Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: FPGAs playing chess--an expert opinion

Author: Vincent Diepeveen

Date: 14:26:44 12/19/99

Go up one level in this thread


On December 18, 1999 at 06:29:22, Tom Kerrigan wrote:

>On December 17, 1999 at 21:50:32, Vincent Diepeveen wrote:
><cut useless, uninformed tirade about DB's eval function>
>>Apart from this discussion great discussion about FPGA. You're however
>>comparing with deep blue. I'm only interested in how much faster my program
>>can get if i for example put my evaluation to FPGA.
>
>I think that putting your eval() on a good FPGA wouldn't be too hard. The main
>problem is that every time you want to run eval(), you have to send the FPGA the
>board position and grab the result. So that's a bottleneck. I think you could
>probably get a net speedup, but it's not like your eval() will run infinitely
>fast.

At a PII450 my program is doing 2000-4000 full evals a second.

How much at an AGP card for example could this be speeded up theoretical?
Giving the board positions isn't much bytes.

I just need to give next at least to it:
integer (so 32 bits or 4 bytes):
  piececount[2]      ==> number of pieces
  CanCastle[2]       ==> can you castle?
  RookA[2]           ==> whether a rook has moved
  RookH[2]           ==> whether h rook has moved
  Piecelist[2][16]   ==> squares pieces are and type of pieces

So in short we need to give it 32+8 = 40 x 4 bytes an integer = 160 bytes.

Currently i'm doing a lot of incremental stuff for datastructure. In fpga
this must be first generated based upon this information. Only after that
one can go further.

If we can do this faster than 2000-4000 times a second (including getting
an integer back which is the evaluation result) then it would be nice
to know an estimation how much faster.

>-Tom



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.