Author: William Bryant
Date: 20:16:04 04/19/00
Go up one level in this thread
On April 19, 2000 at 22:59:04, Robert Hyatt wrote: >On April 19, 2000 at 14:44:00, Wayne Lowrance wrote: > >>On April 19, 2000 at 13:57:26, Dan Andersson wrote: >> >>>>I don't think FPU performance is significant for chess analysis, so the Pentium >>>>appears to be more suitable in this respect. >>> >>>It doesn't follow that the athlon has a substandard integer unit. >>>Price/Performance ratings favour the Athlon on almost all counts. >> >>The latest info on Tom's rates only FPU as superior to Pent III coppermine 1 >>giz. I can not intrepet the Integer performance out of his test data. Can you ? >>Thanks. And from the previous thread maybe FPU is _not_ important to chess >>software, I dunno, have to get the advise from Dr Hyatt or others. >>Wayne > > >FP performance is meaningless for chess programs... although for bitmappers, >the mmx instructions would be useful if there was a portable way of accessing >them... or if we could coax the compilers to emit them to do long long bitwise >operators... Knowing that the IBM PPC processors are not your processor of choice, do you know a way to coax the AltiVec processor in the new G4 processors to do long long bitwise operators. I have looked into this a little bit, and although the bus and register width is 128 bits wide, it _appers_ to be designed for parallel operations on 4, 8, 16, 32 and 128 bit data structures. I may be missing something. An thoughts or comments. William wbryant@ix.netcom.com
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.