Search found 18551 matches

by hgm
Sat May 26, 2007 2:07 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: question about symmertic evaluation
Replies: 37
Views: 23734

Re: question about symmertic evaluation

It does not really matter if the individual scores for black or white are negative, and thus rounded towards -infinity rather than towards zero. As long as you do the same to both, the result will be symmetric.
by hgm
Thu May 24, 2007 11:55 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: question about symmertic evaluation
Replies: 37
Views: 23734

Re: question about symmertic evaluation

Yes, I try to avoid modulo wherever I can. Even unsigned (unless it is by a power of 2, of course). For instance, to generate a number in the range 0-4 from the hash key, I don't use Key%5, but in stead (Key&0xFFFFFFF)*5 >> 28. (I need that to pick a primary entry to hash to within each bucket. The ...
by hgm
Thu May 24, 2007 10:39 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: question about symmertic evaluation
Replies: 37
Views: 23734

Re: question about symmertic evaluation

Indeed, for division by a constant multiplication with the reciprocal seems the obvious choice. More tricky is the case where you have to divide through a variable unknown at compile time. The best way to do that would be to make use of the SSE (or 3DNow!) reciprocal instruction, (a fast table looku...
by hgm
Thu May 24, 2007 8:06 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: question about symmertic evaluation
Replies: 37
Views: 23734

Re: question about symmertic evaluation

The latency of divide instructions is such that they almost always cause a hefty pipeline stall. The re-order buffer is simply not large enough to hide the latency, even if it is not on a critical path. The divide cannot be retired before it has finished, and this means that it will stall the retire...
by hgm
Wed May 23, 2007 6:49 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Had anyone tried EvalHashTable?
Replies: 15
Views: 11216

Re: Had anyone tried EvalHashTable?

I found that for me bigger eval hash is better. Note that I evaluate every node and with small eval hash tables I am afraid I will get almost no hits. I also found that pawn hash table of some mbytes make movei slightly faster relative to no hash. Well, if your evaluation takes longer to compute th...
by hgm
Wed May 23, 2007 2:59 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Had anyone tried EvalHashTable?
Replies: 15
Views: 11216

Re: Had anyone tried EvalHashTable?

The main concern with an Eval Hash is that the evaluation takes a limited amount of time, and that you must make sure that the cost of the EvalHash probes is less than that time. Access to a RAM-based table (so one that does not fit in cache) takes 286 clocks on my (slow) Pentium-M machine, and is l...
by hgm
Tue May 22, 2007 10:19 am
Forum: Computer Chess Club: Tournaments and Matches
Topic: Deadline Extension Early registration WCCC
Replies: 11
Views: 5006

Re: Deadline Extension Early registration WCCC

You are not understanding the message fully. Originally entrants had to pay double after May 21. That has now been shifted to Juny 1. So it is basically a 10-day extension, with an extra good deal for those acting immediately.
by hgm
Tue May 22, 2007 9:44 am
Forum: Computer Chess Club: General Topics
Topic: Participants WCCC 2007
Replies: 108
Views: 51553

Re: Participants WCCC 2007

Yes, that was a shameful display by Joker, and is one of the reasons it doesn't play. I suspect there is an ilusive bug somewhere in Joker, because in that game it completely voluntarily gave the exchange in an extremely obvious way. (It put a Rook that could have gone anywhere and didn't have to be...
by hgm
Tue May 22, 2007 7:50 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Perishing Perft !
Replies: 34
Views: 24608

Re: Perishing Perft !

My makemove code uses the 0x88 delta trick you mentioned to figure out if in check, but I am currently doing this right at the end of makemove(). This is less than optimum as it would be best to do it at the start of makemove(), to save the move make and move undo code. That is much trickier to cod...
by hgm
Mon May 21, 2007 10:00 pm
Forum: Computer Chess Club: General Topics
Topic: Participants WCCC 2007
Replies: 108
Views: 51553

Participants WCCC 2007

Seems micro-Max will finish in the top 10 at the next WCCC! :lol: :lol: :lol:

Code: Select all

Rank Program Country Hardware Title  
0 Diep  NLD 4 core AMD K8 2.4Ghz  
0 IsiChess  DEU   
0 Jonny  DEU Intel Core2Duo@3.4GHz, 1 GB RAM  
0 micro-Max  NLD   
0 Rybka  HUN   
0 Shredder  DEU   
0 The Baron  NLD