Author: Dave Gomboc
Date: 15:24:17 10/04/01
Go up one level in this thread
On October 04, 2001 at 18:18:37, Bruce Moreland wrote: >On October 04, 2001 at 18:08:31, Dave Gomboc wrote: > >>No, that is the point. If your parallel algorithm consistently super-linearly >>outperforms your sequential algorithm, then your sequential algorithm is proven >>to be bad, and you should fix it. > >"Bad" lies on a contiuum, and applies to everything in that sense. > >bruce I'm applying "bad" to a provably inferior algorithm, not everything. You might have thought the older sequential algorithm was good, but your parallel search implementation has shown otherwise, so it's bad now. Yes, "badness" is dynamic. KAISSA and Chess 4.x were world champion programs, but they're bad now, because what's "bad" depends on what you know -- when you learn something better, something that seemed good isn't as good anymore. Dave
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.