Author: Vincent Diepeveen
Date: 14:00:45 12/18/00
Go up one level in this thread
On December 18, 2000 at 15:52:05, Frank Quisinsky wrote: >On December 18, 2000 at 13:43:41, Vincent Diepeveen wrote: > >>On December 17, 2000 at 14:24:13, Frank Quisinsky wrote: > >Hi Vincent, > >we speak about a Chess-Base engine !! >All is not a problem ! >I can adjust threads = 1 and Deep Fritz used 1 processor on my dual machine. So aha cool, didn't know chessbase also used the crafty enhancement to allow to adjust the number of threats. Where in the interface is that, i happen to own those progs and probably never figured out where to do it in the interface. >I can play with ponder engine/engine matches without any problems. Program A and >Deep Fritz (program B in this example) get 1 processor ! > >The reason for me is to see how good is the new version, because Frans say to me >in Leiden that the new version (in Leiden T26 or T27) is much better (also with >1 processor). I saw in my test games that Deep Fritz is better ! > >Please check my News Ticker Page 42 for more information. Here you can find a >picture from the engine settings of Deep Fritz. > >Best >Frank > >> >>I don't know how Frans implemented his parallellism, >>but it doesn't sound like a good idea to run it parallel with >>another engine playing against it, unless the >>other program runs on a different computer. >> >>With 2 processors running 3 program threads/processes means >>that one process in turn will not get system time and will >>get blocked. In about 50% of the total time that will be one >>of the parallel engines. If a parallel engine gets blocked >>very likely real soon locks inside the engine will prevent >>the other process from getting further too. >> >>Running a processor dual means you definitely NEED 2 free cpu's >>(free of work load) otherwise the parallellism works against you. >> >>If you run diep dual on a single cpu for example then the first 30 >>seconds it will search like a couple of hundreds of nodes. >> >>Only when the locks are near the root and the search depth is huge, >>then the problem gets a lot smaller as both can search without >>locking each other too much. >> >>>Hi, >>>in the next 3 months I play with 5 new engines in my CCE tourney >>>(biggest computer chess tourney). >>> >>>01. Tourney "B": WB LambChop 10.x (very strong, I think 2.425 - 2.525 ELO). >>>02. Tourney "B": WB Insomniac 0.69a (older version with 2.450 - 2.475 ELO). >>>03. Tourney "A": CB The Crazy Bishop (I believe 2.375 - 2.425 ELO). >>>04. Tourney "A": CB Nimzo 8 (I believe 20 - 40 ELO stronger as Nimzo 7-32). >>>05. Tourney "A": CB DeepFritz (I believe 20 - 40 ELO stronger as Fritz 6a). >>> >>>We will see ... >>>New results daily ! >>> >>>Dual Pentium III 733 MHz, Ponder = on, 64 MB Hash-Tables, 40 moves in 40 >>>minutes, 4-piece nalimov TBs with 4 MB cache ! >>> >>>More information on my webpage ! >>>http://amateurschach.in-trier.de >>> >>>Have a nice day ! >>> >>>Best >>>Frank
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.