Author: Dann Corbit
Date: 10:04:10 02/15/04
Go up one level in this thread
On February 15, 2004 at 12:20:47, Russell Reagan wrote: >On February 15, 2004 at 07:49:56, Bruce Cleaver wrote: > >>>Why can't the use of floating point calculations be overcome? >> >> >>The use of floating point is intrinsic to the algorithm, involving >>exponentiation & division. IF you mean some sort of processor substitiution or >>hardware trick to speed up FP calcs, well that might be possible. >> >>You really ought to read the paper, Rivest had a beautiful idea. > >I don't know if this would work or not, but I remember reading about the early >days of 3D games when floating point calculations were not very fast, >programmers used to emulate "fixed point" numbers with normal integers. I guess >this is similar to how chess programmers use a score of 100 to represent a pawn, >and when they display that score it is 1.00. The speed of floating point is high enough now that you should not worry about that choice of floating point verses integer math. However, you should defintely not use floating point for MTD(f) evaluation or you will be searching forever. ;-)
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.