Author: Ricardo Gibert
Date: 16:48:15 03/30/05
Go up one level in this thread
On March 30, 2005 at 18:18:28, axel binevskiy wrote: >Pentium 4 is considered to be a top processor these days, so why is it not good >for chess engines testing?...can not calculate moves in the right way or what??? Chess contains a lot of branches that are difficult to predict. The branch misprediction penalty is significant ly higher with the P4. Read the section called Branch Misprediction Penalty at http://www.aceshardware.com/Spades/read.php?article_id=15000186 An illustration of how this is manifested in a chess program like crafty can be seen in the crafty section of cint2000 where the top Athlon 64 gets a 59.5 score http://www.spec.org/osg/cpu2000/results/res2004q4/cpu2000-20040920-03391.html which is much faster than 73.6 for the top P4 http://www.spec.org/osg/cpu2000/results/res2004q4/cpu2000-20041115-03597.html The designers of the P4 made a trade off in order to produce a chip with a higher clock speed. The tradeoff had its advantages and its disadvantages. Good for some apps and bad for others. Chess is among the "others."
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.