Author: Dana Turnmire
Date: 22:30:29 11/30/03
Go up one level in this thread
On November 30, 2003 at 20:20:29, Michael Conway wrote: >ChessGenius 2.0 has native ARM code for its engine and allows hash tables up to >8MB. On my Palm Tungsten C (400 MHz) with ARM selected, it calculates >approximately 65,000 positions per second! This is in contrast to around 2,000 >positions without using ARM - an increase of over 30x!! (These are based on the >positions counter in ChessGenius - new in v2.0). Chess Tiger 15.1 on the other >hand calculates around 1,500 positions per second. > >As a quick test I played four 15 0 games between the two engines using ARM and >8MB hash table for ChessGenius 2.0 and the standard engine with 12MB hash table >for Chess Tiger 15.1. > >ChessGenius 2.0 won all four games. One game was a dead drawn position >(opposite color bishops with ChessGenius up a pawn) but Chess Tiger in a time >crunch played poorly at the end and lost. The other three games were resounding >defeats. > >Needless to say, the addition of hash tables to ChessGenius (along with its much >stronger engine) helps it play MUCH better in the endgame than it used to. > >I plan to play additional 15 0 games (maybe a total of 20) between to two >engines. When I'm done, if there's any interest, I'll post the PGNs. > >It's time for Christophe to write native ARM code for Chess Tiger because for >now, he can no longer claim "Chess Tiger for Palm is the strongest chess program >for the Palm organizers". Now if Mr. Lang will start working on Genius 7.
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.