DB2 UDB V8.1 5.3.0 09/30/04 p5-570 P5+/16 1 2.2 15.2/288 1,025,169.69 4.42 DB2 UDB V8.2 5.3.0 05/31/06 570 P6/4 1 4.7 16/64 404,462.54 3.50 Oracle DB 10g 5.3.0 11/26/07 Page 19
- 20. L2/L3 Processor # Cache $/tmp Avail. Model / # Cores Nodes GHz (MB) tpmC C Database AIX Date 570 P6/16 1 4.7 64/256 1,616,162.84 3.54 DB2 9.1 5.3.0 11/21/07 #p5-595 P5/32 1 1.9 30.4/576 1,601,784.98 5.05 Oracle DB 10g 5.3.0 04/20/05 #p5-595 P5/64 1 1.9 60.8/1152 3,210,540.63 5.07 DB2 UDB V8.2 5.3.0 05/14/05 p5-595 P5+/64 1 2.3 60.8/1152 4,033,378.00 2.97 DB2 9.1 5.3.0 12/20/06 595 P6/64 1 5.0 256/1024 6,085,166.00 2.81 DB2 9.5 5.3.0 12/10/08 Section 4 - TPC-H Published Results TPC-H 1000 GB (1 TB): # $/ Avail. Proc. Nodes MHz QphH QppH QthH QphH AIX Database Date #p5-570 P5 4 1900 26,156 35,789.6 19,115.9 53 5.3.0 DB2 UDB V8.2 12/15/04 TPC-H 3000 GB (3 TB): Proc./# # $/ Avail. Model Cores Nodes MHz QphH QphH AIX Database Date #p5-595 P5/64 1 1900 100,512 53.00 5.3.0 Oracle DB 10g 03/01/06 595 P6/64 1 5000 156.537 20.60 6.1.0 Sybase IQ v.15.1 11/24/09 TPC-H 10000 GB (10 TB): Proc./# # $/ Avail. Model Cores Nodes MHz QphH QphH AIX Database Date p570 P6/4 32 4700 343,551.2 32.89 5.3.0 DB2 Warehouse 9.5 04/15/08 #p5-575 P5/8 8 1900 104,100.1 61 5.3.0 DB2 UDB V8.2 08/15/05 p5-575 P5+/8 16 2200 180,108 47.00 5.3.0 DB2 UDB V8.2 08/30/06 Section 5 - AIX SPECsfs97_R1 Benchmark Results L1 L2/L3 SPEC SPEC SPEC SPEC Proc./ Cache Cache sfs97_R1.v2 sfs97_R1.v2 sfs97_R1.v3 sfs97_R1.v3 Model # Cores MHz (KB) (MB) UDP TCP UDP TCP #p5-510 P5/2 1650 64/32 1.9/36 -- -- -- 42,033 #p5-550 P5/4 1650 64/32 3.8/72 -- -- -- 75,839 #p5-550Q P5+/8 1500 64/32 7.6/144 -- -- -- 118,391 #p5-570 P5/8 1900 64/32 7.6/144 -- -- -- 145,362 p5-570 P5+/8 2200 64/32 7.6/144 -- -- -- 169,786 Section 5a - AIX NotesBench Published Results L2/L3 Processor Cache Response Domin Work Model / # Cores MHz (MB) Users TPM Time $/User oVers. Load #p5-570 P5/8 1500 1.9/36 17,400 14,740 0.270 10.19 6.5.3 R6iNotes R7iNotes # Notes Notes Resp. # Work Bench Mark Time $/Notes # Users/ Model Proc. MHz Load Users (TPM) (ms.) $/User Mark Cores Core #p5-550Q P5+ 1500 DWA7 24,000 20,108 932 5.97 7.13 8 3,000 p5-560Q P5+ 1800 DWA7 55,000 46,193 848 4.89 5.82 16 3,438 Page 20
- 21. Section 6 - AIX Java Benchmarks (SPECjvm98, SPECjbb2000, SPECjbb2005) Published Results L1 L2/L3 SPEC SPEC Proc / MHz Cache Cache jvm98 jbb2000 SPECjbb2005 Model # Cores (KB) (MB) (256MB) ops/sec bops JVM inst. bops/JVM JS20 PPC970/2 2200 64/32 1.0/- -- 39,605 -- -- -- p5-505 P5+/2 2100 64/32 1.9/36 -- -- 41,751 1 41,751 p5-505Q P5+/4 1650 64/32 1.9/72 -- -- 63,544 2 31,772 #p5-510 P5/2 1500 64/32 1.9/36 -- 68,029 -- -- -- #p5-510 P5/2 1650 64/32 1.9/36 -- 76,040 -- -- -- #p5-510 P5+/2 1900 64/32 1.9/36 -- -- 36,039 1 36,039 #p5-510Q P5+/4 1500 64/32 1.9/72 -- -- 54,785 1 54,785 #p5-520 P5/2 1650 64/32 1.9/36 -- 75,607 -- -- -- #p5-520 P5+/2 1900 64/32 1.9/36 -- 99,844 32,820 1 32,820 #p5-550 P5+/4 1650 64/32 3.8/72 -- -- 60,419 1 60,419 p5-550 P5+/4 1900 64/32 3.8/72 -- 190,445 61,789 1 61,789 #p5-550Q P5+/8 1500 64/32 7.6/144 -- 294,315 91,806 1 91,806 p5-550Q P5+/8 1650 64/32 7.6/144 -- -- 127,851 8 15,981 550 P6/8 4200 64/64 32/128 -- -- 333,779 4 83,445 #p5-560Q P5+/16 1500 64/32 15.2/288 -- -- 226,291 8 28,286 p5-560Q P5+/16 1800 64/32 15.2/288 -- -- 278,384 8 34,798 560 P6+/16 3600 64/64 64/256 -- -- 593,904 8 74,238 750 P7/32 3.55 32/32 8/128 -- -- 2,478,929 32 77,467 #p5-570 P5/2 1900 64/32 1.9/36 -- 86,267 -- -- -- #p5-570 P5/4 1900 64/32 3.8/72 -- 170,127 -- -- -- #p5-570 P5/8 1900 64/32 7.6/144 -- 328,996 -- -- -- #p5-570 P5/16 1900 64/32 15.2/288 -- 633,106 224,200 1 224,200 #p5-570 P5/16 1900 64/32 15.2/288 -- -- 244,361 8 30,545 p5-570 P5+/16 2200 64/32 15.2/288 -- -- 326,651 8 40,831 570/32 P6+/32 4200 64/64 128/512 -- -- 1,390,087 16 86,880 570 P6/2 4700 64/64 8/32 -- -- 88,089 1 88,089 570 P6/4 4700 64/64 16/64 -- -- 175,474 2 87,737 570 P6/8 4700 64/64 32/128 -- -- 346,742 4 86,686 570 P6/16 4700 64/64 64/256 -- -- 691,975 8 86,497 570 P6/4 4700 64/64 16/64 -- -- 205.917 2 102,959 570 P6/8 4700 64/64 32/128 -- -- 402,923 4 100,731 570 P6/16 4700 64/64 64/256 -- -- 798,752 8 99,844 570 P6+/16 5000 64/64 64/256 -- -- 867,989 8 108,499 #p5-595 P5/64 1900 64/32 60.8/1152 -- 2,200,162 -- -- -- #p5-595 P5/64 1900 64/32 60.8/1152 -- 2,505,245 -- -- -- 595 P6/64 5000 64/64 256/1024 -- -- 3,435,485 32* 107,359 Section 6a - IBM i Java Benchmarks (SPECjbb2005) Published Results L1 L2/L3 Proc / MHz Cache Cache SPECjbb2005 Model # Cores (KB) (MB) Bops JVM inst. bops/JVM 750 P7/16 3.30 32/32 4/64 976,223 16 61,014 570 P6/8 4700 64/64 32/128 -- -- 345,809 595 P6/32 5000 64/64 128/512 -- -- 1,530,343 Section 6b - AIX SPECjAppServer2004 Benchmark Published J2EE DB Nodes / Nodes / J2EE Server GHz J2EE AppServer Cores DB Server GHz Database Cores JOPS JS22 4.0 WebSphere 6.1 26/4 P5-595 2.1 DB2 9.5 1/40 14,004.42 Page 21
- 22. J2EE DB Nodes / Nodes / J2EE Server GHz J2EE AppServer Cores DB Server GHz Database Cores JOPS p5-505 2.1 WebSphere® 6.0 1/2 p5-505 1.65 DB2 UDB V8.2 1/2 349.11 p5-505 2.1 WebSphere 6.1 1/2 p5-505Q 1.65 DB2 UDB V8.2 1/4 404.88 p5-505Q 1.65 WebSphere 6.1 1/4 p5-550 2.1 DB2 UDB V8.2 1/4 618.38 570 4.7 WebSphere 6.1 1/4 p5-550 2.1 DB2 9.1 1/4 1,197.51 Section 7 - SAP Standard Application Benchmarks Published Results Sales and Distribution – SD 2-Tier - AIX Dialog Fully # Avg. Steps Proc. Core Res Per Line CPU SAP (Proc/ p. Hour Items SAPS Util. ECC Model (GHz) Thread) Users Time (K) Per Hour (K) OS Database % Ver. Cert. # #p5-520 P5 1.65 2 (2/4) 572 1.96 172 57,330 2.87 AIX 5.3 DB2 UDB V8.1 98% 4.70 2004061 p5-505 P5+ 2.1 2 (2/4) 680 1.98 204 68,000 3.4 AIX 5.3 DB2 9.1 98% 5.0 2006047 p5-505Q P5+ 1.65 4 (1/8) 1,100 1.97 331 110,330 5.52 AIX 5.3 DB2 9.1 99% 5.0 2006046 550 P6 5.0 GHz 8 (4/16) 3752 .97 1,231 410,330 20.52 AIX 6.1 DB2 9.5 99% EHP4 2009023 #p5-570 P5 1.9 4 (4/8) 1,313 1.97 395 131,670 6.58 AIX 5.3 DB2 UDB V8.1 99% 4.70 2004042 #p5-570 P5 1.9 8 (8/16) 2,600 1.99 781 260,330 13.02 AIX 5.3 DB2 UDB V8.1 99% 4.70 2004041 #p5-570 P5 1.9 16 (16/32) 5,056 1.99 1,518 506,000 25.3 AIX 5.3 DB2 UDB V8.1 99% 4.70 2004040 p5-570 P5+ 2.2 16 (16/32) 5,520 1.97 1,660 553,330 27.67 AIX 5.3 DB2 UDB V8.2 99% 5.0 2006044 570 P6 4.7 (2 Virt 2 (1/4) 1020 1.99 306 102000 5.1 AIX 6.1 DB2 9.5 99% 6.0 2008080 CPU) 570 P6 4.7 4 (2/8) 2,035 1.99 611 203,670 10.18 AIX 5.3 Oracle DB 10g 99% 6.0 2007037 570 P6 4.7 8 (4/16) 4,010 1.96 1,207 402,330 20.12 AIX 5.3 Oracle DB 10g 99% 6.0 2007038 570 P6 4.7 16 (8/32) 8,000 1.98 2,404 801,330 40.07 AIX 5.3 DB2 9.5 99% 6.0 2007039 570/32 P6+ 4.2 32 (16/64) 14,432 1.99 4,335 1,445,000 72.25 AIX 6.1 DB2 9.5 99% 6.0 2008057 #p5-595 P5 1.9 64 (64/128) 20,000 1.92 6,042 2,014,000 100.7 AIX 5.3 DB2 UDB V8.1 97% 4.70 2004062 p5-595 P5+ 2.3 64 (64/128) 23,456 1.98 7,051 2,350,330 117.52 AIX 5.3 DB2 9.1 99% 5.0 2006045 595 P6 5.0 64 (32/128) 35,400 1.94 10,677 3,559,000 177.95 AIX 6.1 DB2 9.5 99% 6.0 2008019 750 P7 3.55 32 (4/128) 15,600 .99 5,113 1.704,330 85.22 AIX 6.1 DB@ 9.7 99% EHP4 *** * The SAP certification number was not available at press time and can be found at www.sap.com/benchmark. Sales and Distribution – SD 2-Tier Parallel # of Dialog Fully Nodes x Avg. Steps Proc. Avg. Cores Res Per Line Items CPU SAP (Proc/ Users p. Hour Per Hour Util. ECC Model (GHz) Threads) S&D Time (K) (K) SAPS OS Database % Ver. Cert. # p5-570 P5 1.9 2 x 4c (4/8) 2,400 1.95 723 241 12050 AIX 5.3 Oracle 9i 99% 4.7 2005033 570 P6 4.7 5 x 16c (8/32) 36,000 1.76 11,021 3,673.67 183680 AIX.5.3 Oracle 10g r2 99% 6.0 2007066 570 P6 4.7 2 x 16c (8/32) 15,520 1.94 1,559.33 4,678 77,970 AIX 5.3 Oracle 10g r2 98% 6.0 2008010 570 P6 4.7 3 x 16c (8/32) 22,416 1.94 2,252.33 6,757 112,620 AIX 5.3 Oracle 10g r2 99% 6.0 2008011 570 P6 4.7 4 x 16c (8/32) 30,016 1.86 3,036 9,108 151,800 AIX 5.3 Oracle 10g r2 99% 6.0 2008012 570 P6 4.7 5 x 16c (8/32) 37,040 1.86 3,749 11,247 187,450 AIX 5.3 Oracle 10g r2 99% 6.0 2008013 Sales and Distribution – SD 3-Tier Page 22
- 23. Fully # Avg. Proc. Dialog CPU Cores Res Line Items Steps Util. SAP Model (Proc/ Users p. Per Hour Per Hour DB ECC (GHz) Thread) Time (K) (K) SAPS OS Database (%) Ver. Cert. # JS12 (3.8) 2 (1/4) 14,000 1.95 1,406.33 4,219 70,320 AIX 6.1 DB2 9.5 97 6.0 2008031 550 P6 (4.2) 4 (2/8) 32,000 1.89 3,230.33 9,691 161,520 AIX.5.3 DB2 9.5 99 6.0 2008001 #p5-570 P5 4 (4/8) 21,712 1.96 2,178.67 6,536 108,930 AIX 5.3 DB2 UDB V8.2 97 4.70 2004076 #p5-595 P5 32 (32/64) 168,300 1.95 16,896.67 50,690 844,330 AIX 5.3 DB2 UDB V8.2.2 99 4.70 2005021 BI Data Mart # Core Query CPU (Proc/ Steps/ Util. OS Platform Model Thread) Hour (%) Database Release Cert. # #i5-520 1.9 GHz 2 (2/4) 26,224 97% i5/OS V5R4 DB2 for i5/OS V5R4 NW 7.0 (2004s) 2007061 520 P6 4.2 GHz 2 (1/4) 41,297 96% i 6.1 DB2 for IBM i 6.1 NW 7.0 (2004s) 2008020 #i5-570 2.2 GHz 4 (4/8) 51,875 98% i5/OS V5R4 DB2 for i5/OS V5R4 NW 2004s 2007003 #i5-570 2.2 GHz 8 (8/16) 114,687 95% i5/OS V5R4 DB2 for i5/OS V5R4 NW 2004s 2007027 570 P6 4.7 GHz 4 (2/8) 92,716 98% i5/OS V5R4M5 DB2 for i5/OS V5R4M5 NW 7.0 (2004s) 2007047 BI Mixed Load (MXL) # Core Query CPU (Proc/ Steps/ Util. OS Platform Model Thread) Hour (%) Database Release Cert. # 520 P6 4.7 GHz 2 (1/4) 41,090 94% i 6.1 DB2 for IBM i 6.1 NW 7.0 (2004s) 2009002 550 P6 5.0 GHz 4 (2/8) 90,492 97% i 6.1 DB2 for IBM I 6.1 NW 7.0 (2004s) 2009010 550 P6 5.0 GHz 4 (2/8) 90,634 98% i 6.1 DB2 for IBM I 6.1 NW 7.0 (2004s) 2009011 570 P6 5.0 GHz 4 (2/8) 93,468 98% i 6.1 DB2 for IBM i 6.1 NW 7.0 (2004s) 2008069 570 P6 5.0 GHz 8 (4/16) 154,447 97% i 6.1 DB2 for IBM i 6.1 NW 7.0 (2004s) 2009043 750 P7 3.30 Ghz 8 1/16) 241,526 *** i 6.1 DB2 for IBM i 6.1 NW 7.0 (2004s) *** Section 8 - AIX PeopleSoft Benchmarks Published Results HRMS Self-service Online PS # Cores Concurrent Version Model - MHz Users Search Save Database 8.9 p5-570 12 - 1900 4,000 1.74 sec. 1.25 sec. Oracle 10.1.0.3 Global Payroll France – Version 1 PS # Cores Maximum Payroll Payroll Version Model - MHz Payees Non Retro Retro Database 8.9 p570 8 - 4700 200,222 44.95 73.56 Oracle 9i Global Payroll France – Version 2 L2/L3 PS # Cores Cache Memory Disk Version Model - MHz (MB) (GB) (GB) Large Rate Database 8.9 p570 8 - 4700 32/128 84 550 200,222 Oracle 9.2.0.6 North American Payroll - Checks per Hour Page 23
- 24. PS # Cores L2/L3 Cache Memory Disk Large Version Model - MHz (MB) (GB) (GB) Rate Database 8.8 p5-570 8 - 1900 0.95/18 64 3,931 393,000 DB2 V8.1 Section 9 - AIX Oracle e-Business Suite (eBS) Benchmarks Published Results R12 – 12.0.4 Batch Results Cores Batch : (Proc / Batch: Payroll Average Checks Model GHz workers) Kit Employees CPU util. i/o (sec) RUP level /Hour 750 P7 3.30 6 (1/12) Medium 10,000 55% 300.4 4 257,143 11i – 11.5.10 Online : Users / Online : Batch : Order Lines / User load Average Batch : # Batch: Payroll % of Response Batch : Checks Model GHz Cores Kit Employees maximum Time (sec) Lines/Hour /Hour p5-570 P5+ 2.2 8 Medium 2,000 / 50,000 / 10,000 100% (full) 0.983 56,391 51,948 p5-570 P5+ 2.2 8 Medium 1,800 / 50,000 / 10,000 90% 0.857 59,678 56,818 p5-570 P5+ 2.2 8 Medium 1,400 / 50,000 / 10,000 70% 0.712 61,894 65,076 p570 P6 4.7 8 Medium 3000 / 50,000 / 10,000 100% (full) 0.764 94,757 74,257 p570 P6 4.7 8 Medium 2,700 / 50,000 / 10,000 90% 0.702 97,784 84,270 p570 P6 4.7 8 Medium 2,100 / 50,000 / 10,000 70% 0.625 106,838 91,047 11i – 11.5.10 Real Application Clusters (RAC) # Online Nodes Online : Users / User load Average Batch : x# Batch : Order Lines / % of Response Batch : Checks Model GHz Cores Kit Batch: Payroll Employees maximum Time (sec) Lines/Hour /Hour p5-505 P5+ 2.1 2 x 2c Small 1,000 / 10,000 / 5,000 100% (full) 0.780 11,080 13,043 p5-505 P5+ 2.1 2 x 2c Small 900 / 10,000 / 5,000 90% 0.758 12,280 17,493 p5-505 P5+ 2.1 2 x 2c Small 700 / 10,000 / 5,000 70% 0.656 17,207 26,087 11i – 11.5.9 # Average Response Model GHz Cores Users Time (sec) Release #p5-570 P5 1.9 8 15,004 0.553 11.5.9 Section 10 - AIX Siebel Benchmarks Published Results Siebel 7.7 Industry Applications Performance and Scalability Benchmark # Cores / App./Gateway # Cores / DB Server Memory Servers Memory Concurrent Users Database #p5-570 P5 4/32GB 5x p690 16/64GB 12,500 DB2 UDB V8.1 Siebel CRM Release 8.0 Industry Applications Benchmark Gateway / App Server DB Server Application # Cores / Concurrent DB # Cores / Servers Memory Users %CPU Server Memory Database 1x p570 P6 4.7 GHz 8 / 64GB 7,000 84% 1x p570 P6 4.7 GHz 4 / 32GB Oracle 10gR2 Page 24
- 25. Section 11 - AIX Sybase Benchmarks Published Results Sybase Risk Analysis Platform Benchmark Load Test with Query Test: Cumulative # Cores / 16-Stream Binary Data Query Processing Times DB Server Memory (GB/Hour/Core) @ 50 Users (lower is better) Database p5-570 P5+ 2.2 GHz 8 / 32GB 7.407 2,308,842 msec Sybase IQ p570 P6 4.7 GHz 4 / 32GB 6.692 3,376,891 msec Sybase IQ p570 P6 4.7 GHz 8 / 32GB 11.254 1,459,035 msec Sybase IQ Section 12 - AIX Manugistics Benchmarks Published Results Manugistics NetWORKS Fulfillment Benchmark Server # Cores SKUs/Hour SKUs/Hour/Core Release #p5-590 32 38,475,727 1,202,366 7.2 p5-595 2.3 GHz 64 116,500,032 1,820,313 7.3 Section 13 - Linux Published Benchmark Results SPEC CPU2006 Performance SPEC SPEC L1 L2/L3 SPEC int_ SPEC Fp_ Proc / Cache Cache int Base Fp base Linux Model #Cores GHz (KB) (MB) 2006 2006 2006 2006 Version JS12 P6/2 3.8 64/64 16/- 16.1 13.6 17.9 14.2 RHEL5.1 570 P6/1 4.7 64/64 8/32 21.7 17.8 22.5 18.1 RHEL5.1 570 P6/1 4.7 64/64 8/32 21.3 17.5 22.4 17.8 SLES10 SP1 SPEC CPU2000 Performance SPEC SPEC SPEC L1 L2/L3 int_ SPEC fp_ Proc / Cache Cache int base fp Base Linux Model #Cores GHz (KB) (MB) 2000 2000 2000 2000 Version #OP710 P5/1 1.65 64/32 1.9/36 1,144 1,129 1,919 1,828 RHEL AS4 #OP720 P5/1 1.65 64/32 1.9/36 1,138 1,121 1,966 1,865 SLES9 p5-505 P5+/1 2.1 64/32 1.9/36 1655 1594 3293 2773 SLES10 p5-510 P5+/1 2.1 64/32 1.9/36 1655 1594 3293 2773 SLES10 p5-520 P5+/1 2.1 64/32 1.9/36 1655 1595 3283 2772 SLES10 p5-520Q P5+/1 1.65 64/32 1.9/72 1302 1255 2580 2152 SLES10 p5-550 P5+/1 2.1 64/32 1.9/36 1656 1596 3282 2778 SLES10 p5-550Q P5+/1 2.1 64/32 1.9/72 1303 1256 2573 2157 SLES10 p5-575 P5+/1 2.2 64/32 1.9/36 1730 1666 3418 2896 SLES10 p5-575 P5+/1 1.9 64/32 1.9/36 1501 1445 2979 2543 SLES10 Multiuser Performance SPEC CPU2006 SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor/ Cache Cache rate base rate base Linux Model # Cores GHz (KB) (MB) 2006 2006 2006 2006 Version JS12 P6/2 3.8 64/64 16/- 45.9 41.2 42.5 36.2 RHEL5.1 JS22 P6/4 4.0 64/64 16/- 84.7 77.2 75.0 65.7 SLES0 SP1 JS23 P6+/4 4.2 64/64 16/64 110 92.2 92.4 75.8 SLES11 JS43 P6+/8 4.2 64/64 32/128 219 184 184 151 SLES11 Page 25
- 26. SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor/ Cache Cache rate base rate base Linux Model # Cores GHz (KB) (MB) 2006 2006 2006 2006 Version 520 P6/4 4.2 64/64 16/- 89.2 81.2 79.7 69.0 SLES10 SP1 520 P6+/4 4.7 64/64 16/64 122 102 102 83.7 SLES11 550 P6/8 4.2 64/64 32/128 213 182 176 151 RHEL5.1 550 P6+/8 5.0 64/64 32/128 258 216 216 178 SLES11 750 P7/32 3.30 32/32 8/128 1030 924 839 736 SLES 11 750 P7/32 3.55 32/32 8/128 1070 960 865 761 SLES 11 755 P7/32 3.30 32/32 8/128 1030 924 839 736 SLES 11 570 P6/4 4.7 64/64 16/64 122 108 116 98.8 RHEL5.1 570 P6/4 4.7 64/64 16/64 118 105 115 97.5 SLES10 SP1 570 P6/8 4.7 64/64 32/128 243 210 216 185 RHEL5.1 570 P6/8 4.7 64/64 32/128 234 204 215 182 SLES10 SP1 570 P6/16 4.7 64/64 64/256 484 420 430 369 RHEL5.1 570 P6/16 4.7 64/64 64/256 466 407 428 364 SLES10 SP1 570/32 P6+/32 4.2 64/64 128/512 826 673 592 492 RHEL 5.2 575 P6/32 4.7 64/64 128/512 928 809 813 681 RHEL 5.2 595 P6/64 5.0 64/64 256/1024 2,097 1,837 2,108 1,822 RHEL 5.2 595 P6/64 5.0 64/64 256/1024 2,160 1,870 2,180 1,680 SLES11 Multiuser Performance SPEC CPU2000 SPEC SPEC SPEC int_ SPEC fp_ L1 L2/L3 int_ rate_ fp_ rate_ Processor/ Cache Cache rate base rate base Linux Model # Cores GHz (KB) (MB) 2000 2000 2000 2000 Version #OP710 P5/2 1.65 64/32 1.9/36 -- -- 40.2 39.5 SLES9 #OP710 P5/2 1.65 64/32 1.9/36 29.8 29.5 40.1 39.0 RHEL AS4 #OP720 P5/4 1.65 64/32 3.8/72 59.8 58.8 80.8 78.8 SLES9 p5-505 P5+/2 2.1 64/32 1.9/36 43.5 42.4 72.4 66.5 SLES10 p5-510 P5+/2 2.1 64/32 1.9/36 43.5 42.4 72.4 66.5 SLES10 p5-520 P5+/2 2.1 64/32 1.9/36 43.6 42.6 71.7 66.0 SLES10 p5-520Q P5+/4 1.65 64/32 3.8/72 68.2 66.6 99.0 92.9 SLES10 p5-550 P5+/4 2.1 64/32 1.9/72 86.7 85.0 143 131 SLES10 p5-550Q P5+/8 1.65 64/32 7.6/144 136 133 196 183 SLES10 #p5-575 P5/8 1.9 64/32 15.2/288 -- -- 238 229 RHEL AS4 p5-575 P5+/8 2.2 64/32 1.9/288 199 193 370 310 SLES10 p5-575 P5+/16 1.9 64/32 1.9/288 311 305 541 478 SLES10 #p5-595 P5/32 1.9 64/32 30.4/576 -- -- 781 754 SLES9 SP1 Java Performance (SPECjbb2000, SPECjbb2005) SPEC jbb2005 Linux L1 L2/L3 SPEC ops/sec Version Processor / Cache Cache jbb2000 JVM bops / Model # Cores GHz (KB) (MB) ops/sec bops inst. JVM #OP720 P5/4 1.65 64/32 3.8/72 136,167 - - - SLES9 #OP720 P5/4 1.65 64/32 3.8/72 136,261 - - - RHEL AS4 550 P6/8 4.2 64/64 32/128 - 328,343 4 82,086 RHEL5.1 #p5-570 P5/2 DDR2 1.9 64/32 1.9/36 82,615 - - - SLES9 #p5-570 P5/4 DDR2 1.9 64/32 3.8/72 160,995 - - - SLES9 #p5-570 P5/8 DDR1 1.9 64/32 7.6/144 299,197 - - - SLES9 #p5-570 P5/16 DDR1 1.9 64/32 15.2/288 542,145 - - - SLES9 750 P7/32 3.55 32/32 8/128 - 2,410,483 32 75,328 SLES 11 Page 26
- 27. SPEC jbb2005 Linux L1 L2/L3 SPEC ops/sec Version Processor / Cache Cache jbb2000 JVM bops / Model # Cores GHz (KB) (MB) ops/sec bops inst. JVM 570 P6/4 4.7 64/64 16/64 - 169,304 2 84,652 RHEL5.1 570 P6/8 4.7 64/64 32/128 - 335,424 4 83,856 RHEL5.1 570 P6/16 4.7 64/64 64/256 - 664,167 8 83,021 RHEL5.1 570/32 P6+/32 4.2 64/64 128/512 1,243,824 16 77,739 RHEL5.2 #p5-595 P5/32 1.9 64/32 30.4/576 1,076,309 - - - SLES9 Web Serving SPECweb99 and SPECweb99_SSL Performance L1 L2/L3 SPEC Proc./ Cache Cache Encryption SPEC web99 Linux Model # Cores GHz (KB) (MB) card web99 ssl Version #p5-570 P5/4 1.9 64/32 3.8/72 Yes, ICA - 4,970 SLES9 #p5-570 P5/4 1.9 64/32 3.8/72 None 13,500 - RHEL AS3 #p5-570 P5/8 1.9 64/32 7.6/144 None 25,000 - RHEL AS4 Web Serving SEPCweb2005 Performance L1 L2/L3 Proc./ Cache Cache Encryption Linux Model # Cores GHz (KB) (MB) card Result Banking Ecommerce Support Version SLES9 #p5-550 P5+/4 1.9 64/32 3.8/72 None 7,881 12,240 11,820 7,500 SP2 SPECsfs97_R1 Benchmark Results L1 L2/L3 SPEC SPEC Proc./ Cache Cache sfs97_R1.v3 sfs97_R1.v3 Linux Model # Cores GHz (KB) (MB) UDP TCP Version #p5-570 P5/2 1.9 64/32 1.9/36 -- 45,586 SLES9 #p5-570 P5/4 1.9 64/32 3.8/72 -- 81,889 SLES9 SP1 NetBench® Published Results L1 L2/L3 Processor / Cache Cache Model # Cores GHz (KB) (MB) Mbps #p5-520 P5/1 1.65 64/32 1.9/36 787 #p5-520 P5/2 1.65 64/32 3.8/72 1,457 #p5-550 P5+/2 1.9 64/32 1.9/36 2,054 #p5-550 P5+/4 1.9 64/32 3.8/72 3,055 SAP Standard Application Benchmark Published Results Sales and Distribution – SD 2-Tier Fully Proc. Average Dialog Line CPU # Resp. Steps Items Util. Kit Model Core Users Time Per Hour SAPS Per Hour OS Database % Ver. Cert. # #p5-550 DB2 UDB 4 1000 1.97 301,000 5,020 100,330 SLES9 99 5.0 2005040 1.9 GHz P5 V8.2.2 #p5-570 DB2 UDB 8 2000 1.95 603,000 10,050 201,000 SLES9 99 4.70 2004057 1.9 GHz P5 V8.2.2 570 8 3104 1.91 938,000 15,630 312,670 RHEL5.1 DB2 9.5 97 6.0 2008002 4.2 GHz P6 Page 27
- 28. STREAM Benchmarks Standard Tuned L1 L2/L3 STREAM STREAM Processor / Cache Cache Triad Triad Linux Model # Cores GHz (KB) (MB) MB/sec MB/sec Version JS22 P6/4 4.0 64/64 16/- 15,701 -- SLES10 SP1 570 P6/4 4.7 64/64 16/64 29,404 -- RHEL5.1 575 P6/32 4.7 64/64 128/512 158,639 -- RHEL 5.2 TPC-C Version 5.4 Published Results L2/L3 Processor # Cache Linux Avail. Model / # Cores Nodes GHz (MB) tpmC $/tmpC Database Version Date #p5-570 P5/4 1 1.9 3.8/72 197,669.81 3.93 DB2 UDB V8.2 RHEL AS4 02/07/06 p5-570 P5+/4 1 2.2 3.8/72 236,271 2.56 Oracle 10g R2 EE RHEL5 04/04/08 #p5-520 P5+/2 1 1.65 1.9/36 81,439.30 2.99 Sybase ASE SLES9 12/22/06 550 P6/4 1 4.2 16/64 276,383 2.22 Sybase ASE RHEL 5.1 12/16/08 SPECjAppServer2004 Performance # J2EE J2EE J2EE J2EE Nodes / DB Database # DB Model/GHz OS AppServer Cores Model/GHz OS Database Nodes JOPS #p5-550 WebSphere 8 nodes DB2 UDB 1 node SLES 9 p5-570 / 1.9 SLES9 2921.48 1.9 GHz 6.0 32 cores V8.2.3 8 cores Page 28
- 29. Section 14 - Historical Multiuser Performance (2006 to present) Type Product Name Announce Marketing # Model Date Withdrawal Processor MHz Cores rPerf 9110-510 eServer p5 510 2005/02/18 2006/11/13 POWER5 1500 1 3.25 9110-510 POWER5 1500 2 9.13 9110-510 2006/05/31 POWER5 1650 1 5.24 9110-510 POWER5 1650 2 9.86 9110-51A System p5 510 2006/02/14 2007/02/12 POWER5+ 1900 1 6.11 9110-51A POWER5+ 1900 2 11.49 9110-51A System p5 510Q POWER5+ 1500 4 18.75 9111-285 IntelliStation 285 2005/10/04 2007/02/12 POWER5+ 1900 1 9111-285 POWER5+ 1900 2 9111-520 eServer p5 520 2004/07/13 2006/05/31 POWER5 1500 1 3.25 9111-520 POWER5 1500 2 9.13 9111-520 POWER5 1650 2 9.86 9112-265 IntelliStation 265 2002/02/05 2003/12/12 POWER3-II 450 2 N/A 9113-550 eServer p5 550 2004/07/13 2006/05/31 POWER5 1500 1 3.25 9113-550 POWER5 1500 2 9.13 9113-550 POWER5 1500 4 18.20 9113-550 POWER5 1650 2 9.86 9113-550 POWER5 1650 4 19.66 9114-275 IntelliStation 275 2003/06/24 2006/05/31 POWER4+ 1000 1 N/A 9114-275 POWER4+ 1450 1 N/A 9114-275 POWER4+ 1450 2 N/A 9115-505 System p5 505 2005/10/04 2006/01/13 POWER5+ 1500 2 9.13 9115-505 POWER5+ 1650 1 3.51 9115-505 POWER5+ 1650 2 9.86 9116-561 System p5 560Q 2006/02/14 2008/02/25 POWER5+ 1500 4 18.75 9116-561 POWER5+ 1500 8 35.50 9116-561 POWER5+ 1500 16 65.24 9117-570 eServer p5 570 2004/07/13 2006/05/31 POWER5 1500 2 9.13 9117-570 POWER5 1500 4 18.20 9117-570 POWER5 1500 8 34.46 9117-570 POWER5 1650 2 9.86 9117-570 POWER5 1650 4 19.66 9117-570 POWER5 1650 8 37.22 9117-570 POWER5 1650 12 53.43 9117-570 POWER5 1650 16 68.40 9117-570 POWER5 1900 2 11.16 9117-570 POWER5 1900 4 22.26 9117-570 POWER5 1900 8 42.14 9117-570 POWER5 1900 12 60.50 9117-570 POWER5 1900 16 77.45 9118-575 System p5 575 2005/02/08 2006/11/17 POWER5 1500 16 N/A 9118-575 2005/02/08 2006/11/10 POWER5 1900 8 N/A Page 29
- 30. Type Product Name Announce Marketing # Model Date Withdrawal Processor MHz Cores rPerf 9119-590 eServer p5 590 2004/10/15 2007/02/12 POWER5 1650 8 41.68 9119-590 POWER5 1650 16 80.86 9119-590 POWER5 1650 24 116.29 9119-590 POWER5 1650 32 151.72 9119-595 eServer p5 595 2004/10/15 2007/02/12 POWER5 1650 16 80.86 9119-595 POWER5 1650 24 116.29 9119-595 POWER5 1650 32 151.72 9119-595 POWER5 1650 40 182.07 9119-595 POWER5 1650 48 212.41 9119-595 POWER5 1650 56 242.76 9119-595 POWER5 1650 64 273.10 9119-595 POWER5 1900 16 90.67 9119-595 POWER5 1900 24 130.39 9119-595 POWER5 1900 32 170.11 9119-595 POWER5 1900 40 204.14 9119-595 POWER5 1900 48 238.16 9119-595 POWER5 1900 56 272.18 9119-595 POWER5 1900 64 306.21 eServer OpenPower 9123-710 2005/01/25 2006/05/31 POWER5 1650 1 N/A 710 POWER5 1650 2 N/A eServer OpenPower 9124-720 2004/09/14 2006/05/31 POWER5 1500 1 N/A 720 9124-720 POWER5 1500 2 N/A 9124-720 POWER5 1500 4 N/A 9124-720 POWER5 1650 2 N/A 9124-720 POWER5 1650 4 N/A 9131-52A System p5 520 2005/10/04 2007/02/12 POWER5+ 1650 1 3.62 9131-52A POWER5+ 1650 2 10.15 9131-52A POWER5+ 1900 2 11.16 9131-52A System p5 520Q POWER5+ 1500 4 18.75 9133-55A System p5 550 2005/10/04 2007/02/12 POWER5+ 1650 2 10.15 9133-55A POWER5+ 1650 4 20.25 9133-55A 2005/10/04 2007/04/27 POWER5+ 1500 4 18.20 9133-55A POWER5+ 1500 8 34.46 8842-42x BladeCenter JS20 6/28/05 5/31/06 PPC970 1600 1 1.53 PPC970 1600 2 2.65 PPC970 2200 1 1.95 PPC970 2200 2 3.40 9117-570 System p5 570 2/14/06 8/29/08 POWER5+ 1900 2 12.27 POWER5+ 1900 4 24.48 POWER5+ 1900 8 46.36 POWER5+ 1900 12 66.55 POWER5+ 1900 16 85.20 POWER5+ 2200 2 13.83 POWER5+ 2200 4 27.58 POWER5+ 2200 8 52.21 POWER5+ 2200 12 74.95 POWER5+ 2200 16 95.96 Page 30
- 31. Type Product Name Announce Marketing # Model Date Withdrawal Processor MHz Cores rPerf 9118-575 System p5 575 2/14/06 8/29/08 POWER5+ 1900 16 -- POWER5+ 2200 8 -- 9119-590 System p5 590 7/25/06 8/29/08 POWER5+ 2100 8 55.74 POWER5+ 2100 16 108.13 POWER5+ 2100 24 155.51 POWER5+ 2100 32 202.88 8844-31X BladeCenter JS21 6/5/07 12/12/08 PPC970 2700 2 5.31 8844-51X BladeCenter JS21 6/5/07 12/12/08 PPC970 2500 4 8.72 9110-51A System p5 510 7/25/06 12/12/08 POWER5+ 2.1 1 6.63 7/25/06 12/12/08 POWER5+ 2.1 2 12.46 9110-51A System p5 510Q 7/25/06 12/12/08 POWER5+ 1.65 4 20.25 9131-52A System p5 520 7/25/06 12/12/08 POWER5+ 2.1 1 6.63 7/25/06 12/12/08 POWER5+ 2.1 2 12.46 9131-52A System p5 520Q 7/25/06 12/12/08 POWER5+ 1.65 4 20.25 9133-55A System p5 550 10/4/05 12/12/08 POWER5+ 1.9 2 11.16 10/4/05 12/12/08 POWER5+ 1.9 4 22.26 7/25/06 12/12/08 POWER5+ 2.1 2 12.46 7/25/06 12/12/08 POWER5+ 2.1 4 24.86 9133-55A System p5 550Q 7/25/06 12/12/08 POWER5+ 1.65 4 20.25 7/25/06 12/12/08 POWER5+ 1.65 8 38.34 9111-285 IntelliStation 285 7/11/06 1/1/09 POWER5+ 2100 1 -- 7/11/06 1/1/09 POWER5+ 2100 2 -- 7047-185 IntelliStation 185 2/14/06 1/1/09 PPC970 2.5 1 -- 2/14/06 1/1/09 PPC970 2.5 2 -- 9115-505 System p5 505 7/25/06 5/29/09 POWER5+ 1900 1 4.10 7/25/06 5/29/09 POWER5+ 1900 2 11.49 7/25/06 5/29/09 POWER5+ 2100 2 12.46 System p5 505Q 7/25/06 5/29/09 POWER5+ 1650 4 20.25 9119-595 System p5 595 7/25/06 5/29/09 POWER5+ 2.1 16 108.13 7/25/06 5/29/09 POWER5+ 2.1 32 202.88 7/25/06 5/29/09 POWER5+ 2.1 48 284.04 7/25/06 5/29/09 POWER5+ 2.1 64 365.19 7/25/06 5/29/09 POWER5+ 2.1 16 116.53 7/25/06 5/29/09 POWER5+ 2.1 32 218.64 7/25/06 5/29/09 POWER5+ 2.1 48 306.10 7/25/06 5/29/09 POWER5+ 2.3 64 393.55 9116-561 System p5 560Q 2/14/07 1/2/09 POWER5+ 1.8 4 21.72 2/14/07 1/2/09 POWER5+ 1.8 8 41.12 2/14/07 1/2/09 POWER5+ 1.8 16 75.58 7998-60X BladeCenter JS12 4/2/08 -- POWER6 3.8 2 14.71 Page 31
- 32. Type Product Name Announce Marketing # Model Date Withdrawal Processor MHz Cores rPerf 7998-61X BladeCenter JS22 11/6/07 -- POWER6 4.0 4 30.26 BladeCenter JS23 4/28/09 -- POWER6+ 4.2 4 36.28 BladeCenter JS43 4/28/09 -- POWER6+ 4.2 8 68.20 8203-E4A Power 520 1/28/08 -- POWER6 4.2 1 8.39 1/28/08 -- POWER6 4.2 2 15.95 1/28/08 -- POWER6 4.2 4 31.48 4/28/09 -- POWER6+ 4.7 2 20.13 4/28/09 -- POWER6+ 4.7 4 39.73 8204-E8A Power 550 1/28/08 -- POWER6 3.5 2 15.85 1/28/08 -- POWER6 3.5 4 31.27 1/28/08 -- POWER6 3.5 6 45.04 1/28/08 -- POWER6 3.5 8 58.80 1/28/08 -- POWER6 4.2 2 18.38 1/28/08 -- POWER6 4.2 4 36.28 1/28/08 -- POWER6 4.2 6 52.24 1/28/08 -- POWER6 4.2 8 68.20 4/28/09 -- POWER6+ 5.0 2 21.18 4/28/09 -- POWER6+ 5.0 4 41.81 4/28/09 -- POWER6+ 5.0 6 60.20 4/28/09 -- POWER6+ 5.0 8 78.60 Power 560 10/7/08 -- POWER6+ 3.6 4 31.32 10/7/08 -- POWER6+ 3.6 8 57.32 10/7/08 -- POWER6+ 3.6 16 100.30 9117-MMA Power 570 5/22/07 -- POWER6 3.5 2 15.85 5/22/07 -- POWER6 3.5 4 31.69 5/22/07 -- POWER6 3.5 8 58.95 5/22/07 -- POWER6 3.5 12 83.35 5/22/07 -- POWER6 3.5 16 105.75 5/22/07 7/2/09 POWER6 4.2 2 18.38 5/22/07 7/2/09 POWER6 4.2 4 36.76 5/22/07 7/2/09 POWER6 4.2 8 68.38 5/22/07 7/2/09 POWER6 4.2 12 96.68 5/22/07 7/2/09 POWER6 4.2 16 122.67 10/7/08 -- POWER6+ 4.2 4 35.50 10/7/08 -- POWER6+ 4.2 8 64.96 10/7/08 -- POWER6+ 4.2 16 113.68 10/7/08 -- POWER6+ 4.2 24 153.46 10/7/08 -- POWER6+ 4.2 32 193.25 10/7/08 -- POWER6+ 4.4 2 19.08 10/7/08 -- POWER6+ 4.4 4 38.16 10/7/08 -- POWER6+ 4.4 8 70.97 10/7/08 -- POWER6+ 4.4 12 100.35 10/7/08 -- POWER6+ 4.4 16 127.32 5/22/07 7/2/09 POWER6 4.7 2 20.13 5/22/07 7/2/09 POWER6 4.7 4 40.26 5/22/07 7/2/09 POWER6 4.7 8 74.89 5/22/07 7/2/09 POWER6 4.7 12 105.89 5/22/07 7/2/09 POWER6 4.7 16 134.35 10/7/08 -- POWER6+ 5.0 2 21.16 10/7/08 -- POWER6+ 5.0 4 42.32 Page 32
- 33. Type Product Name Announce Marketing # Model Date Withdrawal Processor MHz Cores rPerf 10/7/08 -- POWER6+ 5.0 8 78.71 10/7/08 -- POWER6+ 5.0 12 111.3 10/7/08 -- POWER6+ 5.0 16 141.21 9125-F2A Power 575 4/8/08 -- POWER6+ 4.7 32 -- 9119-FHA Power 595 4/8/08 -- POWER6 4.2 8 75.58 4/8/08 -- POWER6 4.2 16 142.90 4/8/08 -- POWER6 4.2 32 266.51 4/8/08 -- POWER6 4.2 48 373.60 4/8/08 -- POWER6 4.2 64 479.89 4/8/08 -- POWER6 5.0 8 87.10 4/8/08 -- POWER6 5.0 16 164.67 4/8/08 -- POWER6 5.0 32 307.12 4/8/08 -- POWER6 5.0 48 430.53 4/8/08 -- POWER6 5.0 64 553.01 Note: The rPerf projections are based on different levels of AIX and databases. As a result, actual performance may vary. Estimates have been provided where no historical projections were available. Page 33
- 34. Notes on Performance Benchmarks and Values The performance benchmarks and the values shown here were derived using particular, well configured, development-level computer systems. Unless otherwise indicated for a system, the values were derived using external cache if external cache is supported on the system. All performance benchmark values are provided "AS IS" and no warranties or guarantees are expressed or implied by IBM. Actual system performance may vary and is dependent upon many factors including system hardware configuration and software design and configuration. Buyers should consult other sources of information to evaluate the performance of systems they are considering buying and should consider conducting application oriented testing. For additional information about the performance benchmarks, values and systems tested, please contact your IBM local Branch Office or IBM Authorized Reseller or access the following on the Web: SPEC - http://www.spec.org TPC - http://www.tpc.org All performance measurements for the IBM System p, System p5, eServer p5, eServer pSeries and RS/6000 servers were made with systems running AIX or AIX 5L operating systems unless otherwise indicated to have used Linux. For new and upgraded systems, AIX Version 4.3 or AIX 5L were used. All other systems used previous versions of AIX. The SPEC CPU2006 and SPEC CPU2000 benchmarks were compiled using IBM's high performance C, C++, and FORTRAN compilers for AIX 5L and Linux. For new and upgraded systems, the latest versions of these compilers were used: XL C Enterprise Edition V9.0 for AIX, XL C/C++ Enterprise Edition V9.0 for AIX, XL FORTRAN Enterprise Edition V11.1 for AIX, XL C/C++ Advanced Edition V8.0 for Linux, and XL FORTRAN Advanced Edition V10.1 for Linux. The following SPEC benchmarks reflect the performance of the microprocessor, memory architecture and compiler of the tested system: SPECint2006 - New SPEC component-level benchmark that measures integer performance. Result is the geometric mean of twelve tests that comprise the CINT2006 benchmark suite. SPECint_base2006 - The result of the same tests in CINT2006 with the same compiler options that must be used in all twelve tests. SPECint_rate2006 - Geometric average of the twelve SPEC rates from the SPEC integer tests (CINT2006). SPECint_rate_base2006 - The result of the same tests as CINT2006 with the same compiler options that must be used in all twelve tests. SPECfp2006 - New SPEC component-level benchmark that measures floating-point performance. Result is the geometric mean of seventeen tests, all written in FORTRAN and C languages, that are included in the CFP2006 benchmark suite. SPECfp_base2006 - The result of the same tests in CFP2006 with the same compiler options that must be used in all seventeen tests. SPECfp_rate2000 - Geometric mean of the seventeen SPEC rates from SPEC floating-point tests (CFP2006). SPECfp_rate_base2000 - The result of the same tests as CFP2000 with the same compiler options that must be used in all seventeen tests. SPECint2000 - SPEC component-level benchmark that measures integer performance. Result is the geometric mean of twelve tests that comprise the CINT2000 benchmark suite. All of these are written in C language except for one which is in C++. SPECint_base2000 - The result of the same tests in CINT2000 with a maximum of four compiler options that must be used in all twelve tests. SPECint_rate2000 - Geometric average of the twelve SPEC rates from the SPEC integer tests (CINT2000). SPECint_rate_base2000 - The result of the same tests as CINT2000 with a maximum of four compiler options that must be used in all twelve tests. SPECfp2000 - SPEC component-level benchmark that measures floating-point performance. Result is the geometric mean of fourteen tests, all written in FORTRAN and C languages, that are included in the CFP2000 benchmark suite. Page 34
- 35. SPECfp_base2000 - The result of the same tests in CFP2000 with a maximum of four compiler options that must be used in all fourteen tests. SPECfp_rate2000 - Geometric mean of the fourteen SPEC rates from SPEC floating-point tests (CFP2000). SPECfp_rate_base2000 - The result of the same tests as CFP2000 with a maximum of four compiler options that must be used in all fourteen tests. SPEC OMP2001 - Geometric mean 11 compute intensive parallel workload tests, written in Fortran and C languages. SPECweb99 - Number of conforming, simultaneous connections the Web server can support using a predefined workload. The SPECweb99 test harness emulates clients sending the HTTP requests in the workload over slow Internet connections to the Web server. The Web server software is Zeus from Zeus Technology Ltd. SPECweb2005 - Emulates users sending browser requests over broadband Internet connections to a Web server. It provides three new workloads: a banking site (HTTPS), an e-commerce site (HTTP/HTTPS mix); and a support site (HTTP. SPECweb99_SSL - Number of conforming, simultaneous SSL encryption/decryption connections the Web server can support using a predefined workload. The Web server software is Zeus from Zeus Technology Ltd. SPECjvm99 - Contains eight different tests. Each test measures the time it takes to load the program, verify the class files, compile on the fly if a JIT compiler is used, and execute the test. A geometric mean is used to compute a composite score. Test scores are normalized against a reference machine. Higher scores indicate better performance. SPECjbb2000 - Expressed in operations per second; evaluates the performance of servers running typical Java business applications; it represents an order processing application for a wholesale supplier. The benchmark can be used to evaluate performance of hardware and software aspects of Java Virtual Machine (JVM) servers. SPECjbb2005 - Expressed in bops and bops/JVM; evaluates the performance of servers running typical Java business applications; it represents an order processing application for a wholesale supplier. The benchmark can be used to evaluate performance of hardware and software aspects of Java Virtual Machine (JVM) servers. SPECsfs97_R1 - Measures speed and request-handling capabilities of NFS (network file server) computers. SPECjAppServer2004 - Measures the performance of Java Enterprise Application Servers using a subset of J2EE APIs in a complete end-to-end Web application. The following Transaction Processing Council (TPC) benchmarks reflect the performance of the microprocessor, memory subsystem, disk subsystem and some portions of the network: tpmC - TPC Benchmark C throughput measured as the average number of transactions processed per minute during a valid TPC-C configuration run of at least twenty minutes. $/tpmC - TPC Benchmark C price-performance ratio reflects the estimated five year total cost of ownership for system hardware, software and maintenance and is determined by dividing such estimated total cost by the tpmC for the system. QppH - The power metric of TPC-H and is based on a geometric mean of the 17 TPC-H queries, the insert test and the delete test. It measures the ability of the system to give a single user the best possible response time by harnessing all available resources. QppH is scaled based on database size from 30GB to 1TB. QthH - The throughput metric of TPC-H and is a classical throughput measure characterizing the ability of the system to support a multiuser workload in a balanced way. A number of query users is chosen, each of which must execute the full set of 17 queries in a different order. In the background, there is an update stream that runs a series of insert/delete operations. QthH is scaled based on the database size from 30GB to 1TB. QphH is the geometric mean of the power tests (QppH) and the throughput tests (QthH). $/QphH - The price/performance metric for the TPC-H benchmark where QphH is the geometric mean of QppH and QthH. The price is the five year cost of ownership for the tested configuration and includes maintenance and software support. Page 35
- 36. Notes on Performance Estimates rPerf (Relative Performance) is an estimate of commercial processing performance relative to other IBM UNIX systems. It is derived from an IBM analytical model which uses characteristics from IBM internal workloads, TPC and SPEC benchmarks. The rPerf model is not intended to represent any specific public benchmark results and should not be reasonably used in that way. The model simulates some of the system operations such as CPU, cache and memory. However, the model does not simulate disk or network I/O operations. rPerf estimates are calculated based on systems with the latest levels of AIX and other pertinent software at the time of system announcement. Actual performance will vary based on application and configuration specifics. The IBM eServer pSeries 640 is the baseline reference system and has a value of 1.0. Although rPerf may be used to approximate relative IBM UNIX commercial processing performance, actual system performance may vary and is dependent upon many factors including system hardware configuration and software design and configuration. Note that the rPerf methodology used for the POWER6 systems is identical to that used for the POWER5 systems. Variations in incremental system performance may be observed in commercial workloads due to changes in the underlying system architecture. Commercial Processing Workload (CPW) is a relative measure of performance of processors running the IBM i operating system. Performance in client environments may vary. The value is based on maximum configurations. More performance information is available in the Performance Capabilities Reference at: http://www.ibm.com/systems/i/solutions/perfmgmt/resource.html . All performance estimates are provided "AS IS" and no warranties or guarantees are expressed or implied by IBM. Buyers should consult other sources of information, including system benchmarks, and application sizing guides to evaluate the performance of a system they are considering buying. For additional information about rPerf and CPW, contact your local IBM office or IBM authorized reseller. IBM withdrew Relative OLTP (ROLTP). Starting June 2001, IBM will not publish/update ROLTP results. ROLTP results of systems that are withdrawn from the market are left in Section 14, Historical Multiuser Performance. Application Benchmarks SAP - Benchmark overview information: http://www.sap.com/benchmark/ PeopleSoft - To get information on PeopleSoft benchmarks, contact PeopleSoft directly or the PeopleSoft/IBM International Competency Center in San Mateo, CA. Oracle Applications - Benchmark overview information: http://www.oracle.com/apps_benchmark/html/results.html NotesBench - The driver program to test various aspects of Lotus® Notes®. It is designed to execute the commands in customized workload scripts, simulating Notes client actions. Source: http://www.notesbench.org/. Total Users - Number of active users supported in the workload, each producing approximately one transaction/minute. TPM - Transactions per minute (NotesMark) Average Response Time - Average time for a transaction to be completed for an average user action. $/User - Total cost of the hardware and software including discounts quoted by a supplier. Page 36
- 37. © IBM Corporation 2010 IBM Corporation Marketing Communications Systems and Technology Group Route 100 Somers, New York 10589 Produced in the United States of America February 2010 All Rights Reserved This document was developed for products and/or services offered in the United States. IBM may not offer the products, features, or services discussed in this document in other countries. The information may be subject to change without notice. Consult your local IBM business contact for information on the products, features and services available in your area. All statements regarding IBM future directions and intent are subject to change or withdrawal without notice and represent goals and objectives only. IBM, the IBM logo, AIX, AIX 5L, BladeCenter, DB2, eServer, Lotus, Notes, Power, POWER, POWER2, POWER3, POWER4, POWER4+, POWER5, POWER5+, POWER6, POWER PC, Power Architecture, Power Systems, PowerPC 601, PowerPC 604, pSeries, RS/6000, SP, System I, System p, System p5 and WebSphere are trademarks or registered trademarks of International Business Machines Corporation in the United States or other countries or both. A full list of US trademarks owned by IBM may be found at http://www.ibm.com/legal/copytrade.shtml. UNIX is a registered trademark of The Open Group in the United States, other countries or both. Linux is a registered trademark of Linus Torvalds in the United States, other countries or both. Java and all Java-based trademarks and logos are trademarks of Sun Microsystems, Inc. in the United States and/or other countries. TPC-C is a trademark of the Transaction Processing Council (TPPC). SPECint, SPECfp, SPECjbb, SPECweb, SPECjAppServer, SPEC OMP, SPECviewperf, SPECapc, SPEChpc, SPECjvm, SPECmail, SPECimap and SPECsfs are trademarks of the Standard Performance Evaluation Corp (SPEC). NetBench is a registered of Ziff Davis Media in the Untied States, other countries or both. AltiVec is a trademark of Freestyle Semiconductor, Inc. Other company, product, and service names may be trademarks or service marks of others. IBM hardware products are manufactured from new parts, or new and used parts. Regardless, our warranty terms apply. This equipment is subject to FCC rules. It will comply with the appropriate FCC rules before final delivery to the buyer. Information concerning non-IBM products was obtained from the suppliers of these products or other public sources. Questions on the capabilities of the non-IBM products should be addressed with the suppliers. All performance information was determined in a controlled environment. Actual results may vary. Performance information is provided “AS IS” and no warranties or guarantees are expressed or implied by IBM. The IBM home page on the Internet can be found at http://www.ibm.com. The Power Systems home page on the Internet can be found at http://www.ibm.com/systems/power. The BladeCenter home page on the Internet can be found at http://www.ibm.com/ystems/bladecenter. POO03017USEN-08 Page 37