| CPU2006 license: | 20 | Test date: | Dec-2008 |
|---|---|---|---|
| Test sponsor: | Bull SAS | Hardware Availability: | Oct-2008 |
| Tested by: | NEC Corporation | Software Availability: | Nov-2008 |
| Hardware | |
|---|---|
| CPU Name: | Intel Xeon E5450 |
| CPU Characteristics: | 1333 MHz system bus |
| CPU MHz: | 3000 |
| FPU: | Integrated |
| CPU(s) enabled: | 8 cores, 2 chips, 4 cores/chip |
| CPU(s) orderable: | 1,2 chips (fault tolerant, see Platform Notes) |
| Primary Cache: | 32 KB I + 32 KB D on chip per core |
| Secondary Cache: | 12 MB I+D on chip per chip, 6 MB shared / 2 cores |
| L3 Cache: | None |
| Other Cache: | None |
| Memory: | 24 GB (6x4 GB PC2-5300F, 2 rank, CL5-5-5, ECC) |
| Disk Subsystem: | 2x146.5 GB SAS, 15000 RPM, Software RAID Level1 |
| Other Hardware: | None |
| Software | |
|---|---|
| Operating System: | Red Hat Enterprise Linux Server release 5.2 Advanced Platform, Kernel 2.6.18-92.1.13.el5 on an x86_64 |
| Compiler: | Intel C++ Compiler 11.0 for Linux Build 20081105 Package ID: l_cproc_p_11.0.074 |
| Auto Parallel: | Yes |
| File System: | ext3 |
| System State: | Run level 3 (multi-user) |
| Base Pointers: | 32-bit |
| Peak Pointers: | 32/64-bit |
| Other Software: | MicroQuill SmartHeap Library 8.1 ft Server Control Software 6.0.2-198 |
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. | ||||||||||||
| 400.perlbench | 484 | 20.2 | 480 | 20.3 | 481 | 20.3 | 380 | 25.7 | 377 | 25.9 | 380 | 25.7 |
| 401.bzip2 | 589 | 16.4 | 587 | 16.4 | 587 | 16.4 | 543 | 17.8 | 542 | 17.8 | 543 | 17.8 |
| 403.gcc | 547 | 14.7 | 549 | 14.7 | 546 | 14.7 | 431 | 18.7 | 432 | 18.6 | 431 | 18.7 |
| 429.mcf | 369 | 24.7 | 369 | 24.7 | 371 | 24.6 | 369 | 24.7 | 363 | 25.2 | 362 | 25.2 |
| 445.gobmk | 532 | 19.7 | 533 | 19.7 | 533 | 19.7 | 486 | 21.6 | 485 | 21.6 | 486 | 21.6 |
| 456.hmmer | 552 | 16.9 | 552 | 16.9 | 552 | 16.9 | 360 | 25.9 | 360 | 25.9 | 360 | 25.9 |
| 458.sjeng | 612 | 19.8 | 612 | 19.8 | 612 | 19.8 | 575 | 21.0 | 576 | 21.0 | 577 | 21.0 |
| 462.libquantum | 79.9 | 259 | 79.2 | 262 | 80.5 | 258 | 79.9 | 259 | 79.2 | 262 | 80.5 | 258 |
| 464.h264ref | 689 | 32.1 | 690 | 32.1 | 693 | 31.9 | 634 | 34.9 | 632 | 35.0 | 633 | 35.0 |
| 471.omnetpp | 452 | 13.8 | 452 | 13.8 | 452 | 13.8 | 410 | 15.2 | 409 | 15.3 | 409 | 15.3 |
| 473.astar | 496 | 14.2 | 486 | 14.5 | 485 | 14.5 | 414 | 17.0 | 416 | 16.9 | 413 | 17.0 |
| 483.xalancbmk | 273 | 25.3 | 274 | 25.2 | 274 | 25.2 | 273 | 25.3 | 274 | 25.2 | 274 | 25.2 |
Operating System Notes
'ulimit -s unlimited' was used to set the stacksize to unlimited prior to run OMP_NUM_THREADS set to number of cores KMP_AFFINITY set to "physical,0"
Platform Notes
This Express5800/320Fd-MR is a fault-tolerant server. Two modules are installed in this server. Each module physically has "2CPU chips,24GB memory", The total physical configuration is "4CPU chips,48GB memory". Using fault-tolerant lockstep technology, these two modules communicate with each other and execute the same instructions at the same time, The operating system only sees "2CPU chips,24GB memory" as the other components add only redundancy and do not contribute to any performance benefit.
General Notes
The NEC Express5800/320Fd-MR(Intel Xeon E5450) and the Bull NovaScale R630 E1 MR(Intel Xeon E5450, 3.00 GHz) models are electronically equivalent. The results have been measured on a NEC Express5800/320Fd-MR(Intel Xeon E5450) model.
Base Portability Flags
| 400.perlbench: | -DSPEC_CPU_LINUX_IA32 |
| 462.libquantum: | -DSPEC_CPU_LINUX |
| 483.xalancbmk: | -DSPEC_CPU_LINUX |
Base Optimization Flags
C benchmarks:
| -xSSE4.1 -ipo -O3 -no-prec-div -static -parallel -par-runtime-control -opt-prefetch |
C++ benchmarks:
| -xSSE4.1 -ipo -O3 -no-prec-div -opt-prefetch -Wl,-z,muldefs -L/opt/SmartHeap_8.1/lib -lsmartheap |
Peak Compiler Invocation
C benchmarks (except as noted below):
C++ benchmarks:
| icpc |
Peak Portability Flags
| 400.perlbench: | -DSPEC_CPU_LINUX_IA32 |
| 401.bzip2: | -DSPEC_CPU_LP64 |
| 456.hmmer: | -DSPEC_CPU_LP64 |
| 462.libquantum: | -DSPEC_CPU_LINUX |
| 483.xalancbmk: | -DSPEC_CPU_LINUX |
Peak Optimization Flags
C benchmarks:
| 400.perlbench: | -prof-gen(pass 1) -prof-use(pass 2) -xSSE4.1 -ipo -O3 -no-prec-div -static -ansi-alias -opt-prefetch |
| 401.bzip2: | -prof-gen(pass 1) -prof-use(pass 2) -xSSE4.1 -ipo -O3 -no-prec-div -static -auto-ilp32 -opt-prefetch -ansi-alias |
| 403.gcc: | -xSSE4.1 -ipo -O3 -no-prec-div -static -inline-calloc -opt-malloc-options=3 |
| 429.mcf: | -xSSE4.1 -ipo -O3 -no-prec-div -static -opt-prefetch |
| 445.gobmk: | -prof-gen(pass 1) -prof-use(pass 2) -xSSE4.1 -O2 -ipo -no-prec-div -ansi-alias |
| 456.hmmer: | -xSSE4.1 -ipo -O3 -no-prec-div -static -unroll2 -ansi-alias -auto-ilp32 |
| 458.sjeng: | -prof-gen(pass 1) -prof-use(pass 2) -xSSE4.1 -ipo -O3 -no-prec-div -static -unroll4 |
| 462.libquantum: | basepeak = yes |
| 464.h264ref: | -prof-gen(pass 1) -prof-use(pass 2) -xSSE4.1 -ipo -O3 -no-prec-div -static -unroll2 -ansi-alias |
C++ benchmarks:
| 471.omnetpp: | -prof-gen(pass 1) -prof-use(pass 2) -xSSE4.1 -ipo -O3 -no-prec-div -ansi-alias -opt-ra-region-strategy=block -Wl,-z,muldefs -L/opt/SmartHeap_8.1/lib -lsmartheap |
| 473.astar: | -prof-gen(pass 1) -prof-use(pass 2) -xSSE4.1 -ipo -O3 -no-prec-div -ansi-alias -opt-ra-region-strategy=routine -Wl,-z,muldefs -L/opt/SmartHeap_8.1/lib -lsmartheap |
| 483.xalancbmk: | basepeak = yes |