Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: is this the best simulation for processors and chess performance?

Author: Russell Reagan

Date: 16:40:16 07/22/03

Go up one level in this thread


On July 22, 2003 at 17:44:29, Rajen Gupta wrote:

>jokes apart; i have noted that these benchmarks have correlated well with fritz
>marks...

When did they release a 64-bit version of Fritz optimized for the Opteron? Oh, I
guess you're comparing 32-bit executables on 64-bit hardware again, which is
meaningless.


>i have no idea how well the opteron will fare for current chess programmes but
>if these benchmarks are comparable for chess performance for the opteron...
 ^^
 ^^
If, if, if...


>as well
>then (as they have been for other processors) the opteron 244 is only on par
>with and Xp 3200 and i predict a not too bright future for the opteron:

Again, running 32-bit code, you're right. Guess what, no one here cares about
32-bit code on a 64-bit chip. When are you going to understand this? When
someone compiles Crafty with a 64-bit compiler that has optimizations for the
Opteron, then we can discuss whether the Opteron is better than other cpus for
computer chess. Until then, there is nothing but meaningless data to compare.
It's like taking a 16-bit executable and running it on the newest 32-bit chips
to determine which chip is the fastest. That would only measure the chip's
ability to run 16-bit code, and I don't suspect there are many people in the
world who care much about that.


>lets have a bet: 1 yr from now the best performing
>intel will blow the best performing AMD out of the water; i'll bet you a copy of
>the latest chess software ! you can bookmark this message and get back to me on
>21-07 2004!

I am no AMD zealot, and I don't care who is faster (I find it hard to believe
that either side will "blow the other out of the water" though). I care when
someone makes unsound conclusions based upon manipulated or false facts, which
you seem to do once every week or two when you post benchmarks that either
aren't related to computer chess, or didn't use 64-bit executables, etc.

Benchmarks are like statistics, and they can be manipulated to imply whatever
the author desires. Compile crafty with a 64-bit compiler that has optimizations
for the Opteron, and then we'll see how the NPS of the Opteron fairs against 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.