|
|
CFP2000 Result Copyright © 1999-2004 Standard Performance Evaluation Corporation |
|
Advanced Micro Devices Pogo Linux PerformanceWare 3566, AMD Opteron (TM) 870 |
|
| SPEC license # 49 | Tested by: AMD, Austin, TX | Test date: Apr-2005 | Hardware Avail: May-2005 | Software Avail: May-2005 |
![]() |
Benchmark | Base Copies |
Base Runtime |
Base Ratio |
Copies | Runtime | Ratio |
|---|---|---|---|---|---|---|---|
| 168.wupwise | 8 | 84.2 | 176 | 8 | 84.2 | 176 | |
| 171.swim | 8 | 293 | 98.2 | 8 | 293 | 98.2 | |
| 172.mgrid | 8 | 183 | 91.5 | 8 | 183 | 91.5 | |
| 173.applu | 8 | 227 | 86.0 | 8 | 227 | 86.0 | |
| 177.mesa | 8 | 96.0 | 135 | 8 | 90.5 | 144 | |
| 178.galgel | 8 | 160 | 169 | 8 | 146 | 184 | |
| 179.art | 8 | 316 | 76.5 | 8 | 296 | 81.5 | |
| 183.equake | 8 | 135 | 89.7 | 8 | 135 | 89.7 | |
| 187.facerec | 8 | 152 | 116 | 8 | 152 | 116 | |
| 188.ammp | 8 | 246 | 83.1 | 8 | 214 | 95.5 | |
| 189.lucas | 8 | 178 | 104 | 8 | 178 | 104 | |
| 191.fma3d | 8 | 214 | 91.1 | 8 | 201 | 97.1 | |
| 200.sixtrack | 8 | 190 | 53.6 | 8 | 190 | 53.6 | |
| 301.apsi | 8 | 237 | 102 | 8 | 237 | 102 | |
| SPECfp_rate_base2000 | 100 | ||||||
| SPECfp_rate2000 | 103 | ||||||
|
|
||||||||||||||||||||||||||||||||||||||||||
| Notes / Tuning Information |
|---|
+FDO: PASS1=-Qprof_gen PASS2=-Qprof_use
+ACML is linking with AMD Core Math Library V2.1
ONESTEP is set for all peak runs.
ifort is the Intel Fortran compiler, icl is the Intel C++ compiler and
pgf90 is the PGI Fortran compiler.
The Intel C++ 8.0 and the Intel Fortran 8.1 compilers are setup in the following order:
"c:\program files\intel\fortran\compiler80\ia32\bin\ifortvars.bat"
"c:\program files\intel\cpp\compiler80\ia32\bin\iclvars.bat"
To make sure that the correct libraries are selected, the following link option is
added for the peak runs where Intel Fortran 8.1 compiler is used:
LDOPT = -Fe$@ -link -LIBPATH:"c:\program files\intel\fortran\compiler80\ia32\lib"
(denoted by +LIBPATH:INTEL8.1 in the optimization flags listed below)
Portability:
178.galgel: -Mfixed
Baseline: C : icl -fast -arch:SSE2 -QaxW +FDO
Baseline: Fortran: pgf90 -fastsse -Mipa=fast,inline
Peak tuning:
168.wupwise: pgf90 basepeak=yes
171.swim: basepeak=yes
172.mgrid: pgf90 basepeak=yes
173.applu: basepeak=yes
177.mesa: icl -Qipo -arch:SSE2 +FDO -Qunroll1 -Qansi_alias
-Qoption,f,-ip_ninl_max_stats=1500,-ip_ninl_max_total_stats=4500
179.art: icl -Qipo -Zp4 +FDO
183.equake: icl basepeak=yes
178.galgel: pgf90 -fastsse -Mipa=fast,safe RM_SOURCES=lapak.f90 -Munix +ACML
187.facerec: basepeak=yes
188.ammp: icl -Oa -arch:SSE2 -Zp4 -Qansi_alias
189.lucas: basepeak=yes
191.fma3d: ifort -Qipo -QaxN -QxW +FDO -Qansi-alias- +LIBPATH:INTEL8.1
200.sixtrack: pgf90 basepeak=yes
301.apsi: pgf90 basepeak=yes
Bios Rev 2.33.1.1
All memory slots populated on all CPU(s).
Physical Address Extension (PAE) enabled.
The start /b /wait /affinity command is used to bind CPU(s) to processes.
|
First published at SPEC.org on 03-May-2005
Generated on Tue May 3 18:30:23 2005 by SPEC CPU2000 HTML formatter v1.01