Houdini bug(s)?

Discussion about chess-playing software (engines, hosts, opening books, platforms, etc...)
Post Reply
BB+
Posts: 1484
Joined: Thu Jun 10, 2010 4:26 am

Houdini bug(s)?

Post by BB+ » Mon Jan 24, 2011 1:10 am

This is not quite the behaviour I expected. Whether it is a "bug" is a different question.

Code: Select all

position fen r1b1r1k1/ppp2Bpp/8/4p1N1/1B2Q3/P3P1K1/5PPP/n2q4 b - - 0 0
go infinite
The output ends:
info multipv 1 depth 99 seldepth 2 score mate -1  time 11 nodes 100 nps 9000 tbhits 0 hashfull 0 pv g8h8
bestmove g8h8 ponder NULL
Note that it does not give the "full" PV (g8h8 e4h7), but only Black's move with White's mating move being elided.

Also, earlier versions of the UCI protocol were a bit hazy here, but with "go infinite", it is now (April 2004 and later, at the least -- ShredderChess has a 2006 version with the same) said that the search should not exit (as with pondering):
* infinite
search until the "stop" command. Do not exit the search without being told so in this mode!
So the bugs are:
* Failure to report the full PV here (and no ponder move, for that matter)
* Exiting the search prematurely in "go infinite" mode.

Matioupi
Posts: 28
Joined: Mon Aug 09, 2010 8:30 am
Real Name: MP

Re: Houdini bug(s)?

Post by Matioupi » Mon Jan 24, 2011 7:30 am

could this be linked to what i reported here :

http://www.open-chess.org/viewtopic.php?f=3&t=862

for which I got no explainations ?

Regards,

Mathieu

Post Reply