Author: Enrique Irazoqui
Date: 09:06:36 10/19/99
Go up one level in this thread
On October 19, 1999 at 06:02:11, Didzis Cirulis wrote: >On October 19, 1999 at 05:38:26, Enrique Irazoqui wrote: > >>Didzis: Are you sure that playing CM and Tiger in the same machine doesn't hurt >>Tiger? One of the reasons I am asking is that your results and Marcus', posted >>on Gambitsoft forum, are the opposite. Yesterday I played a game CM6K-Tiger and >>my view coincides with what Marcus says. Maybe the explanation is that we use 2 >>machines? > >2 and faster computers may sure result in different results ;-) > >Now, Enrique, back to the game! ;-) I am doing this type of tests for quite a >long time. I usually do some speed testing and calculations before I play. > >In this situation I used Chess Tiger's postition cmb06 at ni=11 for my tests. > >What I did is: > >1) I run CT alone several times to see how much time does it take to reach this >move 11 > >2) Then I started CM 7000, selected my CM 7000 personality (PB OFF), made some >moves to be sure it is no more in book, turned on "Set Up Game" window to stop >CM from taking opponent's CPU - that is exactly the situation as in my test >games! > >3) Now, when there is CM 7000 sitting there in background, I returned to CT and >repeated the same test again: >(ni11, im, cmb06, ENTER ) > >4) what I got is insignificant slowdown that, processed as I described on my web >page at http://open.konts.lv/usr/Didzis/Two_Programs.html , resulted as follows: > Chess Tiger needs 61 minutes for a game while CM 7000 requires 60. > >5) I have sent all my calculations to Christophe and could not find any problem >there. > >Feel free to ask if any questions I checked all this with a middle-game position that can be considered "typical": 3rr1k1/pp1nqp1p/2pb1np1/8/3N4/PP2P2P/1BQ1NPP1/R4RK1 b - - 0 22 CM6K computed at exactly the same speed, with or without Tiger in the background: Ply 04/09 17'' 17'' Ply 04/10 98'' 98'' Ply 05/11 282'' 282'' Tiger was slowed down by 15%, admittedly not all that much, but still a P500 becomes a P435. Both programs ran on a PII-500, PB off and 16MB hashtables. Times/ply for Tiger 12.0, without and with CM6K loaded (and idling). N9 6.7 7.68 N10 13.29 15.21 N11 43.83 50.53 N12 123.19 142.42 NPS 115793 100159 Aside from this, Ed's findings about PB on/off seem quite conclusive to me. All this can contribute to explain why your results playing on one machine and the results of Marcus Kaestner and mine with 2 machines are so different, in fact the opposite. It is true that we are playing Tiger 12 against CM6K, but it has the same engine of CM7K, isn't it? Enrique >Didzis > >> >>Enrique
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.