Author: Robert Hyatt
Date: 19:56:39 10/04/01
Go up one level in this thread
On October 04, 2001 at 18:36:29, Dave Gomboc wrote: >> >You can always rewrite the sequential algorithm, because you can simulate the >parallel one. Any simulation overhead will be swamped by measurement error. > >Dave The overhead could be almost zero on the right hardware anyway. The old xerox sigma 9 could run two threads with no headache, since it had two completely separate sets of 16 registers... The hardware issue is simply a non-issue when talking about speedups... and super-linear speedups in particular...
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.