Author: Shep
Date: 03:01:51 08/05/99
Go up one level in this thread
On August 05, 1999 at 05:46:45, Harald Faber wrote:
>>CM 5555 is a good example; its relative strength decreases when going from
>>40/120 on 233 MHz to 40/120 at 400 MHz.
>
>I don't think so, regarding that in KKKKup CM5555 performed very well.
Really? I thought they used the "plain" 5500!? Hmm, maybe my memory is wrong.
Anyway, 5555 did not play the endgame well against Nimzo 98 in my Long TC
Tournament B (10 min/move on P6-233).
>>CM6666 may suffer from the same problems (it is optimized for 40/120 on my 550
>>MHz machines).
>>
>>My personal opinion is that for overnight analysis, a high selectivity is indeed
>>the best choice.
>
>Also for tournament t.c.
Again depending on the "effective" reflection time, i.e. time control _plus_ CPU
speed.
I will finally be able to put some more efforts into this area as I have written
a little tool that creates CM position files from an EPD file (if anyone is
interested, this is in Perl only (!) and will be up on SCCS next week).
So I can now run several settings through some test suites.
(But don't fear, this is not likely to delay the publication of the 6666
settings. ;-)
I tried 6666 vs. 6666 DEEP ("DEEP" has SS=32!) yesterday on FIN08 and FIN09 of
the LCT2 suite. I ran each position for 3 min only on a 233 MHz machine.
Neither program could solve either position; the DEEP variant search several
plies deeper but spent only little time on any move other than the PV...
---
Shep
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.