Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fritz5

Author: Don Dailey

Date: 14:46:03 12/19/97

Go up one level in this thread


Hi Thorsten,

I see that you have a K6 with limited memory so it was not possible
for you to test any other way.  But I do think it's more fair to test
each program on the platform it runs best on.  My new program completely
sucks on a Pentium of any kind so I never test on a Pentium and I
would consider a pentium test very unfair.  Also my program runs very
fast and therefore needs larger hash tables.  I would consider a test
with tiny hashtables unfair too.   But if I tested your program vs
mine using an emulator for your program on the Alpha, it would hardly
be fair for me to say, "Is it my fault his program is less strong on
my Alpha's?"

This may sound like a silly example to you but keep in mind that the
Pentium is FAR more common than your platform.  You cannot pick out a
platform and then criticize every program that doesn't run well under
it.   If the K6 becomes a huge success and becomes the common platform,
then you may have a reasonable argument.  It would be the same if I
marketed my program (as a pentium program) even though it was designed
and optimized for an Alpha.  Everyone could legitimately claim it was
weak and the burden would be on me to come up with a good pentium
version or only market it for the Alpha's.

But from now on I want to focus my discussions and energies on
algorthims, problems sets and such.  If you want to respond to this
note I'll be glad to read your response but let's close this discussion
soon ok?   A whole lot of this is grey area semantics anyway.

- Don







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.