SPEC Benchmark CFP95rate Summary Base Base Base Peak Peak Peak Benchmarks # Copies Run Time Rate # Copies Run Time Rate ------------ -------- -------- -------- -------- -------- -------- 101.tomcatv 60 42.8 46632 60 39.5 50602 102.swim 60 77.0 60329 60 70.8 65558 103.su2cor 60 60.7 12451 60 37.2 20302 104.hydro2d 60 54.8 23637 60 50.5 25662 107.mgrid 60 44.8 30165 60 40.0 33789 110.applu 60 74.5 15939 60 72.3 16442 125.turb3d 60 109 20239 60 71.6 30939 141.apsi 60 40.9 27696 60 39.2 28894 145.fpppp 60 105 49400 60 98.6 52594 146.wave5 60 47.1 34366 60 44.1 36776 SPECfp_rate_base95 (G. M.) 28697 SPECfp_rate95 (Geom. Mean) 33267 TESTER INFORMATION ------------------ SPEC License #: 2 Tested By: Compaq NH Test Date: Mar-2000 Hardware Avail: Mar-2000 Software Avail: Apr-2000 HARDWARE -------- Model Name: Compaq AlphaServer SC CPU: 667 MHz 21264A FPU: Integrated Number of CPU(s): 64, 16 nodes x 4 each Primary Cache: 64KB(I)+64KB(D) on chip Secondary Cache: 8MB off chip, per CPU Other Cache: none Memory: 4GB per node Disk Subsystem: RAID-0 6x18GB (10,000 RPM), striped across 6x HSG80, attached to node 0 Other Hardware: Ethernet SOFTWARE -------- Operating System: Tru64 UNIX 5.0 AlphaServer SC System Software T1.0-49 Compiler: Fortran V5.3 KAP Fortran V4.2 File System: AdvFS System State: Multi User NOTES ------- base: kf77 -arch ev6 -O5 -transform_loops -fkapargs='-ur=1' tomcatv: kf90 -arch ev6 -O5 -inline none . -fkapargs='-fuselevel=2 -ur=9 -ur2=460 -ur3=30 ' swim: kf77 -arch ev6 -O5 -fkapargs='-ag=a -fuselevel=2 -ur=1' su2cor: kf90 -arch ev6 -O3 -transform_loops -automatic . -fkapargs=' -ag=abfg -r=0 -nointerleave -ur2=260 ' hydro2d: kf90 -arch ev6 -O5 -fkapargs='-fuselevel=2 -ur2=860' mgrid: kf77 -arch ev6 -O5 -speculate all -fkapargs='-fuselevel=2 -ur=1' applu: kf90 -arch ev6 -transform_loops -automatic -fkapargs=' -r=2 -o=0 -ur2=60 ' turb3d: kf90 -arch ev6 -O5 -speculate all -automatic -fkapargs=' -ag=a -o=1 -ur2=760 -ur3=100 ' apsi: kf90 -arch ev6 -O2 -pipeline -automatic . -fkapargs=' -fuselevel=2 -ur=8 -ur2=60 ' fpppp: kf90 -arch ev6 -O2 -transform_loops -fkapargs='-so=0 -o=0' wave5: kf90 -arch ev6 -O5 -transform_loops -unroll 15 . -fkapargs='-r=1 -o=2 -ur2=260 -ur3=30' The SC is 16 nodes, 4 CPUs per node. Nodes 1-15 were used to run benchmarks; node 0 had runspec itself. There is also a front end, where the "Partition Manager" runs; it has 2x 500MHz Alpha 21264 CPUs. submitnnn = echo "%command%" > dobmk; prun -B -c 1 -n 1 -N 1 sh dobmk The SPECfp_rate95 executables were only run on 60 of the 64 CPUs.