bench

Biochemistry Code Performance Results

Date: Sat, 18 Apr 1998 10:44:03 -0700 (PDT)

From: Hartmut Luecke <hudel@anx12.bio.uci.edu>

To: "Herbert W. Hamber" <hamber@cern.ch>

Hi Herbert!

Below are a few benchmarks, but I think multiprocessor performance will be far better since there were other jobs running on node*.ps.

Can't wait to run on 64 procs!

Cheers, Hudel

Some benchmarks (note: as I've seen before for another program, the 300 MHz PII is faster than the R10000!):

single SGI 175 MHz R10000 (no load); Info> TOTAL RUN TIME 786.408; Info> TOTAL CPU TIME 771.62
: single SGI 195 MHz R10000 (no load); Info> TOTAL RUN TIME 636.934; Info> TOTAL CPU TIME 633.75
: single 300 MHz PII (no load); Info> TOTAL RUN TIME 521.92; Info> TOTAL CPU TIME 502.29
: 4 proc 300 MHz PII; Node 0:Info> TOTAL RUN TIME 750.293; Node 0:Info> TOTAL CPU TIME 435.62
: 8 proc 300 MHz PII; Node 0:Info> TOTAL RUN TIME 526.23; Node 0:Info> TOTAL CPU TIME 303.43
: 14 proc 300 MHz PII; Node 0:Info> TOTAL RUN TIME 202.839; Node 0:Info> TOTAL CPU TIME 99.45