SPEC® CFP2006 Result

Copyright 2006-2014 Standard Performance Evaluation Corporation

Hewlett-Packard Company

ProLiant DL785 G6
(2.8 GHz AMD Opteron 8439 SE)

CPU2006 license: 3 Test date: Aug-2009
Test sponsor: Hewlett-Packard Company Hardware Availability: Oct-2009
Tested by: Hewlett-Packard Company Software Availability: Apr-2009
Benchmark results graph
Hardware
CPU Name: AMD Opteron 8439 SE
CPU Characteristics:
CPU MHz: 2800
FPU: Integrated
CPU(s) enabled: 48 cores, 8 chips, 6 cores/chip
CPU(s) orderable: 4,8 chips
Primary Cache: 64 KB I + 64 KB D on chip per core
Secondary Cache: 512 KB I+D on chip per core
L3 Cache: 6 MB I+D on chip per chip
Other Cache: None
Memory: 128 GB (32x4 GB, PC2-6400 CL5)
Disk Subsystem: 4x72 GB 10K SAS
Other Hardware: None
Software
Operating System: Red Hat Enterprise Linux Server release 5.3,
Kernel 2.6.18-128.el5
Compiler: PGI Server Complete Version 8.0
x86 Open64 4.2.2 Compiler Suite
Auto Parallel: No
File System: ext3
System State: Run level 5 (multi-user)
Base Pointers: 64-bit
Peak Pointers: 32/64-bit
Other Software: binutils 2.18

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 48 1755 372 1754 372 1755 372 48 1664 392 1661 393 1665 392
416.gamess 48 1117 841 1120 839 1118 841 48 1054 892 1059 888 1072 877
433.milc 48 1595 276 1595 276 1595 276 48 1595 276 1595 276 1595 276
434.zeusmp 48 794 550 793 551 794 550 48 787 555 786 555 787 555
435.gromacs 48 498 688 500 685 485 707 48 397 864 402 853 400 857
436.cactusADM 48 1034 555 1041 551 1041 551 48 1034 555 1041 551 1041 551
437.leslie3d 48 2045 221 2053 220 2044 221 48 1945 232 1936 233 1939 233
444.namd 48 580 663 579 665 580 664 48 536 718 529 728 526 732
447.dealII 48 622 882 623 881 620 885 48 457 1200 456 1200 456 1200
450.soplex 48 1436 279 1362 294 1354 296 48 1561 256 1261 317 1236 324
453.povray 48 306 835 308 830 308 829 48 263 970 264 969 300 852
454.calculix 48 442 895 446 888 444 892 48 393 1010 394 1010 395 1000
459.GemsFDTD 48 2466 207 2468 206 2467 206 48 2405 212 2405 212 2402 212
465.tonto 48 756 624 755 626 755 626 48 656 720 657 719 660 715
470.lbm 48 3433 192 3434 192 3432 192 48 3428 192 3428 192 3428 192
481.wrf 48 1314 408 1315 408 1315 408 48 1288 416 1285 417 1280 419
482.sphinx3 48 1638 571 1635 572 1624 576 48 1524 614 1520 615 1522 615

Submit Notes

The config file option 'submit' was used.
 'numactl' was  used to bind copies to the cores and local memory
 See the configuration file for details.

Operating System Notes

 'ulimit -s unlimited' was used to set environment stack size
 'ulimit -l 2097152'  was used to set environment locked pages in memory limit

 Set vm/nr_hugepages=21600 in /etc/sysctl.conf
 mount -t hugetlbfs nodev /mnt/hugepages

Platform Notes

  BIOS configuration:
  Power Regulator set to Static High Performance Mode

General Notes

Environment variables set by runspec before the start of the run:
HUGETLB_LIMIT = "450"
LD_LIBRARY_PATH = "/SPECcpu2006_Open64/amd0905is-libs/64:/SPECcpu2006_Open64/amd0905is-libs/32"
PGI_HUGE_PAGES = "450"

The x86 Open64 Compiler Suite is only available from (and supported by) AMD at
http://developer.amd.com/cpu/open64.

Base Compiler Invocation

C benchmarks:

 pgcc 

C++ benchmarks:

 pgcpp 

Fortran benchmarks:

 pgf95 

Benchmarks using both Fortran and C:

 pgcc   pgf95 

Base Portability Flags

410.bwaves:  -DSPEC_CPU_LP64 
416.gamess:  -DSPEC_CPU_LP64 
433.milc:  -DSPEC_CPU_LP64 
434.zeusmp:  -DSPEC_CPU_LP64 
435.gromacs:  -DSPEC_CPU_LP64   -Mnomain 
436.cactusADM:  -DSPEC_CPU_LP64   -Mnomain 
437.leslie3d:  -DSPEC_CPU_LP64 
444.namd:  -DSPEC_CPU_LP64 
447.dealII:  -DSPEC_CPU_LP64 
450.soplex:  -DSPEC_CPU_LP64 
453.povray:  -DSPEC_CPU_LP64 
454.calculix:  -DSPEC_CPU_LP64   -Mnomain 
459.GemsFDTD:  -DSPEC_CPU_LP64 
465.tonto:  -DSPEC_CPU_LP64 
470.lbm:  -DSPEC_CPU_LP64 
481.wrf:  -DSPEC_CPU_LP64   -DSPEC_CPU_CASE_FLAG   -DSPEC_CPU_LINUX 
482.sphinx3:  -DSPEC_CPU_LP64 

Base Optimization Flags

C benchmarks:

 -fastsse   -Msmartalloc=huge   -Mfprelaxed   -Mipa=fast   -Mipa=inline   -tp shanghai-64   -Bstatic_pgi 

C++ benchmarks:

 -fastsse   -Msmartalloc=huge   -Mfprelaxed   --zc_eh   -Mipa=fast   -Mipa=inline   -tp shanghai-64   -Bstatic_pgi 

Fortran benchmarks:

 -fastsse   -Msmartalloc=huge   -Mfprelaxed   -Mvect=short   -Mipa=fast   -Mipa=inline   -tp shanghai-64   -Bstatic_pgi 

Benchmarks using both Fortran and C:

 -fastsse   -Msmartalloc=huge   -Mfprelaxed   -Mipa=fast   -Mipa=inline   -tp shanghai-64   -Mvect=short   -Bstatic_pgi 

Base Other Flags

C benchmarks:

 -Mipa=jobs:4 

C++ benchmarks:

 -Mipa=jobs:4 

Fortran benchmarks:

 -Mipa=jobs:4 

Benchmarks using both Fortran and C:

 -Mipa=jobs:4 

Peak Compiler Invocation

C benchmarks:

 pgcc 

C++ benchmarks (except as noted below):

 openCC 
444.namd:  pgcpp 

Fortran benchmarks (except as noted below):

 openf95 
410.bwaves:  pgf95 
434.zeusmp:  pgf95 
437.leslie3d:  pgf95 

Benchmarks using both Fortran and C (except as noted below):

 pgcc   pgf95 
435.gromacs:  opencc   openf95 

Peak Portability Flags

410.bwaves:  -DSPEC_CPU_LP64 
416.gamess:  -DSPEC_CPU_LP64 
433.milc:  -DSPEC_CPU_LP64 
434.zeusmp:  -DSPEC_CPU_LP64 
435.gromacs:  -DSPEC_CPU_LP64 
436.cactusADM:  -DSPEC_CPU_LP64   -Mnomain 
437.leslie3d:  -DSPEC_CPU_LP64 
444.namd:  -DSPEC_CPU_LP64 
453.povray:  -DSPEC_CPU_LP64 
454.calculix:  -DSPEC_CPU_LP64   -Mnomain 
459.GemsFDTD:  -DSPEC_CPU_LP64 
465.tonto:  -DSPEC_CPU_LP64 
470.lbm:  -DSPEC_CPU_LP64 
481.wrf:  -DSPEC_CPU_LP64   -DSPEC_CPU_CASE_FLAG   -DSPEC_CPU_LINUX 
482.sphinx3:  -DSPEC_CPU_LP64 

Peak Optimization Flags

C benchmarks:

433.milc:  basepeak = yes 
470.lbm:  -fastsse   -Msmartalloc=huge   -Mprefetch=t0   -Mloop32   -Mfprelaxed   -Mipa=fast   -Mipa=inline   -tp shanghai-64   -Bstatic_pgi 
482.sphinx3:  -Mpfi=indirect(pass 1)   -Mpfo=indirect(pass 2)   -Mipa=fast(pass 2)   -Mipa=inline(pass 2)   -fastsse   -Mfprelaxed   -Msmartalloc   -tp shanghai-64   -Bstatic_pgi 

C++ benchmarks:

444.namd:  -Mpfi(pass 1)   -Mpfo(pass 2)   -Mipa=fast(pass 2)   -Mipa=inline(pass 2)   -fastsse   -Munroll=n:4   -Munroll=m:8   -Msmartalloc=huge   -Mnodepchk   -Mfprelaxed   --zc_eh   -tp shanghai-64   -Bstatic_pgi 
447.dealII:  -march=barcelona   -Ofast   -static   -INLINE:aggressive=on   -LNO:opt=0   -Wf,-fno-exceptions   -m32   -OPT:unroll_times_max=8   -OPT:unroll_size=256   -OPT:unroll_level=2   -HP:bdt=2m:heap=2m   -GRA:unspill=on   -CG:cmp_peep=on   -TENV:frame_pointer=off 
450.soplex:  -march=barcelona   -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -O3   -INLINE:aggressive=on   -OPT:IEEE_arith=3   -OPT:IEEE_NaN_Inf=off   -OPT:fold_unsigned_relops=on   -OPT:malloc_alg=1   -CG:load_exe=0   -fno-exceptions   -m32   -HP:bdt=2m 
453.povray:  -march=barcelona   -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -Ofast   -INLINE:aggressive=on   -HP:bdt=2m:heap=2m 

Fortran benchmarks:

410.bwaves:  -fastsse   -Msmartalloc   -Mprefetch=nta   -Mfprelaxed   -Mipa=fast   -Mipa=inline   -tp shanghai-64   -Bstatic_pgi 
416.gamess:  -march=barcelona   -fb_create fbdata(pass 1)   -fb_opt fbdata(pass 2)   -O2   -OPT:Ofast   -OPT:ro=3   -OPT:unroll_size=256   -HP:bdt=2m:heap=2m 
434.zeusmp:  -fastsse   -Mfprelaxed   -Mprefetch=distance:8   -Mprefetch=t0   -Msmartalloc=huge   -Msmartalloc=hugebss   -Mipa=fast   -Mipa=inline   -tp shanghai-64   -Bstatic_pgi 
437.leslie3d:  -Mpfi=indirect(pass 1)   -Mpfo=indirect(pass 2)   -Mipa=fast(pass 2)   -Mipa=inline(pass 2)   -fastsse   -Mvect=fuse   -Msmartalloc=huge   -Mprefetch=distance:8   -Mprefetch=t0   -Mfprelaxed   -tp shanghai-64   -Bstatic_pgi 
459.GemsFDTD:  -march=barcelona   -Ofast   -LNO:fission=2   -LNO:simd=2   -LNO:prefetch_ahead=1   -CG:load_exe=0   -HP 
465.tonto:  -march=barcelona   -Ofast   -OPT:alias=no_f90_pointer_alias   -LNO:blocking=off   -CG:load_exe=1   -IPA:plimit=525   -HP 

Benchmarks using both Fortran and C:

435.gromacs:  -march=barcelona   -Ofast   -OPT:rsqrt=2   -HP:bdt=2m:heap=2m 
436.cactusADM:  basepeak = yes 
454.calculix:  -Mpfi=indirect(pass 1)   -Mpfo=indirect(pass 2)   -Mipa=fast(pass 2)   -Mipa=inline(pass 2)   -fastsse   -Mvect=short   -Msmartalloc=huge   -Mprefetch=t0   -Mpre   -Mfprelaxed   -tp shanghai-64   -Bstatic_pgi 
481.wrf:  -fastsse   -Mvect=noaltcode   -Msmartalloc=huge   -Mprefetch=distance:8   -Mfprelaxed   -tp shanghai-64   -Bstatic_pgi 

Peak Other Flags

C benchmarks:

 -Mipa=jobs:4(pass 2) 

C++ benchmarks:

444.namd:  -Mipa=jobs:4(pass 2) 

Fortran benchmarks:

410.bwaves:  -Mipa=jobs:4 
434.zeusmp:  -Mipa=jobs:4 
437.leslie3d:  -Mipa=jobs:4(pass 2) 

Benchmarks using both Fortran and C:

436.cactusADM:  -Mipa=jobs:4 
454.calculix:  -Mipa=jobs:4(pass 2) 

The flags files that were used to format this result can be browsed at
http://www.spec.org/cpu2006/flags/pgi80_linux_flags.20090901.html,
http://www.spec.org/cpu2006/flags/hp-amd-linux-flags.html,
http://www.spec.org/cpu2006/flags/x86-open64-4.2.2-flags-revE.html.

You can also download the XML flags sources by saving the following links:
http://www.spec.org/cpu2006/flags/pgi80_linux_flags.20090901.xml,
http://www.spec.org/cpu2006/flags/hp-amd-linux-flags.xml,
http://www.spec.org/cpu2006/flags/x86-open64-4.2.2-flags-revE.xml.