Author: Anson T J
Date: 16:17:00 07/21/03
Go up one level in this thread
On July 21, 2003 at 16:35:13, Uri Blass wrote: >On July 21, 2003 at 16:04:58, Rajen Gupta wrote: > >>I think playing against hiarcs is a pointless exercise: i am even surprised that >>it manged to win even a single gme: virtully every modern programme can thrash >>it beyond rcognition by a huge margin: one needs to test junior8 against its >>contemporaries including chessmaster 9000 and tiger 15 first. > >Hiarcs8 is not the best program but it is not a weak program. >20.5-7.5 is a very good result that make me suspect that maybe Junior used more >than one processor because of a bug and I read posts of people who could not use >Junior8 because of that problem(one had to use Deep Junior8 and the ssdf also >use Deep Junior8) > >Uri It might seem that way, but I can assure you that Junior 8 is only using 1 cpu here and Hiarcs 8 has a full cpu worth. I've had a lot of experience with duals. I spotted that the first version of Junior 8 had an affinity bug where it would only use CPU0 and not CPU1 since the update I've had no problems with it. Heres a screenshot from game 26 where you can see the nodes per second of both engines. Hiarcs had 200+ kN/s and is about right for Hiarcs 8 on a AMD 2400+ machine. http://mysite.freeserve.com/intagrand/junior8test/pics/j8-h8pic1.jpg Analysis from Hiarcs 8 (after the game without Junior 8 being loaded) [d]r2qr1k1/pb5p/np1P2p1/2p1bp2/P1B5/2N1B3/1P1Q2PP/R4R1K b - - 0 1 Analysis by Hiarcs 8: 22...Kh8 23.d7 ² (0.30) Depth: 12/37 00:03:07 36665kN 22...Kh8 23.d7 Re7 ² (0.32) Depth: 12/41 00:04:53 57688kN <-- 197kN/s As I've started the match, I'll allow it to finish up to game 50. I didn't expect it to be so one sided. I can only test the engines I own :)
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.