Author: Andrew Wagner
Date: 17:56:09 06/28/04
Go up one level in this thread
Hi Anthony, I'm curious as to whether there's a particular reason you're not requesting tests of games between Crafty-Regular and Crafty-autotuned. I know Bob has stated he doesn't like to tune crafty to play against itself, but for testing purposes, to see if the auto-tuned version is better, wouldn't it make sense? Just wondering. I'm going to see if I can get my P4 2.4 gHz, 512 MB box up and running this weekend. If so, I can give you unlimited processor time on that to run some tests. Andrew On June 28, 2004 at 08:54:00, Anthony Cozzie wrote: >Most of you probably know by now that my "secret project" with Bob has been >evaluation autotuning. I feel that with the improvements in parallel search and >CPU speed, the evaluation is beginning to be become more important than the >search in top programs (e.g. Shredder). As the number of terms increases, the >number of relations increases quadratically and the problem becomes harder. I >don't think that autotuning will ever be as good as hand tuning, but it is a >good way to get a first guess for the value of a parameter, as well as seeing >whether a new parameter is worth anything. This weekend, aside from watching >massive amounts of "Hajime no Ippo", I have finished a run with the latest >tweaks to the algorithm. > >The basic idea (which actually dates to my senior year in college, when I took >CAD tools, 18760) is to use simulated annealing. Simulated annealing makes >random changes and accepts them probablistically: > >(V1 > V0) : accept change >(V1 < V0) : accept if rand() > exp((V1 - V0)/T) > >Where T is the temperature, and it gradually gets smaller during the anneal. >This has the effect of ignoring small changes initially, and concentrating on >the big picture. This is the best link I can find in a few minutes of google >search; I will explain more if people are still unclear: > >http://members.aol.com/btluke/simann1.htm > >I am looking for some testers to run games with the latest crafty against >various engines. Basically, run N games against challenger X with the standard >settings, and then N games with the new settings. I am only really interested >in longer timecontrols: 20 min + on an Athlon 2.0G or so (70 min on P-650, etc), >and at least 10 games. If this is ever published, I will include everyone who >runs some test games in the acknowledgements section. > >anthony > >The crafty evaluation script: > >evaluation 1 100 >evaluation 2 300 >evaluation 3 300 >evaluation 4 500 >evaluation 5 900 >evaluation 11 100 >evaluation 12 0 >evaluation 13 100 >evaluation 14 100 >evaluation 15 100 >evaluation 16 100 >evaluation 17 120 >evaluation 21 10 >evaluation 22 3 >evaluation 23 16 >evaluation 24 1 >evaluation 25 50 >evaluation 26 8 >evaluation 27 13 >evaluation 28 525 >evaluation 29 150 >evaluation 30 50 >evaluation 31 100 >evaluation 32 10 10 10 10 10 10 10 10 9 22 -2 -11 21 50 15 10 13 20 16 17 26 21 >25 16 7 4 3 11 12 16 10 6 -1 1 7 11 7 13 8 0 -1 -1 4 9 4 6 5 0 -3 -5 -5 -7 -9 4 >4 -2 -3 -3 -3 -3 -3 -3 -3 -3 >evaluation 33 0 2 2 2 23 43 212 212 >evaluation 34 0 0 0 0 0 0 7 7 >evaluation 35 0 1 2 12 23 53 117 117 >evaluation 36 0 6 6 10 13 22 43 43 >evaluation 37 0 5 15 36 49 55 63 63 63 >evaluation 38 0 9 28 44 68 89 89 89 89 >evaluation 39 0 3 9 9 9 9 9 >evaluation 40 0 0 6 24 40 64 80 94 96 >evaluation 41 0 0 0 0 12 60 100 0 >evaluation 42 45 41 20 19 18 13 13 9 9 9 9 9 6 6 6 6 5 4 3 3 3 3 3 3 3 3 3 3 3 >3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 >3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 >3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 0 >evaluation 43 72 36 36 36 34 31 30 30 27 27 24 24 21 21 18 18 15 15 12 12 10 10 >10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 >10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 >10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 >10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 >evaluation 51 6 6 4 3 1 0 0 0 >evaluation 52 -32 -16 -8 -8 -8 -8 -16 -32 -24 -24 4 4 4 4 -24 -24 -8 2 6 6 6 6 >2 -8 -8 2 6 6 6 6 2 -8 -8 2 4 4 4 4 2 -8 -8 2 2 2 2 2 2 -8 -8 -8 0 0 0 0 -8 -8 >-8 -8 -8 -8 -8 -8 -8 -8 >evaluation 53 0 0 0 0 0 0 0 0 0 0 6 12 12 6 0 0 0 0 12 18 18 12 0 0 0 0 12 15 >15 12 0 0 0 0 0 6 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >evaluation 61 10 >evaluation 62 2 >evaluation 63 12 >evaluation 64 4 >evaluation 65 175 >evaluation 66 -12 -18 6 19 18 22 11 -2 -13 >evaluation 67 7 7 6 3 1 0 0 0 >evaluation 68 -16 -16 0 0 0 0 -16 -16 -16 2 4 4 4 4 2 -16 -4 2 6 6 6 6 2 -4 -4 >2 6 6 6 6 2 -4 -4 2 4 4 4 4 2 -4 -4 2 2 2 2 2 2 -4 -16 2 0 0 0 0 2 -16 -16 -16 >-8 -8 -8 -8 -16 -16 >evaluation 71 16 >evaluation 72 3 >evaluation 73 4 >evaluation 74 9 >evaluation 75 17 >evaluation 76 10 >evaluation 77 20 >evaluation 78 23 >evaluation 79 10 >evaluation 80 9 9 6 4 2 0 0 0 >evaluation 81 6 6 2 1 1 0 0 0 >evaluation 82 0 0 4 6 6 4 0 0 0 0 4 6 6 4 0 0 0 0 4 6 6 4 0 0 0 0 4 6 6 4 0 0 0 >0 4 6 6 4 0 0 0 0 4 6 6 4 0 0 0 0 4 6 6 4 0 0 0 0 4 6 6 4 0 0 >evaluation 91 22 >evaluation 92 6 >evaluation 93 50 >evaluation 94 6 >evaluation 95 8 >evaluation 96 10 10 10 4 0 0 0 0 >evaluation 97 9 9 5 0 0 0 0 0 >evaluation 98 0 0 0 0 0 0 0 0 -15 0 4 5 5 4 0 -15 0 2 4 10 10 4 2 0 0 2 10 12 >12 10 2 0 -10 2 10 12 12 10 2 -10 -10 -10 4 10 10 4 -10 -10 -10 2 8 8 8 8 2 -10 >-10 -8 0 0 0 0 -8 -10 >evaluation 101 15 >evaluation 102 96 >evaluation 103 600 >evaluation 104 7 >evaluation 105 -86 >evaluation 106 3 >evaluation 107 0 0 0 0 0 0 0 0 1 1 1 2 2 4 4 4 4 7 11 13 13 18 20 20 27 27 30 37 >37 37 37 37 37 43 43 43 47 47 47 47 53 54 57 59 61 61 61 61 61 61 64 65 65 65 66 >68 69 72 72 75 80 80 87 87 97 98 99 108 154 165 184 217 245 249 286 397 426 553 >689 799 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 >922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 922 >922 922 922 922 922 922 922 922 922 922 >evaluation 108 8 8 8 8 8 8 8 8 7 7 7 7 7 7 7 7 6 6 6 6 6 6 6 6 5 5 5 5 5 5 5 5 4 >4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 1 1 2 3 3 2 1 1 1 0 2 3 3 2 0 1 >evaluation 109 -50 -30 -30 -30 -30 -30 -30 -50 -30 -30 0 0 0 0 -30 -30 -30 -10 >30 40 40 30 -10 -30 -30 -10 30 40 40 30 -10 -30 -30 -10 20 30 30 20 -10 -30 -30 >-10 0 10 10 0 -10 -30 -30 -20 -10 0 0 -10 -20 -30 -50 -40 -30 -20 -20 -30 -40 >-50 >evaluation 110 -90 -70 -50 -30 -30 -30 -30 -30 -70 -50 -30 -30 0 0 0 -10 -70 -50 >-30 -10 30 30 30 0 -70 -50 -30 -10 30 30 30 0 -70 -50 -30 -10 30 30 30 0 -70 -50 >-30 -10 20 20 20 0 -70 -50 -30 -20 -10 -10 -10 -10 -90 -70 -50 -40 -30 -20 -20 >-30 >evaluation 111 -30 -30 -30 -30 -30 -50 -70 -90 -10 0 0 0 -20 -30 -50 -70 0 30 30 >30 -10 -30 -50 -70 0 30 30 30 -10 -30 -50 -70 0 30 30 30 -10 -30 -50 -70 0 20 20 >20 -10 -30 -50 -70 -10 -10 -10 -10 -20 -30 -50 -70 -30 -20 -20 -30 -40 -50 -70 >-90 >evaluation 112 0 3 15 16 >evaluation 113 0 12 13 16 >evaluation 114 0 4 13 20 21 25 31 35 35 39 40 40 40 48 52 56 56 56 56 56 56 56 >56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 >56 56 56 56 56 56 56 56 56 56 56 56 56 56 56 >evaluation 115 16 16 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 >26 26 27 27 28 28 29 29 30 30 31 31 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 >32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 >evaluation 121 10 >evaluation 122 7 >evaluation 123 12 >evaluation 124 20 >evaluation 125 2 >evaluation 126 3 >evaluation 127 20
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.