Author: Tom Kerrigan
Date: 21:33:21 03/22/00
Go up one level in this thread
On March 21, 2000 at 19:39:32, Peter Skinner wrote: >So is a Celeron faster than a PIII? > >How would that make sense? The Celeron has a faster (and smaller) L2 cache. If your program does not bang on very much memory, the Celeron will be faster. >The PIII has the SSE instructions, and a bigger cache. SSE instructions do not count for much. >And how would a PII 450 out perform a PIII 450? Wouldn't the extra instrustions >make the PIII faster? Simply because a processor has some extra instructions does not mean programs will use these instructions. In many cases, it wouldn't make sense anyway. If you need to add two numbers, you do not use a multiply instruction. Regardless, the SSE instructions are floating point, and chess programs do not do floating point. >In test outside the realm of computer chess, the PIII out performs almost >everything I have tested ( except the Athlon, but I do not like the AMD Then your tests do not agree with anybody else's... >from office applications, to online gaming. Within Computer Chess, I see that >Crafty produces more nodes/second, and also reaches deaper ply searches, and >really isn't that what computer chess is all about? Crafty bangs on a lot of memory, so it benefits from the larger cache. There are many chess programs that fit nicely in the Celeron's cache, and therefore run faster on the Celeron. -Tom
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.