Author: Vincent Lejeune
Date: 03:12:56 06/14/99
Go up one level in this thread
On June 14, 1999 at 02:00:20, Howard Exner wrote: >On June 13, 1999 at 19:26:40, Vincent Lejeune wrote: > >>On June 13, 1999 at 19:02:18, Howard Exner wrote: >> >>>Here's a request about the different hardware that will be used >>>at the upcoming WCCC. Once the hardware is listed for each program >>>could some kind soul sort it out in terms of chess computational strength. >>>The single processor variety of Intel and AMD I'm ok with but >>>have no clue on the Quad Xeon. How does a Quad Xeon compare to >>>a p3-500 for chess speed? Could someone also put into perspective how much >>>faster the big hardware(Cilkchess for example) is compared to the single >>>processors? >> >>A Pentium Xeon processor is about 10 % faster than Pentium for chess programs >>(at same frequency) >> >>It's not easy to estimate the power of Quad because it depend on quality of >>parallelisation of the algorithm ... Quad can produce between 2.5x and 3.5x >>speed up > >Poor programmers. It's not enough that they have to tweak their software, >now they have to fiqure out smart ways to squeeze out the the full potential >of parallel processing techniques. It's a part of tweaking software : where split the tree and ensure that there's no bug (very difficult thing as some programmers have reported) The hardware give a 3.95% speed improvement but technical difficulties (mainly splitting the tree) can't produce this 3.95x in node/second ...
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.