Author: Robert Hyatt
Date: 06:37:17 04/29/00
Go up one level in this thread
On April 29, 2000 at 05:34:22, Jeremiah Penery wrote: >On April 29, 2000 at 01:02:29, Robert Hyatt wrote: > >>On April 28, 2000 at 23:39:10, Jeremiah Penery wrote: >> >>>On April 28, 2000 at 09:41:29, Chessfun wrote: >>> >>>>I think ponder=on on one cpu just don't work. >>>>IMHO that is :-) >>> >>>As long as both programs are getting 50% of the CPU, it's (roughly) the same as >>>a ponder-on match on 2 machines at half the time control. >> >> >>It is really like a ponder-on match on 2 machines at about 3/4 of the time >>control. Programs will use 50% of their total time properly. They will use >>the other 50% of their total time for pondering, but roughly 50% of that will >>be useful, 50% will be lost due to incorrect pondering. >> >>A game in 1 minute ponder=on is probably pretty close to game in 45 seconds, >>ponder=off. > >True. But I was pointing out that this: > >2 min/game, ponder=ON, 50% CPU (I.e., using one CPU.) > >is approximately the same as this: > >1 min/game, ponder=ON, 100% CPU (I.e., using 2 CPUs.) > >Jeremiah :) That is correct. The original post was discussiong ponder=off at 1 min being the same as ponder=on at 2 mins. That isn't quite right. The difference there is the 3/4 I gave,
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.