Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Extension of the UCI protocol

Author: Russell Reagan

Date: 06:27:02 04/17/04

Go up one level in this thread


On April 17, 2004 at 02:58:10, Andrey Popov wrote:

>>* multipv <num>
>>this for the multi pv mode.
>>for the best move/pv add "multipv 1" in the string when you send the pv.
>>in k-best mode always send all k variants in k strings together.

>As far as I understood, the engine must silently accumulate
>the information (pv,score etc.) about all moves.
>At the end of each ply the engine must sort all these moves by score
>and send to the GUI something like the following:
>info multipv 1 depth 2 score 25 time 100 nodes 33000 nps 33 pv e2e4 e7e5
>info multipv 2 depth 2 score 23 time 100 nodes 33000 nps 33 pv d2d4 g8f6
>info multipv 3 depth 2 score 22 time 100 nodes 33000 nps 33 pv c2c4 c7c5
>All "time","nodes" and "nps" must from the end of the ply.
>Therefore, they must be equal for all moves. Am I right?

I don't think that is accurate. You don't have to accumulate information about
all moves. You only have to accumulate information about the k-best moves. You
could do a search and get the best move. Exclude that move from subsequent
searches. Do another search, get the best move (2nd best move overall), and so
on. During each search you won't search the same number of nodes or time, and
probably won't get the exact same nps.



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.