Author: Robert Hyatt
Date: 19:26:09 08/27/01
Go up one level in this thread
On August 27, 2001 at 16:19:00, Uri Blass wrote: >On August 27, 2001 at 14:43:34, Robert Hyatt wrote: >> ><snipped> >>What is there to understand? A potentially open file is a very concrete >>thing, just like an open file or a half-open file is. No confusing definitions. >>No multiple meanings. > >I do not understand and we need an exact definition because >every file can be opened by capturing the pawns in the file. If I have pawns at a4, b3 and c4, and you have pawns at a5, b6 and c5, then the b file is _not_ "potentially open". Yes if both sides hang the b pawns it would become open. The idea is that if I have a pawn at a2 and b2, and you have a pawn at a7 and no b or c pawn, then the b file is potentially open for me (and you). But only when _I_ choose to make it open. So for me, it may (or may not) be useful. For you it will definitely not be useful because I won't open it to your advantage, ever. I have seen this explained in several books over the years... > ><snipped> >>Not "difficult to do". I believe I said "impossibly slow". There _is_ a >>difference. Everything they do in parallel, you would get to do serially. >>All the special-purpose things they do in a circuit, you get to use lots of >>code to emulate. I estimated a slow-down of 1M. I don't think I would change >>this. Cray Blitz lost a factor of 7,000 from a Cray to a PC of the same >>time period. > >It was at that time period but I guess that today >the speed difference is not so big becaise the computers of today >are more than 10 times faster than the computers of that time. The difference is still significant. 7M on Cray Blitz was significantly stronger than 1M on my quad xeon. If you factor in that CB hasn't been modified in 7-8 years, we could call that "old". Yet it is _still_ very strong by today's program standards. And 10 minutes of programming would make interesting changes happen, like R=2/3 null-move for one thing. ><snipped> >>We know how DB (single-chip) did when slowed to 1/10th its nominal speed >>and played against top commercial programs. That was reported by me first, >>then others asked about it at lectures by the DB team and we got even more >>information from those reports. > >We never saw the games and it cause me to doubt the results. If you tell me you play a match with Crafty vs Program X, and you got a result of 9-1, I'm not going to doubt your results at all. I know the DB guys well enough to give them the same courtesy. > >> >>I am _certain_ that taking DB from hardware to software would cost a lot. >>You would lose a factor of 480 because of the chess chips. You would lose >>a factor of 32 because of the SP. You would lose a factor of something due >>to the cost of doing Make/UnMake/Generate/Evaluate in software during the >>software part of the search, rather than getting to use the hardware they >>had to handle these mundane parts of the software search. 32 X 500 is over >>10,000 already. And it is only going to get worse. > > >I only know that it is not clear to me how much you lose because the >loss from not doing parallel search is not clear and >it is not linear in the number of processors. In my testing so far, Crafty has been linear. I have run lots of tests from 1-8, and the formula speedup=1+(N-1)*.7 fits the data very well. I had a _few_ test results from a 16 cpu machine, but not enough runs to be sure it was what I would call reliable data. It fit pretty well too. > ><snipped> >>When your data is flawed, you need more. Crafty lost one game at a time >>handicap. Ed then played more games with crafty at the same time control, >>but with rebel at that time limit also. And the result was much different. >>Which suggests that the first (and only) handicap game was a fluke, which >>is certainly the most likely truth. > >There is one problem with the comparison and it is the fact that in the first >experiment old versions of both programs were used. > >correct comparison is to continue with exactly the same Rebel and the >same Crafty from the first experiment. > >Uri Why does it matter? Have I magically "gained ground"?? Or am I still "ten years" behind?
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.