Author: Ernst A. Heinz
Date: 01:54:55 05/29/98
Go up one level in this thread
On May 29, 1998 at 04:42:06, Ernst A. Heinz wrote: > >The long-standing rule of thumb in the computer-chess literature >states that each doubling of hash-table size gives you about >7% better performance (see e.g. Hsu's Ph.D. thesis). Forgot to mention that this only holds *on average*, of course. With just a small sample of positions anything can happen, i.e. you need at least hundreds of positions to really test the influence of hash-table size on "Fritz5" performance. =Ernst=
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.