Author: Dann Corbit
Date: 11:41:08 07/15/99
Go up one level in this thread
On July 15, 1999 at 11:19:16, John R. Menke, Sr. wrote:
>Greetings,
>
>On my old i486 dx2 66MHz: Hiarcs 7.32 is about 1 kN/sec. Fritz 5.32 is about
>16 kN/sec. (Incidentally, CM6000 is also approximately 1 kN/sec on this
>system.)
>
>So Fritz is about 16 times faster than Hiarcs (or CM6000). Is that normal?
What you have discovered is that there is more than one sort of chess program.
There are fast searchers, like Fritz and Crafty, and there are slow searchers
like Hiarcs and CS-Tal. Since a slow and a fast searcher top the SSDF list and
are within one point of each other for x-bar, we cannot make any guess about
which is best even for state-of-the-art programs.
A much more accurate (but still imperfect) measure of the power of a chess
program is the depth in plies {half-moves} reached in a given time for a given
position.
Two programs which have reached and finished an equivalent depth in plies will
have a very similar quality of answer {barring bugs}.
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.