Author: Robert Hyatt
Date: 21:21:33 01/13/06
Go up one level in this thread
On January 13, 2006 at 18:11:47, George Sobala wrote: >On January 13, 2006 at 11:46:19, Robert Hyatt wrote: > >>On January 13, 2006 at 03:52:34, Gian-Carlo Pascutto wrote: >> >>>On January 13, 2006 at 01:07:52, Aloisio Ponti Lopes wrote: >>> >>>>Did someone test this? >>>> >>>>A. Ponti >>> >>>I saw CEGT is now testing dual programs and had Shredder 9.12x64 SMP on top of >>>the ratingslist (but with few games). >>> >>>I hope they can more as it will be interesting to see the actual ELO gain from >>>SMP. >>> >>>-- >>>GCP >> >> >>I don't see why it would be any different at all, than just comparing to a CPU >>that is X times faster, since that is all SMP does. I'd expect my dual 2.8 to >>be about as fast as a single 5ghz processor, assuming everything else (memory, >>cache, processor internals, etc) are identical... > >I'm surprised you say that with your experience with Deep Sjeng. > >In my experience with Deep Shredder on 1-4 processors, the behaviour and >performance of the multiprocessor program cannot be predicted or simply >extrapolated from the behaviour of the program running as a single thread. In >test positions, the 4-processor version may "find the move" anywhere between x1 >and x20 as fast as a single thread, and then come up with completely different >times when run on the same position again. 4x2.5GHz do not equal 1x10GHz (or >1x8Ghz!). I'm not sure (a) what you mean by "my experience with deep sjeng"; and (b) what the rest has to do with my comment. So what if one position is far faster, and another is slower? If I average going a ply deeper, I'm going to play stronger. You can see the _same_ effect with one processor. Double the speed and sometimes you get another ply, sometimes nothing because the search explodes. But over the long-run, if a dual-processor program runs 1.7X faster, it will be about equal to a single CPU that is 1.7X faster than the original box as well. Even if a program were to be so erratic that the SMP search goes one ply deeper three moves, and one ply shallower one move, I'd still rather have that program because the average depth is deeper, and it will play stronger overall. The "noise" in the non-deterministic search doesn't change the overall result. So _overall_ (speaking only for my program) two processors is about 1.7X faster overall. And it is stronger overall. About as strong as if you somehow overclocked the original processor 70% to get it to 1.7X faster...
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.