Last night I upgraded all my Folding@home clients to the new version.
Since the new version has timestamps, it's easy to calculate exactly
how fast our boxes are churning out frames.
I wrote a couple of nifty scripts to read FAHlog.txt and print min,
max, mean, and standard deviation of frame calculation time.
Here are the results for my boxen. (The scripts will be attached
in my next email.)
# 3/6/2002, 11:00 PM PST
HOSTNAME WORK FRAME CPU CPU CACHE FSB
UNITS MEAN MHz KB MHz
tivopc-cpu0 321 6:07 Athlon 1400 384 266
tivopc-cpu1 322 6:07 Athlon 1400 384 266
alviso 95 16:28 PIII 600 512 133
jogger-egg 83 20:43 PII 450 512 100
chezgeek 70 21:21 Celeron 466 128 66
Anne's PC 60 PII 450 512 100
fold 58 19:22 Celeron 466 128 66
central-services 52 19:36 Celeron 466 128 66
VAIO 50 22:42 MobilP2 400 256 100
Observations:
Celeron, PII and PIII are pretty similar. Athlon is 10%
faster, corrected for clock speed. Cache size and memory bus
speed doesn't seem to matter much.
Anne's PC and jogger-egg are identical hardware, but one runs
Linux 2.2 and one runs Windows 98. Linux has done 38% more
work units.
I sure have a lot of CPUs between 400 and 466 MHz. (-:
--
Bob Miller K<bob>
kbobsoft software consulting
http://kbobsoft.com [EMAIL PROTECTED]
