Author: Robert Hyatt
Date: 21:53:53 02/20/03
Go up one level in this thread
On February 20, 2003 at 19:09:18, Tom Kerrigan wrote: >It seems you know a lot more than I do about instruction latencies. How do you >explain Crafty's disproportionate speedup on IA-64, then? And why would you >think Crafty's performance is a good predictor of other chess programs on IA-64 >when Crafty is so much different from many other programs? This ought to be obvious. IA32 requires two or more operations to do all the boolean stuff. IA64 requires just one. At the _same_ clock speed, that would be a gain of 2x by itself for all the 64 bit stuff that is done. And 64 bit stuff is _everywhere_. Move generation. Attack detection. Evaluation. SEE. you name it. And it _has_ to speed up. > >Of course, all we have is Hyatt's word that Crafty does well on IA-64, although >he's never seen it in person, and people who have seen Crafty run on IA-64 in >person seem to contradict him... > >-Tom People == person. Only _one_ has reported a bad IA64 NPS for crafty, for reasons unknown. _two_ have reported good NPS values for IA64, Eugene is one. So "people" is wrong as that is plural, and there is definitely only one bad IA64 number out there so far... Or, as usual, I suppose I could be making these numbers up. No 800K on the 600mhz alpha. Mann is nuts. No big nps on Mckinley. Eugene and the folks at Intel are nuts. Or even better, I made it all up on my own, for whatever reason you care to attribute it to. Or it could just be raw, factual data. Nah. Couldn't be that. That would wreck too many theories.
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.