Author: Robert Hyatt
Date: 18:58:18 05/04/01
Go up one level in this thread
On May 04, 2001 at 18:14:20, Bertil Eklund wrote: >On May 04, 2001 at 15:02:04, Robert Hyatt wrote: > >>On May 03, 2001 at 11:10:51, Bertil Eklund wrote: >> >>>On May 03, 2001 at 10:45:45, Mogens Larsen wrote: >>> >>>>On May 03, 2001 at 06:37:52, Larry Proffer wrote: >>>> >>>>>>The so called impartial experts should recommend the "best" programs not all >>>>>>programs in the world that in theory can be better. In April I believed that it >>>>>>was Fritz, Junior, Shredder and Tiger(s) (alphabtical order). I don't believe it >>>>>>is a coincidence that the best programs ARE commercial. >>>>> >>>>>But this was not a pre-condition as suggested elsewhere by others .... >>>> >>>>That one was killed a long time ago IIRC, by various reports from the involved >>>>parties. The only requirements stated then and since was the SMP capability and >>>>naturally significant (superior) strength of the engines involved. Of course you >>>>can choose to invent interests of the sponsors and the obviousness of the >>>>superior strength of commercial programs. Some have tried that without much >>>>success IMO. >>>> >>>>The real problem is the experts opinion of the strength and availability issue. >>>>Forming an opinion on the strongest programs without even a rudimentary >>>>investigation is nonsensical by default. AFAIK you do not develop psychic >>>>abilities after being involved with computer chess for 25 years. The organizers >>>>would have a real explanatory problem if they included the SMP Tiger. >>> >>> >>>It wasn't time to play around and invite 200 programs. Instead you can check the >>>results from the SSDF-list, tournaments from a lot of people and so on. >>>Therefore it is quite easy to see that the four best (comp-comp) programs are >>>Fritz, Junior, Shredder and Tiger(s). Of course you can suggest 100 other ways >>>to check the strength, if you had all the money and all the time in the world. >>> >>>Which programs are better then the above? Give me five, so can we see how many >>>that agrees with you. >> >> >>That is simply _wrong_. SSDF doesn't test with SMP hardware, so using the >>SSDF list to choose the best SMP programs is meaningless. >> >>As far as programs better than the above, how about Crafty on a 64 cpu Alpha >>system? >> >>Or ferret since he can almost use that hardware (it would take some changes but >>not a rewrite). >> >>You are using a tiddly-winks rating list to choose who is going to a >>discus-throw event. The best indicator would be the last WCCC tournament >>which had SMP and non-SMP programs in it. >> >>Rather than logic, we get this nonsense... > >Ok! Give me these five programs that are better then the above mentioned. >Of course the above programs can't use 2-4 or 8 cpus, it is just a gimmick from >the programmers. I have seen some tests from these programs that "fooled" me to >believe that they could use multi-processors. Ok, Crafty, Ferret and a few >others are better with multi-cpus but Fritz, Junior and Shredder are still the >same as with one cpu. > >Bertil I can't answer for Fritz, Junior and Shredder. But I can definitely answer for Crafty and can say with no hesitation that it is not the same on a single cpu and a quad-cpu machine. I really don't think that Fritz or Junior are the same either. Both should get reasonable performance and a factor of even 3 on an 8-cpu machine makes a _huge_ difference as it is a ply or more.
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.