Author: Paul Doire
Date: 14:40:14 12/08/00
Go up one level in this thread
On December 08, 2000 at 03:53:11, Didzis Cirulis wrote:
>On December 08, 2000 at 00:39:58, Didzis Cirulis wrote:
>
>>On December 07, 2000 at 21:01:24, Paul Doire wrote:
>>
>>>Maybe threr is hope for same PC matches.
>>>I recently downloaded BinaryWork CPU Controller 1.0 from
>>>www.tweakbase.com.
>>>
>>>It allows you to set cpu control for different applications
>>>on a scale from minimum to maximum-not recommended,normal and high
>>>are other settings.
>>>
>>>I have used in in matches with CM6000 a very cpu hungry program
>>>with some success.
>>>
>>>I hope other same PC match hopefuls will download and try this,
>>>then post their results here.
>>>
>>>Didziz I hope you read this, try it,if you haven't already and provide feedback
>>>here at CCC.
>>>
>>>Regards,
>>>Paul
>>
>>OK, I will. Wait for my comments in next 24 hours. :-)
>>
>>Didzis Cirulis
>
>It can be used, and here is how:
>
>First of all I have to say that CM8000 is a very tricky program regarding the
>CPU usage. When pondering disabled and no other chess program running, it takes
>some 1 or 2% of CPU. That can be tolerated, BUT (and this is a relly big BUT) as
>soon as some other chess program has to calculate its move, the Chessmaster.exe
>silently jumps on CPU and takes 8 to 14 % of it!!! As the result of this
>"malicious" activity, the other program (Gambit Tiger in this test) has only
>about 86% instead of usual 97 or 98%. The Tiger's interface is much more polite
>and has no influence on CM.
>
>Now about the CPU control utility and its application in this situation:
>
>1) Start CM as usually
>2) Start the CPU control utility
>3) Use the utility to start the Tiger AND when it is running, assign it the
>Highest priority!
>
>This stops CM from getting CPU while it is Tiger's turn.
>
>Data:
>
> Computer: Celeron 500 with 128 RAM
>
>Programs:
>Gambit Tiger 1.0 48MB RAM for Tiger. No book. No pondering. Fixed search depth
>= 13 ply.
>
>CM8000. Special test personality = no pondering and no book. Level = Game in 999
>minutes. Test depth = 10 ply
>
>Test move = e2e4
>
>
>1) Tiger running alone. Time to complete the task = 4 min 34 sec.
>2) Now the same for Tiger + CM8000 in background. Both programs started normally
>= no CPU utility used. Tiger completes the same task in 4 min 48 seconds.
>3) Programs started as described above = CM as usually and Tiger - using CPU
>utility and with the Highest priority. Now Tiger completes the same task in 4
>min 35 sec !!!
>
>Now from the CM perspective:
>
>Running alone. The first time when the PV for depth = 10 ply shows up in the
>Thinking Lines window = 41 sec
>Running with Tiger (started using the CPU utility) = the same 41 sec !!!
>
>This was very first and superficious test. I will let you know if something
>intersting shows up.
>
>P.S. Previous Chessmasters were much easier to handle for matches on one
>computer in Windows 98 environment. Tricky, but easier :-(
>P.P.S. If my CM could run in Windows 2000, there would be no problem, as W2K is
>perfect for suppressing the background programs. Unfortunately my CM8000 keeps
>thinking that it is running outside of its GUI and answers at 3 ply depth at any
>time control.
>
>Have a great chess!
>
>Didzis Cirulis
Thanks Didziz
Mw testing is as follows:
Contestants-Gambit Tiger1.0-Brain OFF-Hash 1MB-No Book
CM8000-Ponder Off-Hash(default 1Meg)-No Book
Tiger set to 12 ply fixed depth search
CM8000 set time=999minutes-watch thinking lines window for 12 ply
FIRST GT's PERSPECTIVE
No Utility
Alone time=44secs WithCM8000 in background=45secs
With Utility
GT High/CM Normal= 45secs
GT Normal/CM Minimum= 45secs
GT High/CM High= 45secs
1st conclusion CM8000 in background makes GT take 2.272%longer
regardless of cpu priority set by utility.Utility seems to have no effect!
CM'S PERSPECTIVE
Alone time=4.05 With Gambit Tiger13 in background=4.06
With Utility
CM High/GT Normal= 4.06
CM Normal/GT Minimum= 4.06
CM High/GT Minimum= 4.09??!!
2nd conclusion GT in background makes CM take .246% longer regardless of cpu
priority set by utility.Utility seems to have no effect as tested,except for
Unusual response from CM high/GT minimum.
CM's presence has more effect on GT than vice versa,but not as much as I
would have thought.I will test GT against CM6000 a notorious cpu hog and post
results.
Need more testing.Have played games with ponder and the utility seems to
make a much bigger difference in allocation of ponder time.
Regards,
Paul
This page took 0.01 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.