Author: Robert Hyatt
Date: 21:43:36 01/27/98
Go up one level in this thread
On January 27, 1998 at 22:41:12, Richard A. Fowell (fowell@netcom.com) wrote: >On January 26, 1998 at 09:31:21, Robert Hyatt wrote: > ><massive snip> > >> >>from my perspective, after playing hundreds of games vs both on the >>various >>chess servers, cm5000 plays better than hiarcs against Crafty. Of >>course, >>CM5000 is more tactical than hiarcs, and maybe the faster search in >>Crafty >>makes this tactical difference between hiarcs and CM5000 more important. >> >>But if you play an extended match crafty vs hiarcs, and crafty vs >>cm5000, >>IMHO CM5000 will win more games than hiarcs... *against* crafty... > >So, for blitz chess against Crafty, then: > >CM5000 > HIARCS 6.0 >? Crafty > MacChess, perhaps, >although transitivity isn't a given. I'm not sure. I doubt hiarcs is going to do real well against Crafty at blitz. Search speed is too different, and at shallow depths the extra couple of plies from Crafty may well be too much. At least based on the few games I have watched on servers. I can't comment about longer games. There are *far* more cm5000 users than hiarcs users on the servers, so data is more scarce. Most common program (after crafty) is genius. Next is fritz. the others are used far less frequently. I'd assume due to their blitz skill since most games are blitz... > >On another topic, I have another result in the >Crafty 14.5 "a3" test, that is to say: >(book off, ponder off, hash 6M, hashp 2M, sd 9, a3) > >84654 nps - Mac G3/300 (cpu 100%) >41608 nps - Mac 7300/180 (604e) (cpu 100%) > >Last I checked, the Alphas were doing the best on this test - what's >an Alpha number like these days? at least double the above...a3 my P6/200 hits around 43K here. the alpha/500 we used in Paris was almost exactly 3.1X faster. But there are now *lots* of alphas faster than 500mhz... so a new one might well be 5x faster... > >(terminal output for the G3 was:) >time: 35.81 cpu:100% mat:0 n:3031491 nps:84654 > ext-> checks:51451 recaps:19782 pawns:24 1rep:44871 > predicted:0 nodes:3031491 evals:1220247 > endgame tablebase-> probes done: 0 successful: 0 > hashing-> trans/ref:20% pawn:92% used:w96% b92%
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.