Author: Tim Foden
Date: 08:10:24 03/27/02
Go up one level in this thread
On March 27, 2002 at 10:51:57, Matthias Gemuh wrote: > >Hi Experts, >I generate moves at 10 million/sec on Athlon 1.4 GHz using unrotated bitboards >(these include the illegal moves which would leave me in check). Are these 10 >million "pseudolegal"? Is this slow? How many would I generate with rotated >bitboards? Well, these simplest way for you to compare is to run Crafty's perf command on example positions, as Crafty is a rotated bit-board program. E.g. After e4 e5 d4 d5 Nf3 Nf6 Nc3 Nc6 we get this position: [D]r1bqkb1r/ppp2ppp/2n2n2/3pp3/3PP3/2N2N2/PPP2PPP/R1BQKB1R w KQkq - 4 5 On my machine (AXP1700, 1.47GHz) I get this (with an old version of Crafty): EPD Kit revision date: 1996.04.21 unable to open book file [./book.bin]. book is disabled unable to open book file [./books.bin]. Crafty v17.11 White(1): easy pondering disabled. White(1): setb r1bqkb1r/ppp2ppp/2n2n2/3pp3/3PP3/2N2N2/PPP2PPP/R1BQKB1R w KQkq - 4 5 White(1): perf generated 37000000 moves, time=1.48 seconds generated 24949426 moves per second generated/made/unmade 37000000 moves, time=5.56 seconds generated/made/unmade 6657070 moves per second White(1): So the rotated bitboards generate 24.9 million moves per second here. However, this doesn't tell the whole story, as rotated bitboards are slower in MakeMove, but can be faster in the static evaluation (depending on what you have there). Cheers, Tim. > >Regards, >Matthias.
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.