| CPU2006 license: | 11 | Test date: | Aug-2010 |
|---|---|---|---|
| Test sponsor: | IBM Corporation | Hardware Availability: | Sep-2010 |
| Tested by: | IBM Corporation | Software Availability: | Aug-2010 |
| Hardware | |
|---|---|
| CPU Name: | POWER7 |
| CPU Characteristics: | TurboCore mode |
| CPU MHz: | 4256 |
| FPU: | Integrated |
| CPU(s) enabled: | 128 cores, 32 chips, 4 cores/chip, 4 threads/core |
| CPU(s) orderable: | 48 - 128 cores |
| Primary Cache: | 32 KB I + 32 KB D on chip per core |
| Secondary Cache: | 256 KB I+D on chip per core |
| L3 Cache: | 4 MB I+D on chip per core |
| Other Cache: | 16 MB I+D on chip per chip |
| Memory: | 2 TB (256x8 GB) DDR3 1066 MHz |
| Disk Subsystem: | 17x146.8 GB Raid0 SAS SFF 15K RPM |
| Other Hardware: | None |
| Software | |
|---|---|
| Operating System: | SUSE Linux Enterprise Server 11 SP1 (ppc64), Kernel 2.6.32.12-0.7-ppc64 |
| Compiler: | IBM XL C/C++ for Linux, V11.1 IBM XL Fortran for Linux, V13.1 |
| Auto Parallel: | No |
| File System: | xfs |
| System State: | Run level 5 (multi-user) |
| Base Pointers: | 32-bit |
| Peak Pointers: | 32/64-bit |
| Other Software: | -Post-Link Optimization for Linux on POWER, Version 5.5.0-3 -MicroQuill SmartHeap 9 -Apache C++ Standard Library V4.2.1 |
Results Table
| Benchmark | Base | Peak | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Copies | Seconds | Ratio | Seconds | Ratio | Seconds | Ratio | Copies | Seconds | Ratio | Seconds | Ratio | Seconds | Ratio | |
| Results appear in the order in which they were run. Bold underlined text indicates a median measurement. | ||||||||||||||
| 410.bwaves | 512 | 802 | 8680 | 804 | 8650 | 806 | 8630 | 512 | 802 | 8680 | 804 | 8650 | 806 | 8630 |
| 416.gamess | 512 | 2210 | 4540 | 2213 | 4530 | 2212 | 4530 | 512 | 2105 | 4760 | 2100 | 4770 | 2110 | 4750 |
| 433.milc | 512 | 566 | 8300 | 570 | 8250 | 568 | 8270 | 512 | 566 | 8300 | 570 | 8250 | 568 | 8270 |
| 434.zeusmp | 512 | 973 | 4790 | 993 | 4690 | 981 | 4750 | 512 | 973 | 4790 | 993 | 4690 | 981 | 4750 |
| 435.gromacs | 512 | 994 | 3680 | 996 | 3670 | 994 | 3680 | 512 | 767 | 4760 | 757 | 4830 | 764 | 4780 |
| 436.cactusADM | 512 | 1075 | 5690 | 1069 | 5720 | 1067 | 5740 | 512 | 866 | 7070 | 874 | 7000 | 861 | 7110 |
| 437.leslie3d | 512 | 1091 | 4410 | 1099 | 4380 | 1096 | 4390 | 128 | 267 | 4510 | 265 | 4530 | 266 | 4530 |
| 444.namd | 512 | 699 | 5870 | 701 | 5860 | 699 | 5880 | 512 | 682 | 6020 | 681 | 6030 | 683 | 6010 |
| 447.dealII | 512 | 674 | 8690 | 680 | 8610 | 676 | 8660 | 512 | 526 | 11100 | 523 | 11200 | 524 | 11200 |
| 450.soplex | 512 | 1161 | 3680 | 1160 | 3680 | 1162 | 3670 | 256 | 569 | 3750 | 570 | 3740 | 573 | 3730 |
| 453.povray | 512 | 540 | 5040 | 548 | 4970 | 543 | 5020 | 512 | 524 | 5200 | 520 | 5230 | 519 | 5240 |
| 454.calculix | 512 | 989 | 4270 | 955 | 4420 | 957 | 4410 | 512 | 961 | 4400 | 953 | 4430 | 958 | 4410 |
| 459.GemsFDTD | 512 | 1431 | 3800 | 1437 | 3780 | 1443 | 3760 | 512 | 1322 | 4110 | 1318 | 4120 | 1330 | 4080 |
| 465.tonto | 512 | 1571 | 3210 | 1571 | 3210 | 1572 | 3200 | 512 | 897 | 5620 | 855 | 5890 | 855 | 5900 |
| 470.lbm | 512 | 713 | 9860 | 707 | 9950 | 705 | 9980 | 512 | 713 | 9860 | 707 | 9950 | 705 | 9980 |
| 481.wrf | 512 | 1312 | 4360 | 1294 | 4420 | 1312 | 4360 | 512 | 1055 | 5420 | 1025 | 5580 | 1026 | 5580 |
| 482.sphinx3 | 512 | 1730 | 5770 | 1730 | 5770 | 1732 | 5760 | 512 | 1700 | 5870 | 1701 | 5870 | 1694 | 5890 |
Peak Tuning Notes
fdpr binary optimization tool used for:
433.milc 435.gromacs 436.cactusADM 450.soplex 482.sphinx3
with options -O4 -nodp
434.zeusmp
with options -O4 -vrox -nodp
437.leslie3d 444.namd
with options -O3 -lu -1 -nodp -sdp 9
465.tonto
with options -O4
Submit Notes
The config file option 'submit' was used. Benchmarks bound to a processor using numactl on the submit command.
Operating System Notes
ulimit -s (stack) set to 1048576.
Large pages reserved as follows by root user:
echo 34304 > /proc/sys/vm/nr_hugepages
The following environment varibles were set before the runspec command:
export HUGETLB_VERBOSE=0
export HUGETLB_MORECORE=yes
export HUGETLB_ELFMAP=W
export XLFRTEOPTS=intrinthds=1
General Notes
447.dealII (peak): "apache_stdcxx_4_2_1" src.alt was used. The Apache C++ Standard Library V4.2.1 was installed from http://stdcxx.apache.org/download.html using: gmake BUILDTYPE=8d CONFIG=gcc.config
Base Compiler Invocation
C benchmarks:
| xlc -qlanglvl=extc99 |
C++ benchmarks:
| xlC |
Fortran benchmarks:
| xlf95 |
Benchmarks using both Fortran and C:
| xlc -qlanglvl=extc99 xlf95 |
Base Portability Flags
| 410.bwaves: | -qfixed |
| 416.gamess: | -qfixed |
| 434.zeusmp: | -qfixed |
| 435.gromacs: | -qfixed -qextname |
| 436.cactusADM: | -qfixed -qextname |
| 437.leslie3d: | -qfixed |
| 454.calculix: | -qfixed -qextname |
| 481.wrf: | -DNOUNDERSCORE |
| 482.sphinx3: | -qchars=signed |
Base Optimization Flags
C benchmarks:
| -O5 -qarch=pwr7 -qtune=pwr7 -lhugetlbfs |
C++ benchmarks:
| -O5 -qarch=pwr7 -qtune=pwr7 -qrtti -lhugetlbfs |
Fortran benchmarks:
| -O5 -qarch=pwr7 -qtune=pwr7 -qsmallstack=dynlenonheap -qalias=nostd -B/usr/share/libhugetlbfs/ -tl -Wl,--hugetlbfs-align |
Benchmarks using both Fortran and C:
| -O5 -qarch=pwr7 -qtune=pwr7 -qsmallstack=dynlenonheap -qalias=nostd -B/usr/share/libhugetlbfs/ -tl -Wl,--hugetlbfs-align |
Base Other Flags
C benchmarks:
| -qipa=noobject -qipa=threads |
C++ benchmarks:
| -qipa=noobject -qipa=threads |
Fortran benchmarks:
| -qipa=noobject -qipa=threads |
Benchmarks using both Fortran and C:
| -qipa=noobject -qipa=threads |
Peak Compiler Invocation
C benchmarks:
| xlc -qlanglvl=extc99 |
C++ benchmarks:
| xlC |
Fortran benchmarks:
| xlf95 |
Benchmarks using both Fortran and C:
| xlc -qlanglvl=extc99 xlf95 |
Peak Portability Flags
| 410.bwaves: | -qfixed |
| 416.gamess: | -qfixed |
| 434.zeusmp: | -qfixed |
| 435.gromacs: | -qfixed -qextname |
| 436.cactusADM: | -DSPEC_CPU_LP64 -qfixed -qextname |
| 437.leslie3d: | -qfixed |
| 453.povray: | -DSPEC_CPU_LP64 |
| 454.calculix: | -qfixed -qextname |
| 481.wrf: | -DNOUNDERSCORE |
| 482.sphinx3: | -qchars=signed |
Peak Optimization Flags
C benchmarks:
| 433.milc: | basepeak = yes |
| 470.lbm: | basepeak = yes |
| 482.sphinx3: | -Wl,-q -qpdf1(pass 1) -qpdf2(pass 2) -O4 -qarch=pwr7 -qtune=pwr7 -lhugetlbfs |
C++ benchmarks:
| 444.namd: | -Wl,-q -qpdf1(pass 1) -qpdf2(pass 2) -O5 -qarch=pwr7 -qtune=pwr7 -lhugetlbfs |
| 447.dealII: | -O4 -qarch=pwr7 -qtune=pwr7 -qrtti -qcpp_stdinc=/root/stdcxx421/include/ansi:/root/stdcxx421/include -lsmartheap -lhugetlbfs -L/root/stdcxx421/lib -R/root/stdcxx421/lib -lstd8d |
| 450.soplex: | -Wl,-q -qpdf1(pass 1) -qpdf2(pass 2) -O3 -qtune=auto -qarch=pwr5 -lhugetlbfs |
| 453.povray: | -Wl,-q -qpdf1(pass 1) -qpdf2(pass 2) -O4 -qarch=pwr7 -qtune=pwr7 -qsimd -q64 -lsmartheap64 |
Fortran benchmarks:
| 410.bwaves: | basepeak = yes |
| 416.gamess: | -qpdf1(pass 1) -qpdf2(pass 2) -O5 -qarch=pwr7 -qtune=pwr7 -qalias=nostd -lhugetlbfs |
| 434.zeusmp: | basepeak = yes |
| 437.leslie3d: | -Wl,-q -O5 -qarch=pwr7 -qtune=pwr7 -q64 -B/usr/share/libhugetlbfs/ -tl -Wl,--hugetlbfs-align |
| 459.GemsFDTD: | -O4 -qarch=pwr7 -qtune=pwr7 -qsimd -B/usr/share/libhugetlbfs/ -tl -Wl,--hugetlbfs-align |
| 465.tonto: | -Wl,-q -qpdf1(pass 1) -qpdf2(pass 2) -O5 -qarch=pwr7 -qtune=pwr7 -qsimd -lhugetlbfs |
Benchmarks using both Fortran and C:
| 435.gromacs: | -Wl,-q -qpdf1(pass 1) -qpdf2(pass 2) -O4 -qarch=pwr7 -qtune=pwr7 -qsimd -lhugetlbfs |
| 436.cactusADM: | -Wl,-q -O4 -qarch=pwr7 -qtune=pwr7 -qsimd -qnostrict -q64 -lhugetlbfs |
| 454.calculix: | -qpdf1(pass 1) -qpdf2(pass 2) -O5 -qarch=pwr7 -qtune=pwr7 -B/usr/share/libhugetlbfs/ -tl -Wl,--hugetlbfs-align |
| 481.wrf: | -O3 -qarch=pwr7 -qtune=pwr7 -q64 -lhugetlbfs |