| CPU2006 license: | 20 | Test date: | Mar-2011 |
|---|---|---|---|
| Test sponsor: | Bull SAS | Hardware Availability: | Apr-2011 |
| Tested by: | Dell Inc. | Software Availability: | Apr-2011 |
| Hardware | |
|---|---|
| CPU Name: | Intel Xeon E7-4820 |
| CPU Characteristics: | Intel Turbo Boost Technology up to 2.27 GHz |
| CPU MHz: | 2000 |
| FPU: | Integrated |
| CPU(s) enabled: | 32 cores, 4 chips, 8 cores/chip |
| CPU(s) orderable: | 2,4 chips |
| Primary Cache: | 32 KB I + 32 KB D on chip per core |
| Secondary Cache: | 256 KB I+D on chip per core |
| L3 Cache: | 18 MB I+D on chip per chip |
| Other Cache: | None |
| Memory: | 512 GB (64 x 8 GB 4Rx8 PC3-8500R-7, ECC, running at 978 MHz) |
| Disk Subsystem: | 1 x 500 GB 7200 RPM SAS 6Gb |
| Other Hardware: | None |
| Software | |
|---|---|
| Operating System: | SUSE Linux Enterprise Server 11 SP1 (x86_64), Kernel 2.6.32.12-0.7-default |
| Compiler: | Intel C++ and Fortran Intel 64 Compiler XE for applications running on Intel 64 Version 12.0 Update 3 |
| Auto Parallel: | Yes |
| File System: | ext3 |
| System State: | Run level 3 (multi-user) |
| Base Pointers: | 64-bit |
| Peak Pointers: | 32/64-bit |
| Other Software: | None |
Results Table
| Benchmark | Base | Peak | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Seconds | Ratio | Seconds | Ratio | Seconds | Ratio | 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 | 37.9 | 358 | 40.3 | 337 | 45.3 | 300 | 43.9 | 309 | 44.3 | 307 | 38.1 | 357 |
| 416.gamess | 1333 | 14.7 | 1349 | 14.5 | 1334 | 14.7 | 1042 | 18.8 | 1041 | 18.8 | 1042 | 18.8 |
| 433.milc | 447 | 20.5 | 455 | 20.2 | 454 | 20.2 | 447 | 20.5 | 455 | 20.2 | 454 | 20.2 |
| 434.zeusmp | 106 | 85.8 | 106 | 85.5 | 107 | 85.3 | 106 | 85.8 | 106 | 85.5 | 107 | 85.3 |
| 435.gromacs | 479 | 14.9 | 479 | 14.9 | 481 | 14.9 | 446 | 16.0 | 449 | 15.9 | 449 | 15.9 |
| 436.cactusADM | 32.8 | 364 | 34.0 | 351 | 50.0 | 239 | 32.8 | 364 | 34.0 | 351 | 50.0 | 239 |
| 437.leslie3d | 58.8 | 160 | 64.2 | 146 | 60.0 | 157 | 58.8 | 160 | 64.2 | 146 | 60.0 | 157 |
| 444.namd | 608 | 13.2 | 608 | 13.2 | 608 | 13.2 | 597 | 13.4 | 597 | 13.4 | 597 | 13.4 |
| 447.dealII | 457 | 25.0 | 456 | 25.1 | 456 | 25.1 | 457 | 25.0 | 456 | 25.1 | 456 | 25.1 |
| 450.soplex | 510 | 16.4 | 508 | 16.4 | 503 | 16.6 | 510 | 16.4 | 508 | 16.4 | 503 | 16.6 |
| 453.povray | 253 | 21.0 | 253 | 21.0 | 253 | 21.0 | 202 | 26.4 | 201 | 26.4 | 203 | 26.3 |
| 454.calculix | 437 | 18.9 | 438 | 18.9 | 440 | 18.7 | 373 | 22.1 | 374 | 22.1 | 374 | 22.1 |
| 459.GemsFDTD | 137 | 77.2 | 120 | 88.4 | 127 | 83.4 | 110 | 96.4 | 105 | 101 | 108 | 98.2 |
| 465.tonto | 659 | 14.9 | 661 | 14.9 | 665 | 14.8 | 436 | 22.6 | 434 | 22.7 | 435 | 22.6 |
| 470.lbm | 29.2 | 470 | 28.4 | 483 | 37.8 | 363 | 53.4 | 257 | 57.2 | 240 | 53.6 | 256 |
| 481.wrf | 426 | 26.2 | 430 | 26.0 | 426 | 26.2 | 426 | 26.2 | 430 | 26.0 | 426 | 26.2 |
| 482.sphinx3 | 552 | 35.3 | 549 | 35.5 | 550 | 35.4 | 493 | 39.6 | 493 | 39.5 | 492 | 39.6 |
Operating System Notes
'ulimit -s unlimited' was used to set the stacksize to unlimited prior to run 'mount -t hugetlbfs nodev /mnt/hugepages' was used to enable large pages echo 900 > /proc/sys/vm/nr_hugepages export HUGETLB_MORECORE=yes export LD_PRELOAD=/usr/lib64/libhugetlbfs.so
Platform Notes
BIOS Settings: Power Management = Maximum Performance (Default = Active Power Controller) Memory Interleaving = Enabled (Default = Disabled)
General Notes
The Dell PowerEdge R910 and the Bull NovaScale R480 F2 models are electronically equivalent. The results have been measured on a Dell PowerEdge R910 model. OMP_NUM_THREADS set to number of cores Binaries were compiled on RHEL5.5
Base Compiler Invocation
C benchmarks:
| icc -m64 |
C++ benchmarks:
| icpc -m64 |
Fortran benchmarks:
| ifort -m64 |
Benchmarks using both Fortran and C:
| icc -m64 ifort -m64 |
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 -nofor_main |
| 436.cactusADM: | -DSPEC_CPU_LP64 -nofor_main |
| 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 -nofor_main |
| 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:
| -xSSE4.2 -ipo -O3 -no-prec-div -static -parallel -opt-prefetch -ansi-alias |
C++ benchmarks:
| -xSSE4.2 -ipo -O3 -no-prec-div -static -opt-prefetch -ansi-alias |
Fortran benchmarks:
| -xSSE4.2 -ipo -O3 -no-prec-div -static -parallel -opt-prefetch |
Benchmarks using both Fortran and C:
| -xSSE4.2 -ipo -O3 -no-prec-div -static -parallel -opt-prefetch -ansi-alias |
Peak Compiler Invocation
C benchmarks:
| icc -m64 |
C++ benchmarks:
| icpc -m64 |
Fortran benchmarks:
| ifort -m64 |
Benchmarks using both Fortran and C:
| icc -m64 ifort -m64 |
Peak Portability Flags
Same as Base Portability Flags
Peak Optimization Flags
C benchmarks:
| 433.milc: | basepeak = yes |
| 470.lbm: | -xSSE4.2(pass 2) -prof-gen(pass 1) -ipo(pass 2) -O3(pass 2) -no-prec-div(pass 2) -prof-use(pass 2) -parallel -ansi-alias -static -auto-ilp32 |
| 482.sphinx3: | -xSSE4.2 -ipo -O3 -no-prec-div -unroll2 -ansi-alias -parallel |
C++ benchmarks:
| 444.namd: | -xSSE4.2(pass 2) -prof-gen(pass 1) -ipo(pass 2) -O3(pass 2) -no-prec-div(pass 2) -prof-use(pass 2) -fno-alias -auto-ilp32 |
| 447.dealII: | basepeak = yes |
| 450.soplex: | basepeak = yes |
| 453.povray: | -xSSE4.2(pass 2) -prof-gen(pass 1) -ipo(pass 2) -O3(pass 2) -no-prec-div(pass 2) -prof-use(pass 2) -unroll4 -ansi-alias -B /usr/share/libhugetlbfs/ -Wl,-melf_x86_64 -Wl,-hugetlbfs-link=BDT |
Fortran benchmarks:
| 410.bwaves: | -xSSE4.2 -ipo -O3 -no-prec-div -opt-prefetch -parallel -static |
| 416.gamess: | -xSSE4.2(pass 2) -prof-gen(pass 1) -ipo(pass 2) -O3(pass 2) -no-prec-div(pass 2) -prof-use(pass 2) -unroll2 -inline-level=0 -scalar-rep- -static |
| 434.zeusmp: | basepeak = yes |
| 437.leslie3d: | basepeak = yes |
| 459.GemsFDTD: | -xSSE4.2(pass 2) -prof-gen(pass 1) -ipo(pass 2) -O3(pass 2) -no-prec-div(pass 2) -prof-use(pass 2) -unroll2 -inline-level=0 -opt-prefetch -parallel -B /usr/share/libhugetlbfs/ -Wl,-melf_x86_64 -Wl,-hugetlbfs-link=BDT |
| 465.tonto: | -xSSE4.2(pass 2) -prof-gen(pass 1) -ipo(pass 2) -O3(pass 2) -no-prec-div(pass 2) -prof-use(pass 2) -inline-calloc -opt-malloc-options=3 -auto -unroll4 -B /usr/share/libhugetlbfs/ -Wl,-melf_x86_64 -Wl,-hugetlbfs-link=BDT |
Benchmarks using both Fortran and C:
| 435.gromacs: | -xSSE4.2(pass 2) -prof-gen(pass 1) -ipo(pass 2) -O3(pass 2) -no-prec-div(pass 2) -prof-use(pass 2) -static -auto-ilp32 -ansi-alias |
| 436.cactusADM: | basepeak = yes |
| 454.calculix: | -xSSE4.2 -ipo -O3 -no-prec-div -auto-ilp32 -ansi-alias |
| 481.wrf: | basepeak = yes |