Author: Robert Hyatt
Date: 07:23:06 01/28/04
Go up one level in this thread
On January 27, 2004 at 22:44:21, enrico carrisco wrote: >On January 26, 2004 at 21:46:23, Robert Hyatt wrote: > >>On January 26, 2004 at 17:53:15, enrico carrisco wrote: >> >>>On January 26, 2004 at 16:06:13, Robert Hyatt wrote: >>> >>>>On January 26, 2004 at 15:59:59, Rolf Tueschen wrote: >>>> >>>>>On January 26, 2004 at 15:47:56, Robert Hyatt wrote: >>>>> >>>>>>This is probably the best result I have ever seen from Crafty. I am simply >>>>>>running some "sanity checks" to make sure everything is functioning, after a few >>>>>>recent changes. Here is my first test, WAC at 1 second per move: >>>>>> >>>>>>total positions searched.......... 300 >>>>>>number right...................... 299 >>>>>>number wrong...................... 1 >>>>>>percentage right.................. 99 >>>>>>percentage wrong.................. 0 >>>>>>total nodes searched.............. 109706792 >>>>>>average search depth.............. 4.6 >>>>>>nodes per second.................. 6522401 >>>>>> >>>>>>It misses #230. >>>>>> >>>>>>I re-compiled with Dann's favorite -DDETECTDRAW and at one second per move, >>>>>>blasted out the following result: >>>>>> >>>>>>total positions searched.......... 300 >>>>>>number right...................... 300 >>>>>>number wrong...................... 0 >>>>>>percentage right.................. 100 >>>>>>percentage wrong.................. 0 >>>>>>total nodes searched.............. 103251953 >>>>>>average search depth.............. 4.5 >>>>>>nodes per second.................. 5713998 >>>>>> >>>>>>I will try to run some other tests as I continue looking at things. I am also >>>>>>once again experimenting with the -DFUTILITY stuff that Jeramiah wrote, to see >>>>>>how it affects things on this really monster box. >>>>>> >>>>>>BTW for those that have asked, "crafty" is playing on ICC most of this week >>>>>>using this box. Scrappy is running on my normal dual xeon 2.8 all the time... >>>>> >>>>> >>>>> >>>>>I love you as a scientist, buddy, but how will you do as a participant at the >>>>>CTG6? You are doing a good job in training your own opponents IMO. But this is >>>>>the most decent I've seen for years here around! I wish that Crafty still takes >>>>>the event in passing by. You deserved it. Fingers crossed. :) >>>>> >>>>>Rolf >>>> >>>> >>>>You mean the actual event? >>>> >>>>Hint: Crafty will _not_ be using the same book I am playing with on ICC right >>>>now. :) That's all I'll say. Except that it will not be using a commercial >>>>book either. :) >>> >>>You'll need to send that book over for inspection immediately. >>>My email address is.... ;) >>> >>>-elc. >> >> >>I'll mail it to you Sunday night. :) >> >>My old Cray Blitz book partner is coming down Wednesday for a day or two of book >>tuning. Like the good old days. :) > >If I didn't know better I'd say someone wants to win the tournament. This >appears to be more effort than the past CCTs (4 & 5), no? > >-elc. No. I do some book preparation every time. Remember that I won the first one and finished in a 3-way tie for first in the last one... I take all such events seriously, but I'm hardly paranoid. I casually asked AMD about the opteron I was using to help them evaluate the "scaling" issue for Crafty on their NUMA boxes, and they said "probably not" and then with no further prompting said "here's a box to use..." I don't lose sleep over these events, but do try to be reasonably well prepared. I called bert and he was interested in coming down as we used to do this for weeks in the old CB days. :)
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.