Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: 0x88 and move generator speed

Author: Dan Newman

Date: 17:21:32 01/29/01

Go up one level in this thread


On January 29, 2001 at 16:32:19, David Rasmussen wrote:

>On January 29, 2001 at 12:28:10, Severi Salminen wrote:
>
>>>  that, if I'm not wrong, is call "Vincent's test" (because Vincent Diepeveen
>>>created it), where you have to generate all moves 2,000,000 times, I was getting
>>>about 9,500,000 moves per second (8.5 seconds for the whole test) in my AMD
>>>Athlon 550, with my old board[64].
>>
>>So in this test (I have BTW used this position allmost exclusively for testing)
>>you generate 2000000x moves for white only? No move makeing and unmakeing? I'll
>>give my results tomorrow. Could you post also the result of Crafty for
>>comparison. Run perf for this position. It does 1000000x movegeneration.
>>
>>>  Tonight I'll continue with my make/unmake functions, which were my bottleneck
>>>in Averno. I realized that the inCheck detection determined the speed of all the
>>>program. Is there any "smart" trick for make fast in check detection with 0x88?
>>
>>I use bitboards so I have no experience in 0x88. It is quite fast with bitboards
>>as you don't have to loop anything. You just check where king is and then a few
>>ANDs and ORs and voila! But it would be interesting to see how a version without
>>any incheck routines would do. So basically you assume that a move is legal and
>>at next ply return if you can capture King. I think a "return beta;" could be
>>enough. (somebody correct me).
>>
>>Severi
>
>You should know that bitboards aren't especially fast on 32-bit machines. While
>many of us like the idea for it's simplicity, it is still not the fastest
>approach. As far as I know, none of the really fast programs (not necesarilly
>best programs), are using bitboards.

I'm currently getting about 1040 knps on a P3 running at 980 MHz with Shrike
which uses non-rotated bitboards.  I've been stuffing a lot of eval in too.
A lot for me anyway.  If I turn off lazy eval it drops to about 890 knps.
(All this measured on WAC.)

I Used to use 0x88 (and have tried a lot of other things too) but am finding
bitboards to be somewhat faster for some reason...

-Dan.



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.