Author: Will Singleton
Date: 23:40:26 09/17/99
Go up one level in this thread
On September 18, 1999 at 01:19:48, KarinsDad wrote: >On September 17, 1999 at 22:31:57, Will Singleton wrote: > >[snip] >> >>So, as I understand it, you are suggesting that the time it takes to generate 2 >>complete pv's (or more) is less than incorporating a pawn structure eval in a >>normal (single pv) eval. >> >>It would seem to me that in order to get 2 good pv's to the same depth would >>require twice the time as one pv. Of course, I'm on my second vodka & tonic, so >>maybe I'm missing something. >> >>Will > >Yes, that is basically what I am suggesting. However, since my search algorithm >is different, I'm assuming that it will not take twice as long for a second PV, >but maybe 2% to 5% longer (this is a guess not based on any empirical data yet) >since the definition of a PV in my algorithm is merely the best n moves found in >the normal search (which takes x amount of time regardless of which moves are >finally determined to be PV and sub-PV moves) extended somewhat in an attempt to >ensure some measure of correctness (or stability). > >And, as you state, some of the overall time can be saved by doing the more >complex and slower pawn structure evals just for the leaves of the PVs >calculated and not for every node evaluated (although any simple quick pawn >structure evals probably belong in the normal eval). > >I am not sure how standard Alpha Beta works for PVs, but if you tell me that >searching a second PV takes just as long as searching the first PV, I will take >your word for it. > >KarinsDad :) Are you attempting to pull my leg, KD? (Now on third drink.) Will
This page took 0.01 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.