Author: Will Singleton
Date: 00:39:58 01/30/02
Go up one level in this thread
On January 28, 2002 at 17:43:29, Jay Scott wrote: >On January 28, 2002 at 13:17:20, Vincent Diepeveen wrote: > >>Photoshop and such rendering programs can make use of the velocity >>engine in the G4, which is of no use for computerchess > >I disagree. > > If Cray Blitz was able to get benefits from the Cray's long vectors with slow >startup times, then a chess program written for AltiVec should get bigger >benefits from the low-overhead short vectors of AltiVec. The architecture looks >like a good match to me. Hello Jay, If you have any specific ideas, let me know. I plan to test a bitboard implementation, seems like a 128-bit integer vector could possibly be quick. However, the main benefit of the vector (as you know) is in a parallel speedup accomplished by single instructions on multiple data, like adding two 16-byte arrays in the same time as the regular integer unit adds one byte. Has obvious application in video processing, and other tasks where large amounts of data must be converted or processed in repetitive ways. How this can be applied effectively to chess programming I'm not sure. I will try to find out. But if you have ideas, please let me know. Will
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.