Author: Tom Kerrigan
Date: 01:42:02 02/14/00
Go up one level in this thread
On February 14, 2000 at 03:50:18, Michael Neish wrote: > >Thanks Tom for the great effort. > >Now that you've eliminated searching all moves in response to check in qsearch, >has it made much difference to the number of nodes searched? I ask because I >noticed some time ago that about 90-95% of the nodes searched by TSCP are in >qsearch. I have no idea what this value should be for a "strong" program. Do >you happen to know? > >Cheers, > >Mike. In many positions, TSCP's search is now noticably faster. The big problem is that if you do check evasions in the quiescence search, you end up searching some very long lines. TSCP's move stack is 32 moves, and many 6-ply searches were exceeding that limit. Unfortunately, I didn't have any check for that. That's why TSCP was so unstable during the ICC tournament. All of the arrays were getting overrun. A few years ago, many strong programs were spending about 80% of their time in qsearch. This may be a little bigger or smaller now, due to new techniques. -Tom
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.