Author: Tord Romstad
Date: 14:03:48 04/27/05
Go up one level in this thread
On April 27, 2005 at 16:39:57, Steven Edwards wrote: >On April 27, 2005 at 12:11:33, Matthew Hull wrote: >>On April 27, 2005 at 12:04:58, Tord Romstad wrote: >>>On April 27, 2005 at 11:29:07, John Sidles wrote: > >>>The PowerMac G5 is an awesome machine, but if the "ultimate desktop chess >>>computer" is what you are looking for, you should look elsewhere. >> >>Unless someone could figure out how to use the vector processor for chess >>knowledge. Bob did that on Cray machines, but the "altivec" type processor is >>probably much more primitive and less flixible than the Cray. > >I spent some time with this a while back. Unfortunately, AltiVec (a >coprocessor, really) doesn't have any instructions that are of interest to a >typical chess program. I arrived at the same conclusion the last time I looked at AltiVec. Perhaps it would be possible to make a very fast program if you design it from the ground up with AltiVec in mind, but it wouldn't be easy, and of course such a program would be difficult to port. I have read that gcc 4.0 is supposed to support "auto-vectorization" on the G4 and G5, but it remains to be seen how well this works in practise. I would be extremely impressed if it actually gives a noticable performance boost for chess programs. Tord
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.