Author: William Bryant
Date: 06:35:10 01/01/05
Go up one level in this thread
On December 31, 2004 at 23:42:59, Richard A. Fowell wrote: >On December 31, 2004 at 10:58:33, Roy Eassa wrote: > >>On December 30, 2004 at 03:35:13, Richard A. Fowell wrote: >> >>>My copy of Chessmaster 9000 for Macintosh showed up today ( I had pre-ordered it >>>through Amazon.com in the spring). It comes on DVD-ROM (and will not launch >>>without the DVD-ROM in the drive) and requires OS 10.2 as a minimum. >>> >>>Compared to Chessmaster 6000 Mac, the new version: >>> >>>- Runs native in OS X >>>- Has "The King 3.23" (7/7/2002) engine >>>- Apparently has endgame tablebases (though I have yet to get it to use them). >>> >>>Probably a few other things too. >>> >>>-Richard >> >> >>Please let us know how strong it plays, approximately, compared to other Mac >>programs and to Windows programs on similar-speed hardware. > >My guess is: > >- Chessmaster 9000 is weaker than the imminent HIARCS 9.6 for Mac > release, but significantly stronger than any other Mac chess engine. > >- On similar Windows hardware, several Windows programs are significantly >stronger > than CM9000 - up to 90 points, in the case of Shredder. > >My reasoning: > >It is notoriously difficult to measure the strength of chess programs accurately >without >playing a lot of games between them (which I don't plan on). > >I'm basing my my guess based on this month's SSDF rating list. >Here are the 11 "non-sibling" engines on this months SSDF rating list: > > THE SSDF RATING LIST 2004-12-31 100049 games played by 267 computers > Rating + - Games Won Oppo > ------ --- --- ----- --- ---- >1 Shredder 8.0 CB 256MB Athlon 1200 MHz 2805 23 -22 1075 71% 2645 >2 Deep Fritz 8.0 256MB Athlon 1200 MHz 2791 25 -24 896 72% 2628 >3 Junior 8.0 256MB Athlon 1200 MHz 2772 25 -24 832 66% 2656 >4 Gandalf 6.0 256MB Athlon 1200 MHz 2749 49 -46 225 63% 2658 >5 Hiarcs 9.0 256MB Athlon 1200 MHz 2739 22 -21 1068 62% 2656 >6 Chess Tiger 2004 256MB Athlon 1200 MHz 2723 27 -26 710 59% 2657 >7 Chessmaster 9000 256MB Athlon 1200 MHz 2717 36 -36 376 55% 2679 >8 Rebel 12.0 256MB Athlon 1200 MHz 2684 32 -32 484 57% 2633 >9 Ruffian 1.0.1 256MB Athlon 1200 MHz 2675 27 -27 684 52% 2657 >10 Deep Sjeng 1.5a 256MB Athlon 1200 MH 2673 31 -31 493 52% 2662 >11 Crafty 18.12/CB 256MB Athlon 1200 MHz 2617 27 -27 687 50% 2617 > >Of these 11 engines, four are now (or will soon be) available on the Macintosh: > >5 Hiarcs 9.0 256MB Athlon 1200 MHz 2739 22 -21 1068 62% 2656 >7 Chessmaster 9000 256MB Athlon 1200 MHz 2717 36 -36 376 55% 2679 >9 Ruffian 1.0.1 256MB Athlon 1200 MHz 2675 27 -27 684 52% 2657 >11 Crafty 18.12/CB 256MB Athlon 1200 MHz 2617 27 -27 687 50% 2617 > >Two questions about using these results to predict performance on the Macintosh: >- how similar is the chess engine source code between the Mac and Windows >versions? >- how efficiently does the chess engine object code run on Intel vs. PowerPC >CPUs? > >In the past, I've probed at these questions by running a problem test like the >LCT-II test >on both platforms. If the results are the same, but the timing is different, I >conclude >that the engines are probably the same, but the CPU efficiency is different. > >With respect to the level of similarity: > >HIARCS: I would think HIARCS 9.6 is likely to be somewhat stronger than HIARCS >9.0 on Windows: >- HIARCS 9.6 Macintosh is presumably believed stronger than the earlier version >of HIARCS 9.0 >- In the past, the chess engine source code for HIARCS has been identical >between Mac and Windows >- In the (quite distant) past, older HIARCS engines had better nps/MHz speed on >PowerPC than Intel > >(http://web.archive.org/web/20001118051800/http://gambitsoft.com/chess/games/hiar6me.htm) > >Chessmaster 9000 >- Both Chessmaster 9000 for PC and for Mac use the "The King 3.23" engine. >- I have no benchmarks comparing "The King's" CPU efficiency between Intel and >Mac. > >Ruffian >- The Mac OS/X Version is 2.02, so it may be better than version 1.01 above >- I have no benchmarks comparing Ruffian's CPU efficiency between Intel and Mac. > >Crafty >- I think (though I don't have it) Crafty 18.12 has been compiled for the Mac >- CPU benchmarks have indicated Crafty does better on Intel vs. PowerPC CPUs. > In particular, the spec2000 benchmark 186.crafty runs 30% faster per MHz on > a 2600 MHz AMD Athlon vs. a 2200 MHz PowerPC 970. > >-Richard Richard, I will be glad to send you a binary of Crafty 19.17 optomized for the 970. I can get 3M NPS when run of a duel platform. I runs as a UNIX process (from the terminal). Let me know if you would like me to eMail ti to you. William
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.