Author: Robert Hyatt
Date: 13:30:10 09/03/02
Go up one level in this thread
On September 03, 2002 at 15:28:03, Eugene Nalimov wrote: >Probably I don't understand something, but for me it looks that Crafty copies >much less than 44k when splitting a tree. > >Yes, sizeof(TREE) == 44k. But if you'll look at the function CopyToSMP() in file >utility.c and count size of the copied data it will be something like 3k. > >So, here you miscalculated 15x. I did not read rest of your calculations. You >are 15x wrong in the first one, why the next ones should be more credible? We have been around this bush before. it started with Vincent telling GCP to "copy crafty's efficient way of splitting". I told vincent that I _could_ copy up to xxK (I don't remember xx but I got it the same way you apparently did, by printing sizeof(TREE).) He then got on the opposite horse. You can't possibly produce speedups of 1.7 using _that_ kind of copying. Even though I don't copy it all. Of course he is also on the bandwagon that "smp_lock" is killing my performance, even after I told him I only use it 3-4 thousand times per 3 minute search. But that really doesn't matter. :) > >Once again: I'd recommend you to replace "it's absolutely impossible" by "for me >it's absolutely impossible" in your posts. > >BTW, I mentioned your so-called "problem" with threads and global variables to >Andrew Kadatch. In less than 10 seconds he gave *second* solution to the >"problem", that differs from one I suggested. :-) > >Thanks, >Eugene That is impossible. :) We have already been told that "there is no solution, threads are bad". Even though they work just fine for me. Of course, I'm the one faking the data so I am really running slower with 4 cpus than with one, but am just too damned stubborn to admit it and use just one all the time. :)
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.