Author: David Blackman
Date: 20:44:51 09/03/00
Go up one level in this thread
For a normal chess engine, it just won't help at all. There are several slightly unconventional ideas that might make use of these instructions. One idea would be to use a neural net for the static eval function. This would be a lot slower than the static evals most people use now, but if it was also better, it might be ok. SSE or 3DNOW! would be quite helpful for making the neural net calculations go faster, both when you run the program, and when you do the very long, ofline, calculations to optimise the parameters.
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.