Author: Gregor Overney
Date: 17:48:58 05/19/99
Go up one level in this thread
Let's hope that's the case. I am very interested to see such a chip in action. The division in a fast and slow evaluation function is very intriguing. Feng-hsiung mentioned that during early tests, because of a hardware fault, the chess CPUs were actually running at 7% to 14% of their actual strength/speed. When he played against top commercial programs, a one-chip-design won all 10 games. This suggests that even playing with the same "performance" (roughly 200,000 nodes/sec), his chip wins. I must assume that his "hardwired" evaluation functions must be very sophisticated indeed. Gregor
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.