Author: Robert Hyatt
Date: 19:20:23 10/05/99
Go up one level in this thread
On October 05, 1999 at 21:37:42, Christophe Theron wrote:
>
>
>It has been said from the beginning of this thread that we were talking about 2
>programs, both PB off, running on the same computer.
>
>This is mentionned on top of this message, BTW.
>
I realized that. But every time we start discussions about two programs,
one computer, thinking on opponent's time disabled, someone always asks
"if that is a problem, what about 2 programs, 1 computer, pondering _on_?
How can that cause problems?"
I tried to head that off at the pass..
>
>
>> If PB is
>>off, then timing changes cause difficulties.
>>
>>
>>>1) That you don't use 100% CPU time when probing tablebases? In this case it's
>>>up to you to do whatever you can to use 100% of your CPU...
>>
>>Care to tell me how to do a read and then not wait on the result? :) I
>>_must_ know the result of a TB probe before I go on and start any work at
>>this node...
>>
>>
>>>2) That your opponent is going to use some CPU resource while you are accessing
>>>the endgame databases? I don't think so. Either the other program runs in a DOS
>>>box that is STOPPED when it is in the background, or the other program is
>>>guaranteed to be idle when not thinking. So you can access as many files as you
>>>want or even sleep for a while, but your opponent is not going to do any
>>>computation in your back.
>>>
>>
>>if PB if off, the above doesn't happen, of course, _UNLESS_ you are playing
>>one of the chessmaster programs that seem to like to burn the cpu when they
>>are sitting idle.
>
>
>Didzis takes care of that. Before playing two programs, he makes sure that one
>is not stealing CPU time from the other.
>
>This is also mentionned on top of this message.
>
How? IE it would be easy to notice that pondering is disabled, and at random
times, burn a fair number of cpu cycles in a loop. If no one looks carefully,
they would _never_ notice.
>
>
>>Testing on _one_ machine is a bad practice, period. There are too many
>>variables. way too many.
>
>
>But for me it looks quite simple.
>
>Each program runs when the other is TOTALLY idle. Permanent brain is OFF in each
>program. Careful tests are conducted before the game to make sure one program
>doesn't steal CPU time from the other one.
>
See above. If one author is dishonest, your 'careful test' means nothing
at all.
>I have been using this way of testing since years and it works fine.
>
>You'd better simply say that you don't want to invest some time to make your
>time management work correctly in this case.
>
>But don't bring discredit upon Didzis because your time management is
>inappropriate.
>
I wasn't trying to "bring discredit upon anyone". I was simply pointing out
that _my_ program is not designed to be tested in a 'broken' environment. Ed
said the same thing about Rebel. If anyone wants to do it anyway, more power
to them. But it doesn't mean a lot, IMHO...
IE if I say that Crafty needs at least a P6/200 to run decently, someone can
_still_ run it on a 486/33 if they want. I simply choose to not put much
significance into the games...
>
>
>It sounds like you are implying that testers using only one computer are stupid,
>as if they were doing the thing you describe just above.
>
>Thanks Dr Hyatt, I'm one of them.
>
>
>
Why so defensive? read my lips:
"Crafty is not designed nor tested for playing games with ponder=off.
I see _zero_ reasons to test/develop/tune the program in that mode,
because I will _never_ play a serious game in that mode."
That is all I have said. I did _not_ say anyone was stupid. I only said
the _results_ don't mean a lot if one of the two programs is _mine_. Who
better to know that?
>
>>>So to be clear you suggest that Crafty should be left out of tournaments played
>>>on one computer only?
>>>
>>>
>>
>>yes...
>
>
>Testers using Winboard to play tournaments, as well as Fritz user playing
>tournaments inside Fritz, are certainly going to be very pleased to read this.
>
>
>
> Christophe
I don't distribute crafty to please anyone. I distribute it to run on a single
computer, by itself, and play chess against an opponent, whether it be a human
or another computer. If someone wants to run it on an abacus with 4 other
programs at the same time, they can feel free to do so. I reserve the right to
point out that for _my_ program, this doesn't result in a representative level
of play. It simply is _never_ tested in that mode. I have no time to test it
and tune it in that mode. I have no _interest_ in tuning it for that mode. Any
more than I have any interest in rewriting it in Java...
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.