Author: Vincent Diepeveen
Date: 13:42:26 05/27/02
Go up one level in this thread
On May 27, 2002 at 14:58:05, Jorge Pichard wrote: >On May 27, 2002 at 13:15:36, Ron Langeveld wrote: > >>On May 27, 2002 at 12:38:19, Jorge Pichard wrote: >> >>> AMD Still the one to beat, unbelievable but true, or do we need more testing? >>> >>>Jorge. >> >>It's not a question of more testing, it's a question of clarity on test methods. >>I used default settings, and a common motherboard, the Abit KR7A with the KT266A >>chipset. Maybe Slater used a bad performing chipset or memory, or fiddled with >>settings in a dubious way, I don't know. Only he can give us this insight. > >I wrote the statement "unbelievable but true, or do we need more testing? >Simply because a Pentium IV 2.53 is faster than any AMD 1.73 Ghz, and your AMD >1.73 Ghz is NOT performing a better fritzmarks. Sorry, but I can NOT believe >your report. I have been several times with Ron and i know how scientific accurate he's doing his tests. I am already amazed the P4 gets that close to the K7. When i do similar tests for diep, the P4 is lightyears slower. I would prefer a bench of junior by the way. Not fritz. Fritz is P3 assembly with MMX. Not exactly 'an average chessprogram'. Just like yace, also not programmed with newer processors in mind (kind of K6 optimized). Hopefully changes soon when Dieter gets a K7 in his hands to toy with. Junior is C++ and a small program (so on paper favouring the small P4 cpu's on chip stuff) and i assume it is pretty deterministic i hope. I am however pretty amazed the P4 is getting so close just because of a faster access times to RAM and a bit bigger L2 cache. Crafty is an important benchmark simply because it is in specint. However let's look how official testers compare the things: http://www.specbench.org/osg/cpu2000/results/res2002q1/cpu2000-20020311-01272.html fastest AMD MP2000: crafty needing 103 seconds base time. http://www.specbench.org/osg/cpu2000/results/res2002q2/cpu2000-20020430-01347.html fastest P4-Xeon 2.4Ghz: crafty needing 125 seconds base time. In short the athlon is 125/103 = 21% faster than the P4. That answers questions? >Jorge.
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.