Author: Jeremiah Penery
Date: 17:09:09 09/13/03
Go up one level in this thread
On September 13, 2003 at 19:02:00, Sune Fischer wrote: >On September 13, 2003 at 18:50:31, Jeremiah Penery wrote: > >>Can't you just calculate the speedup (in time to solution, or whatever measure >>you want) and figure effective NPS from that? > >Well, you would need to run a serial search to compare with, so you won't get >on-the-fly nps. That's true. >Still, you'd have to be careful in how you compute it if you want to compare >with a serial search. >You might be tempted to use the ratio of times to solution, but this wouldn't be >the nps I'm interested in, move ordering plays a role here. >Ie. on a dual you could theoreticly see greater than two speedups, but you >should _never_ see greater than two cpu-loads / nps, simply theoreticly >impossible. > >So depends what kind of "efficiency" you want, and I want the effective nps. :) If you want to calculate the ratio of effective nodes/time, I'm not really sure how you plan to do it. How do you define an effective node, first of all? The serial search duplicates nodes often enough on its own that I don't see a way to calculate a difference between wasted serial nodes and wasted parallel nodes. Or am I just completely off-base here?
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.