Author: Robert Hyatt
Date: 18:20:19 11/01/99
Go up one level in this thread
On November 01, 1999 at 21:08:52, leonid wrote: >On November 01, 1999 at 19:59:43, Peter Kappler wrote: > >>On November 01, 1999 at 19:50:12, leonid wrote: >> >>> >>>Hi! >>>I have the Crafty game in my Hiarcs 7.32 package. When I plyed with Crafty I was >>>impressed and puzzled. Impressed with the game that I found quick and good but >>>doubtful about the central part of the game. In what language it was really >>>written? Core part in Assembler, or 100% on C? >>> >>>Just can't retain me to not mention what really puzzled me in the game. It was >>>one idea that never went away. Why to write so good game in so slow language? To >>>use the C in this case is like to do the premeditated suicide. Like willingly >>>leaving your first place to somebody else. But maybe I rush to the conclusion. >>> >>>Leonid. >> >> >>Where did you get the idea that C is a "slow language"? >> >>Well-written C and a good compiler will produce very fast code. >> >>--Peter > >In despite of everything that could be said about the good compiler, and so >like, Assembler is the speediest one possible. If Crafty, that is already more >that good, could have just one version written in Assembler (if its core part >not done already on Assembler) it could be speeded somewhere between 3 and 8 >times. This way game will start seeing extra ply. More that enough to put it >from already good position to the best one. I can see only the portability the >reason why game was done on C. > >Leonid. Yes you can get another ply if the thing is totally rewritten from scratch around the X86 architecture. But it becomes extremely difficult to modify, and harder to debug after modification. The wins aren't worth the pains...
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.