[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ccp4bb]: Benchmarks

***  For details on how to be removed from this list visit the  ***
***    CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html    ***

Yesterday I posted an informal benchmark I had done on some of the
computers in our department.  The primary interest was to show the
performance of a real crystallography program (SOLVE) on a Linux based
PC computer compared to SGI and DEC Alpha workstations.  A number of you
have written asking for clarification or details.  Allow me to expand.

The test program is Tom Terwilliger's SOLVE program.   I downloaded all
the binaries from the SOLVE home page and so do not know what
optimizations Tom used.  The test data is a 3 wavelength MAD experiment
to 2.8 A that solves for 10 sites.

I ran this programs on:
SGI Indigo II, MIPS R10000 195 MHz, 256 MB RAM, Irix 6.2    102 min
SGI Octane, MIPS R10000 250 MHz, 256 MB RAM, Irix 6.4    83 min
DEC Alpha 600AU, Alpha 21164 600 MHz, 512 MB RAM, Dec Unix 4.0E    56
PC, Athlon 800 MHz, 256 MB RAM, RedHat 6.2    106 min
(This was built for me by www.affordablecomputers.com in Ann Arbor, MI
who did a very nice job for <$2k)

All trials were run in the foreground (nice = 0) with no competing
activity.  But, as we say in America, Your Mileage my vary.

Nic Steussy

P.S. There is an extensive evaluation of computers for scientific uses