Author: Robert Hyatt
Date: 10:18:05 12/17/02
Go up one level in this thread
On December 17, 2002 at 12:11:34, Vincent Diepeveen wrote: >On December 17, 2002 at 12:05:19, Gian-Carlo Pascutto wrote: > >>On December 17, 2002 at 11:58:21, Gian-Carlo Pascutto wrote: >> >>>On December 17, 2002 at 11:27:18, Matt Taylor wrote: >>> >>>>Crafty gets better results with HT, >>> >>>In addition to what Vincent said, the data we currently >>>have is saying exactly the opposite. >> >>Gah - forget I said this. I was looking at the wrong logfile. >> >>I don't have data for Crafty and Diep for a 2 CPU without and >>4 CPU with hyperthreading setup yet. > >yes there is a test done diep without 2 cpu's and without HT. >In diep's log i always write down the number of cpu's it's tested >at. (2) means 2 cpu's. > >for crafty there is indeed no log yet of 2 cpu's without versus >2 cpu's with HT. > >I gave you that 18.15 version which prints out exact node counts >for crafty (possibly it could print out only a wrong number when >it splits in the root which provable is not possible when it is >still searching the first rootmove, but at the first PV and the >end of a ply it will print out the correct number of nodes >as one can proof easily). Your "proofs" leave a lot to be desired. Facts, in particular. By the time it finishes searching the first root move, it _can_ already be searching other root moves. Your node counts were wrong. They will continue to be wrong. And as far as the SMT vs no-SMT discussion, they are irrelevant as well. > >Best regards, >Vincent > >>-- >>GCP
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.