Author: Slater Wold
Date: 17:32:18 08/20/01
Go up one level in this thread
On August 20, 2001 at 17:36:43, Dann Corbit wrote: >On August 20, 2001 at 17:26:32, Slater Wold wrote: > >>I am at work now, but just looking at the position and your eval, I would almost >>guarantee that on my 2x1.7Ghz, it'll find it, and in a reasonable amount of >>time. What it took your p200 to do in 11 hours, will take my computer 41 >>minutes. (Just a mathematical ESTIMATION.) If it takes 5 hours for yours, >>should take mine, at the MAX, 20 minutes. And I am guessing it won't take half >>that time. >> >>I will post the eval around 6pm central time. > >You also get another interesting effect with multiple CPU's. >Sometimes, it will solve much faster than a single CPU (far more than simple >math would explain). Strangely, it never seems to take perversely longer. I've >yet to figure that one out. I agree. I did a test with Crafty on one of Sune's positions a day or two ago, and found that Crafty on 1 of my CPU's solves the problem in 3:33 minutes, while 2 CPU's take 1:52. It took more nodes, and more ply. But I know exactly what you're talking about, SMP follows no rules of math, that's for sure. Here is my result: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ,S - Shredder 5.32 2b5/7p/P2k4/1P4p1/2Ppp3/5pP1/1N3K1P/8 b - - 0 1 Analysis by Shredder 5.32: 1...Bxa6 2.bxa6 h5 +- (3.23) Depth: 1/2 00:00:00 1...Bxa6 2.bxa6 h5 +- (3.23) Depth: 1/2 00:00:00 1...h5 +- (2.64) Depth: 1/2 00:00:00 1...h5 +- (1.66) Depth: 1/2 00:00:00 1...h5 +- (1.66) Depth: 1/2 00:00:00 1...Kc5 2.h3 h6 3.Ke1 Bxh3 4.a7 ± (0.81) Depth: 1/2 00:00:00 1...Kc5 2.h3 h6 3.Ke1 Bxh3 4.a7 ± (0.81) Depth: 1/2 00:00:00 1...Kc5 2.h3 h6 3.Ke1 Bxh3 4.a7 ² (0.52) Depth: 2/4 00:00:00 1...Kc5 2.h3 h6 3.Ke1 Bxh3 4.a7 = (0.12) Depth: 2/4 00:00:00 1...Kc5 2.Ke1 e3 3.h4 Bh3 4.a7 ³ (-0.28) Depth: 3/6 00:00:00 1...Kc5 2.Ke1 e3 3.h4 Bh3 4.a7 ³ (-0.28) Depth: 3/6 00:00:00 1...Kc5 2.Ke1 e3 3.h4 Bh3 4.a7 ³ (-0.28) Depth: 4/8 00:00:00 1...Kc5 2.Ke1 e3 3.h4 Bh3 4.a7 ³ (-0.28) Depth: 5/10 00:00:00 1...Kc5 2.Ke1 e3 3.h4 gxh4 4.gxh4 Kb6 5.Na4+ Ka7 ³ (-0.28) Depth: 6/12 00:00:00 6kN 1...Kc5 2.Na4+ Kxc4 3.Nb6+ Kb4 4.Nxc8 Ka3 = (-0.03) Depth: 7/14 00:00:00 17kN 1...Kc5 2.Na4+ Kb4 3.Nb6 Bg4 4.Ke1 e3 5.h4 gxh4 ² (0.40) Depth: 7/14 00:00:00 21kN 1...Kc7 2.Ke1 Kb8 3.c5 e3 4.h4 gxh4 5.gxh4 ² (0.39) Depth: 7/14 00:00:00 27kN 1...Kc7 2.Ke1 Kb8 3.h4 e3 4.Nd3 gxh4 5.gxh4 = (0.04) Depth: 7/14 00:00:00 36kN 1...Kc7 2.c5 Kb8 3.h4 Be6 4.b6 Bb3 5.hxg5 = (0.01) Depth: 7/14 00:00:00 47kN 1...g4 2.Na4 Kc7 3.b6+ Kc6 4.b7 Bxb7 5.axb7 Kxb7 = (0.00) Depth: 7/14 00:00:00 76kN 1...g4 2.Kf1 Kc5 3.a7 Bb7 4.h4 ³ (-0.27) Depth: 7/14 00:00:00 82kN 1...g4 2.Kf1 Kc5 3.Na4+ Kxc4 4.Kf2 Kxb5 5.a7 Bb7 6.Nc3+ dxc3 7.Ke3 ³ (-0.52) Depth: 8/16 00:00:00 107kN 1...g4 2.Ke1 f2+ 3.Kxf2 e3+ 4.Ke1 Bxa6 5.bxa6 Ke6 6.Nd3 ³ (-0.52) Depth: 8/16 00:00:00 117kN 1...g4 2.Na4 e3+ 3.Kf1 Kc7 4.c5 Kb8 5.Nb6 Be6 6.Ke1 ³ (-0.39) Depth: 9/18 00:00:00 207kN 1...g4 2.Na4 e3+ 3.Kf1 Kc7 4.c5 Kb8 5.h3 gxh3 6.Ke1 h6 ³ (-0.64) Depth: 10/20 00:00:00 406kN 1...g4 2.Na4 e3+ 3.Kf1 Bf5 4.Nb2 Kc7 5.c5 Kb8 ³ (-0.65) Depth: 10/20 00:00:00 479kN 1...g4 2.Kf1 Kc5 3.Na4+ Kxc4 4.Nb6+ Kd3 5.Nxc8 e3 6.a7 e2+ 7.Ke1 f2+ 8.Kxf2 e1Q+ 9.Kxe1 ³ (-0.40) Depth: 11/22 00:00:01 875kN 1...g4 2.Ke1 Kc5 3.Na4+ Kxc4 4.Nb6+ Kxb5 5.Nxc8 Kxa6 6.Nd6 e3 7.Nf5 Kb5 8.Nxd4+ Kc4 = (-0.25) Depth: 11/22 00:00:01 931kN 1...h5 2.Na4 Bg4 3.c5+ Ke6 4.a7 e3+ 5.Kf1 Bh3+ 6.Ke1 f2+ 7.Kd1 f1Q+ 8.Kc2 ³ (-0.26) Depth: 11/22 00:00:02 1773kN 1...h5 2.h4 gxh4 3.gxh4 Kc5 4.Na4+ Kb4 5.Nb6 e3+ 6.Kxf3 Bg4+ 7.Kg3 d3 ³ (-0.33) Depth: 11/22 00:00:03 2070kN 1...h5 2.h3 g4 3.hxg4 Bxg4 4.Ke1 e3 5.Nd3 Kc7 6.Kd1 ³ (-0.56) Depth: 12/24 00:00:05 3596kN 1...Kc5 2.Na4+ Kxc4 3.Nb6+ Kd3 4.Nxc8 e3+ 5.Kxf3 e2 6.h4 e1Q 7.Nd6 ³ (-0.57) Depth: 12/24 00:00:09 5727kN, tb=1 1...Kc5 2.g4 Kb4 3.h3 h6 4.Ke1 Kb3 5.Na4 Kxa4 6.Kf1 µ (-0.82) Depth: 12/24 00:00:09 6123kN, tb=1 1...Kc5 2.Na4+ Kxc4 3.Nb6+ Kd3 4.g4 Bxg4 5.a7 e3+ 6.Kg3 f2 7.a8Q f1N+ 8.Kxg4 h5+ 9.Kxh5 e2 10.Kxg5 Ne3 ³ (-0.57) Depth: 13/25 00:00:11 7546kN, tb=1 1...Kc5 2.Na4+ Kxc4 3.Nb6+ Kd3 4.g4 e3+ 5.Kxf3 Bxg4+ 6.Kg3 Kc2 7.Nc4 Be6 8.Nd6 ³ (-0.28) Depth: 13/25 00:00:12 7864kN, tb=1 1...h5 2.h3 g4 3.hxg4 Bxg4 4.Ke1 e3 5.Nd3 Kc7 6.Kf1 Bh3+ 7.Kg1 f2+ 8.Nxf2 exf2+ 9.Kxf2 ³ (-0.29) Depth: 13/25 00:00:13 8328kN, tb=1 1...h5 2.h3 g4 3.hxg4 Bxg4 4.Ke1 e3 5.Nd3 Kc7 6.Kd1 ³ (-0.64) Depth: 13/25 00:00:13 8634kN, tb=1 1...h5 2.h3 g4 3.hxg4 Bxg4 4.Kf1 e3 5.Nd3 Bf5 6.c5+ Kd5 7.Nf2 e2+ 8.Ke1 µ (-0.88) Depth: 13/25 00:00:15 9707kN, tb=1 1...h5 2.h4 g4 3.Ke1 Kc5 4.Na4+ Kxc4 5.Nb6+ Kd3 6.Nxc8 e3 7.a7 f2+ 8.Kf1 e2+ 9.Kxf2 e1Q+ ³ (-0.63) Depth: 14/26 00:00:23 14678kN, tb=39 1...h5 2.h4 g4 3.Ke1 Kc5 4.Na4+ Kxc4 5.Nb6+ Kxb5 6.Nxc8 Kxa6 7.Nd6 e3 8.Nf5 Kb5 ³ (-0.36) Depth: 14/26 00:00:23 14902kN, tb=39 1...h5 2.h4 g4 3.Ke1 Kc5 4.Na4+ Kxc4 5.Nb6+ Kxb5 6.Nxc8 Kxa6 7.Nd6 e3 8.Nf5 Kb5 ³ (-0.36) Depth: 15/27 00:00:52 31992kN, tb=217 1...Kc5 2.Na4+ Kxc4 3.Nb6+ Kd3 4.g4 e3+ 5.Kxf3 Bxg4+ 6.Kxg4 e2 7.a7 e1Q 8.a8Q Qe6+ 9.Kxg5 Qg6+ 10.Kf4 Qxb6 11.Kg4 ³ (-0.37) Depth: 15/27 00:01:02 37440kN, tb=282 1...Kc5 2.Na4+ Kxc4 3.Nb6+ Kd3 4.g4 e3+ 5.Kxf3 e2 6.Kf2 Be6 7.Ke1 Ke3 8.Nd5+ Bxd5 9.h4 µ (-0.72) Depth: 15/27 00:01:04 39160kN, tb=292 1...Kc5 2.g4 Bxg4 3.a7 e3+ 4.Kg3 f2 5.a8Q f1Q 6.Qa3+ Kb6 7.Na4+ Ka7 8.b6+ Kb8 9.Qd6+ Ka8 10.Kxg4 Qxc4 µ (-0.72) Depth: 15/27 00:01:18 46770kN, tb=505 1...Kc5 2.g4 Bxg4 3.a7 e3+ 4.Kg3 f2 5.a8Q f1Q 6.Qa3+ Kb6 7.Na4+ Ka7 8.b6+ Kb8 9.Qd6+ Ka8 10.Kxg4 Qd3 11.Kxg5 Qc2 12.Qxd4 ³ (-0.47) Depth: 16/28 00:01:34 55362kN, tb=918 1...Kc5 2.g4 Kb4 3.h3 h6 4.Kg3 Ka5 5.Kf2 Kb4 6.Kg3 = (0.00) Depth: 16/28 00:01:57 68412kN, tb=1178 1...h5 2.h4 g4 3.Ke1 Kc5 4.Na4+ Kxc4 5.Nb6+ Kxb5 6.Kd1 Kxb6 7.a7 Kxa7 = (-0.01) Depth: 16/28 00:02:02 71491kN, tb=1210 1...h5 2.h4 g4 3.Ke1 Kc5 4.Na4+ Kxc4 5.Nb6+ Kxb5 6.Kd1 Kxb6 7.a7 Kxa7 ³ (-0.36) Depth: 16/28 00:02:15 78771kN, tb=1292 1...h5 2.h4 g4 3.Ke1 Kc5 4.Na4+ Kxc4 5.Nb6+ Kxb5 6.Nxc8 e3 7.Ne7 Kxa6 8.Nf5 Kb5 ³ (-0.36) Depth: 16/28 00:02:19 81430kN, tb=1314 1...h5 2.h4 gxh4 3.gxh4 Kc5 4.Na4+ Kxc4 5.Nb6+ Kd3 6.a7 e3+ 7.Kg3 Bb7 8.a8Q Bxa8 9.Nxa8 Ke4 ³ (-0.44) Depth: 17/29 00:04:23 145705kN, tb=3953 1...h5 2.h4 gxh4 3.gxh4 Kc5 4.Na4+ Kxc4 5.Nb6+ Kd3 6.a7 e3+ 7.Kg3 Bb7 8.a8Q Bxa8 ³ (-0.69) Depth: 18/30 00:07:11 237810kN, tb=11749 1...h5 2.h4 gxh4 3.gxh4 Kc5 4.Na4+ Kxc4 5.Nb6+ Kd3 6.a7 e3+ 7.Kg3 Bb7 8.a8Q Bxa8 9.Nd5 µ (-1.19) Depth: 18/30 00:08:21 278230kN, tb=14779 1...h5 2.h4 gxh4 3.gxh4 Kc5 4.Kg3 Bg4 5.a7 Kb4 6.a8Q f2 7.Kxf2 e3+ 8.Ke1 Bh3 9.Na4 Kxc4 µ (-1.19) Depth: 18/30 00:09:18 310567kN, tb=16077 1...g4 2.Ke1 Kc5 3.Na4+ Kb4 4.Nb6 e3 5.Nxc8 d3 6.Nb6 Kc3 7.Nd5+ Kd4 8.Kd1 d2 µ (-1.20) Depth: 18/30 00:22:32 775459kN, tb=19012 1...g4 2.Ke1 Kc5 3.Na4+ Kb4 4.Nb2 e3 5.c5 Kxc5 6.Na4+ Kxb5 7.a7 f2+ 8.Kf1 Bb7 9.h3 -+ (-1.54) Depth: 18/30 00:23:29 809434kN, tb=19448 1...g4 2.Ke1 Kc5 3.Na4+ Kb4 4.Nb2 e3 5.c5 Kxc5 6.Na4+ Kxb5 7.a7 f2+ 8.Kf1 Bb7 9.h3 -+ (-1.79) Depth: 19/31 00:27:38 952873kN, tb=23785 1...g4 2.Ke1 Kc5 3.Na4+ Kb4 4.Nb2 e3 5.c5 Kxc5 6.Na4+ Kxb5 7.a7 Bb7 8.Nb2 -+ (-2.29) Depth: 19/31 00:28:59 1001172kN, tb=24311 1...g4 2.Ke1 Kc5 3.Na4+ Kb4 4.a7 Bb7 5.Nb6 d3 6.Nd5+ Kxc4 7.Ne3+ Kxb5 8.Nxg4 Kb6 9.Kd2 h5 -+ (-3.43) Depth: 19/31 00:37:25 1286120kN, tb=31943 (, 20.08.2001) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ It took almost the exact same route as Uri's. However, mine of course, did it a lot faster. Also, I had TB's, and he did not. Not sure what kind of effect that would have on it. I cannot say for sure whether or not Shredder would find this in a tournament game, soley based on the fact that while playing 120/40 games I've seen Shredder take 20-30 minutes for 1 move. Also, there is always the fact of pondering in a tournament game. Perhaps DS would have seen this move in 1 second, had it stored this move in hash a few moves back. Hard to say, but it for sure takes a while for it to find it. And I too have ALWAYS regarded Shredder as have 10x any other programs endgame. But all programs have their problems, and perhaps this is just a problem Shredder has, for whatever reason. BTW, this was run on a 2x1.7Ghz with 128MB hash. Also, all 3/4/5 piece TB's. Slate
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.