Author: Robert Hyatt
Date: 08:28:40 05/16/04
Go up one level in this thread
On May 16, 2004 at 10:31:10, Frank Phillips wrote: >On May 16, 2004 at 05:36:00, Frank Phillips wrote: > >>On May 16, 2004 at 01:32:21, Robert Hyatt wrote: >> >>>On May 14, 2004 at 18:40:13, Frank E. Oldham wrote: >>> >>>>here's the speedup test results from dual G5 (2 GHz) >>>>crafty-19.10f >>>> >>>>best speedup = 3.32 >>>>worst speedup = 0.89 >>>> >>>>mt=2a mt=2b mt=2c mt=2d >>>> speedups >>>>==================== >>>>1.88 1.52 1.87 1.54 >>>>2.04 1.91 1.96 1.96 >>>>1.77 1.79 1.77 1.76 >>>>1.96 1.60 1.48 1.78 >>>>2.92 2.92 2.77 2.92 >>>>2.20 1.84 1.92 1.78 >>>>1.87 1.87 1.87 1.87 >>>>1.56 1.62 1.73 1.64 >>>>1.76 1.62 1.79 1.69 >>>>1.83 1.77 1.92 2.04 >>>>2.03 2.32 1.90 2.17 >>>>1.67 1.52 1.92 1.85 >>>>1.12 1.08 1.45 1.99 >>>>3.26 3.39 3.27 2.29 >>>>1.29 1.20 1.30 1.31 >>>>1.33 1.57 1.42 1.45 >>>>1.51 1.55 1.99 2.08 >>>>1.71 1.91 1.36 2.10 >>>>1.54 2.67 2.21 3.32 >>>>2.28 2.28 2.69 2.86 >>>>1.45 2.17 1.55 2.00 >>>>1.40 2.01 0.89 1.56 >>>>1.53 1.69 1.65 1.52 >>>>2.28 1.43 1.43 2.41 >>>>==================== >>>>1.84 1.89 1.84 2.00 mean speed-up >>>>0.50 0.53 0.51 0.49 stdev of mean speed-up >>>> >>>>results seem to match expected values; perhaps slightly higher than the Xeon >>>>tests >>>>because the memory interface is faster >>>> >>>>Frank >>> >>>Please don't post faked data here. We know it is faked because Vincent has said >>>that _everybody_ that tests crafty on duals and quads can't get anywhere my >>>predicted 1.7x and 3.1x speedups... >> >>I played some games against Brettermeier on FICS >> >>1: Crafty v19.13 (8 cpus) >> 2: Sun Fire V880, 8x Ultrasparc III 900 MHz, 16GB RAM >> >>Might be worth contacting him, if you want data on an 8 cpu machine (?) >> >>I was not paying much attention during the games, but he seemed to be getting >>around 3.5 M nps at times. Not shabby at all. >> >>Frank > >A random selection of statistics: > >cpu=4748% 1.88M >cpu=4834% 1.93M >cpu=5094% 3.24M >brettermeier(C)(2466)[73] whispers: ply=12; eval=Mat08; nps=2.86M; time=18.56; >\ cpu=4530%; egtb=0 > >Not sure how to interpret the 50 hundred % > >Frank Me either. That would be good for a 64 CPU machine of course, max=6400% there. But on an 8-way box I assume something in the CPU time is wrong. In fact, I seem to remember that Sun did something odd. IE in normal threads the CPU time returned is the time per thread. But on Solaris (IIRC) CPU is the sum for all threads. But Crafty does the summing internally. That would make the above numbers make sense. Divide by 8 since the CPU time is off by a factor of 8, and the numbers look like 593%, 604%, 637% and so forth...
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.