Author: Sune Fischer
Date: 13:01:44 08/08/03
Go up one level in this thread
On August 08, 2003 at 05:51:21, Robert Allgeuer wrote: > >>>I know this as well, your posting is not particularly helpful. >>>Still, knowing the speed-up for pure 32 bit code on an Opteron does make sense >>>for those who might want to buy an Opteron machine for chess today, when most >>>programs have NOT been compiled or released for 64 bit, yet. >>> >>>Robert >> >>So it doesn't make sense to buy a P4 either nor a K7, because the software they >>run at it is optimized in a much better way than the software you go install at >>it. >> >>You understanding what i say here? >> >>If you guess you do. Then first reread it carefully. >> > >A comparison of 32 and 64 bit code running on the same Opteron hardware is >beneficial for several reasons, for technical and practical ones. If you have an old program and want to know how it does on the new chips, then it's interesting. But if these programs are in any way speed criticial, then it probably won't take long for an updated version to arrive. I guess it's just a transition phase, today you don't really see anyone caring about 16 bit performance, same will be for 32 bit performance before long. Naturally some will transcend faster than others :) >It appears >that I am not the only one who thinks like that, please read the PS section in >Eugene's post a few lines above at >http://www.talkchess.com/forums/1/message.html?310380 Eugene is pointing out that clock for clock is a meaningless measure. It's not directly a 64/32 bit issue. -S. >Robert
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.