Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: WMCCC Hardware

Author: Robert Hyatt

Date: 08:44:11 10/17/97

Go up one level in this thread


On October 17, 1997 at 09:54:08, Chris Whittington wrote:

>
>So the argument boils down to:
>
>Others do it, so its ok, more than ok, for me as well.
>
>Just supposing we collectively created a new title: WMCC-PC-Champion. We
>could do this. It could have cachet ....
>
>Would you then be persuaded to run on one of the fast PC's instead ? So
>we could make the tourney a 'fair' platform one as a result .... ?
>
>Probably we'ld need to get Bob to agree as well ...
>
>Chris
>
>
>



First I don't think you need me to agree on anything.  I'm but one
person.  I've been doing this longer than the rest of you by a wide
margin, and since my first ACM event about 25 years ago, *nothing*
has changed.  This has *always* been about both a good program *and*
a hot machine to run it on.  Chess 4.x is a good example, always
running on the fastest CDC machine around.  Cray Blitz is another
good example, running on $60 million machines every year.  The only
reason why I try to get the fastest machine I can is simple:  I don't
want to go to Paris and repeatedly smash my face into a brick wall,
just because it feels so damned good when I stop.  :)  To avoid the
smashing, I want to be relatively "equal" in terms of hardware, to see
how the program can play.  If everyone would cut the assembly bullshit
out, as many of us have, a "uniform platform" event could be done, where
*everyone* gets a quick platform to run on.  The new Mac machines are
better than the pentium II's, and getting better all the time.

However, if anyone organizes a "uniform platform" event, I'd go.  But
the WMCCC has *never* been uniform.  Years ago the dedicated machines
were running at clock rates way faster than anything that could be
bought.  So long as anyone is playing in the horsepower race, I will do
my best to play there as well.  I don't like it one bit, because it
takes
a lot of time to make these arrangements, test on the new hardware
before
the event, figure out how to transfer gigabytes of endgame databases and
opening books and stuff after arriving, and so forth.  I'd rather use
that
time to run and analyze games on ICC for example.

This event is only partially about the best program, it has a hardware
factor tossed in.  I don't think that I'd even bet money that one of the
766 alphas is going to win...  but I would be more on them than I would
on
a K6/233, because of the speed factor.  It's important.  How important,
I
don't know.  And I'd prefer *not* to find out by running on something
that
is 2-3X slower than the fastest hardware there.

You call this "self-serving"...  I tend to think of it as
"self-preservation".
Believe me, My first 10 years of ACM events came on hardware way slower
than
everyone elses.  I finally joined the arms race, and did much better.
Yes,
CB blew out the programs on slow machines.  But it also *competed* with
the
programs on fast machines.  I would suppose that if everyone wanted to
do so,
the ICCA could be convinced to make this a uniform platform event.  I
can
run on anything from a blender to a Cray T90.  I prefer the PC because I
have
one in my office and one at home, where I can run Crafty whenever I want
and
for however long I want.  Getting a T90 is a bit harder, although I
might
try for next year's WCCC perhaps.  And perhaps I'll simply show up with
my
P5/233 notebook and have fun.



This page took 0.04 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.