Author: Vincent Diepeveen
Date: 10:01:48 11/01/02
Go up one level in this thread
On November 01, 2002 at 12:10:25, Brian Richardson wrote: It is a VERY relevant comparision because basically intel claims that it is faster at especially DATABASE serving, because many lightweight threads there who now and then execute can make use of SMT they claim. Like my tests confirm it hardly works simply at the P4! The workload of databases is more relevant than computerchess with regard to selling servers :) Of course for computerchess reality is we can't use it at all despite some nonsense BOB is posting. If your program works great on a NUMA machine, yes even cluster then of course the chance is a lot bigger that one day in the future it might be able to profit from a better cpu than the P4 ;) >On November 01, 2002 at 05:14:10, Jorge Pichard wrote: > >> >>http://www.anandtech.com/IT/showdoc.html?i=1641&p=6 > >Fine, but this was a database test with a workload that looks nothing like >computer chess. > >My dual AMD shows Crafty SMP scalability efficiency of only about 1.4x and >others have posted Xeon SMP results like 1.8 or 1.9, IRRC
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.