Author: Bruce Moreland
Date: 14:58:04 01/26/99
Go up one level in this thread
On January 26, 1999 at 16:23:50, Dann Corbit wrote: >I missed some of the earlier posts. > >What is the time control for these tests? > >What is the hardware used for these tests? > >How does the time control used compare with the time control used for the >original matches in question? > >How does the memory & CPU usage compare with the original equipment? > >I think that these are key questions. The original game involved beefy hardware at a long time control. Most of us don't have hardware that strong. Also, the amount of time taken per move wasn't known to me at the time. It may be possible to figure it out now, but if we're getting down to that level of detail there will be permanent problems, since basically we'd have to do it during the right phase of the moon, etc. So to be practical I am using a much shorter time control, but one that still lets the programs search as long as I think is practical. You pick the program. Run it for a minute per position. A match is a match if the move matches at the end of one minute. Report program, version, hardware, and hash memory, and how many from each suite your program matched, and I'll write it up. I posted the EPD's, if someone wants them emailed I am at brucemo@seanet.com, and I'll have them to you quickly. If someone wants to run this at more than a minute per position, I'll write that up and report it seperately. If someone wants to break their results up into one-second bins, I'll report that as well. I don't care, my goal is to get a lot of data and report it in some sort of organized fashion. The test as I've designed it takes about 14 hours to run. bruce
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.