Author: William Penn
Date: 21:51:17 08/15/03
Go up one level in this thread
I have now run Shreddermarks and kN/s tests, all with 64MB hash on the Shredder
7 (Chessbase) GUI. The two computers used were a 500MHz Celeron/256MB, and
2.6GHz Celeron/512MB.
I let the Shreddermarks run for quite awhile, 50-100 positions, and they didn't
seem to vary any more:
500MHz, Shreddermark=318 (85kN/s)
2.6GHz, Shreddermark=856* (230kN/s)
ratio: 2.7x
[*This Shreddermark alternated between 856 (230kN/s) and 928 (241kN/s) for
unknown reasons. Sometimes it was the one, sometimes the other.]
Tests were also run for 10 minutes each with different engines, using infinite
analysis mode on the starting chessboard position:
Shredder 7 Shredder 7.04 UCI Fritz 5.32
500MHz 76 kN/s 72 kN/s 416 kN/s
2.6Ghz 194 kN/s 170 kN/s 1433 kN/s
ratios: 2.6x 2.4x 3.4x
I know from experience that the ratio with the Shredder 7.04 UCI engine is
usually about 2.2x in typical opening/middlegame positions.
Maybe I learned something(?). It looks like the faster Fritz 5.32 engine
benefits most from the faster processor. I'm curious whether this would hold
true with AMD processors?
WP
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.