Author: Ernst Walet
Date: 08:52:52 03/03/01
Go up one level in this thread
On March 03, 2001 at 10:08:52, Enrique Irazoqui wrote: >On March 03, 2001 at 09:21:57, Ernst Walet wrote: > >>On March 03, 2001 at 09:11:23, Ernst Walet wrote: >>> >>>I don't understand, with me the ply depth was 18 with 128MB hash. I'll try it >>>again, maybe somthing went wrong. >>> >>>Ernst. >>> >>> >> >>I tried it again, and no f4 at ply 16, but at ply 18 again. Beats me. >>For the record, I used a hashtable of 128MB. >> >>Ernst. > >There is nothing wrong with your setup. What happens is that Deep Fritz will >find f4 at a different depth when changing the hash size. With 256MB hash >running on 1 P933 it finds f4 at ply 18, like on your computer with 128Mb. With >184MB hash, DF sees it at ply 16. In other positions it may be the other way >around. Not so uncommon. > >8/p1p2pp1/1pP3p1/3Pb2k/4P3/5P1P/8/3K2B1 w - - 0 1 > >Analysis by DEEP FRITZ : > >2.Be3 g5 3.Kd2 a5 > ³ (-0.47) Depth: 4/7 00:00:00 >2.Be3 g5 3.Ke2 a5 4.Kd3 > ³ (-0.47) Depth: 5/9 00:00:00 1kN >2.Be3 g5 3.Ke2 a5 4.Kd3 > ³ (-0.47) Depth: 5/9 00:00:00 1kN >2.Be3 g5 3.Ke2 a5 4.Kd3 > ³ (-0.47) Depth: 5/9 00:00:00 1kN >2.Be3 g5 3.Ke2 a5 4.Kd3 > ³ (-0.47) Depth: 5/9 00:00:00 1kN >2.Be3 g5 3.Ke2 a5 4.Kd3 > ³ (-0.47) Depth: 5/9 00:00:00 1kN >2.Be3 g5 3.Ke2 a5 4.Kd3 > ³ (-0.47) Depth: 5/9 00:00:00 1kN >2.Be3 g5 3.Ke2 a5 4.Kd3 > ³ (-0.47) Depth: 5/9 00:00:00 1kN >2.Be3 g5 3.Kc2 a5 4.Bd2 a4 > µ (-0.72) Depth: 6/9 00:00:00 2kN >2.Be3 g5 3.Kc2 a5 4.Bd2 a4 > µ (-0.72) Depth: 6/9 00:00:00 2kN >2.Be3 g5 3.Kc2 a5 4.Bd2 a4 > µ (-0.72) Depth: 6/9 00:00:00 2kN >2.Be3 g5 3.Kc2 a5 4.Bd2 a4 5.Bb4 > ³ (-0.69) Depth: 7/13 00:00:00 7kN >2.Be3 g5 3.Kc2 a5 4.Kb3 b5 5.Ka2 Kh4 > µ (-0.78) Depth: 8/14 00:00:00 14kN >2.Be3 g5 3.Kc2 a5 4.Kc1 a4 5.Kb1 b5 6.Ka2 > µ (-0.81) Depth: 9/14 00:00:00 36kN >2.Ke2! > µ (-0.78) Depth: 9/17 00:00:00 74kN >2.Ke2 Kh4 3.Be3 Bd6 4.Bd4 f6 5.Ke3 Kxh3 6.f4 a5 > µ (-0.78) Depth: 10/17 00:00:00 124kN >2.Be3! > µ (-0.75) Depth: 10/17 00:00:00 143kN >2.Be3 g5 3.Bf2 a5 4.Kc2 a4 5.Kb1 b5 6.Bc5 > µ (-0.91) Depth: 11/19 00:00:00 284kN >2.Bf2! > µ (-0.87) Depth: 11/20 00:00:00 379kN >2.Bf2 a5 3.Kc2 a4 4.Be1 a3 5.Kb3 Bd6 6.Bc3 Kh4 7.e5 Bc5 > ³ (-0.66) Depth: 12/21 00:00:01 744kN >2.Be3! > ³ (-0.62) Depth: 12/21 00:00:01 1016kN >2.Be3 g5 3.Bf2 a5 4.Kc2 a4 5.Be1 Bd6 6.Bc3 f6 7.Be1 a3 > µ (-0.72) Depth: 13/22 00:00:02 1872kN >2.Be3 f6 3.f4 Bd6 4.Ke2 a5 5.Kf3 a4 6.Bd4 g5 7.e5 Bc5 > µ (-0.72) Depth: 14/25 00:00:06 4670kN >2.Be3 g5 3.Bc1 Bd6 4.e5 Bxe5 5.Ba3 Kh4 6.d6 g6 7.d7 Bf6 > ³ (-0.56) Depth: 15/28 00:00:11 7716kN >2.Be3 g5 3.Bc1 Bd6 4.e5 Bxe5 5.Ba3 Bf6 6.d6 Bd8 7.d7 Kh4 > ³ (-0.62) Depth: 16/28 00:00:18 12863kN >2.Be3 g5 3.Bc1 Bd6 4.e5 Bxe5 5.Ba3 Bf6 6.Bf8 Bd8 7.d6 cxd6 > ³ (-0.62) Depth: 17/32 00:00:32 22662kN >2.Bf2! > ³ (-0.59) Depth: 17/33 00:01:06 45660kN >2.Bf2 a5 3.Be1 f6 4.Bc3 Bd6 5.f4 Bxf4 6.e5 fxe5 7.Bd2 Bg3 > ³ (-0.56) Depth: 18/33 00:01:49 75635kN >2.f4! > ³ (-0.53) Depth: 18/36 00:02:54 121186kN, tb=1 >2.f4! Bxf4 3.Bd4 f6 4.e5 fxe5 5.Be3 Bg3 6.Bg5 e4 7.Be7 g5 > ³ (-0.47) Depth: 18/36 00:02:57 123879kN, tb=1 >2.f4! > = (0.00) Depth: 19/37 00:03:14 135836kN, tb=2 >2.f4 Bxf4 3.Bd4 f6 4.e5 fxe5 5.Be3 Bg3 6.Bf2 Bf4 7.Be3 Bg3 > = (0.00) Depth: 20/41 00:04:05 173125kN, tb=6 >2.f4 Bxf4 3.Bd4 f6 4.e5 fxe5 5.Be3 Bg3 6.Bf2 Bf4 7.Be3 Bg3 > = (0.00) Depth: 21/40 00:05:55 253403kN, tb=14 > >(Irazoqui, Cadaqués 03.03.2001) Hi Enrique, Sure I'm familiar with hash table side effects, but in message www.icdchess.com/forums/1/message.shtml?156940 Sune reported that QUOTE I can add here that on my PIII 800, 128 MB hash used, Deep Fritz finds 1.f4 in 73 seconds - with a draw score. Sune UNQUOTE and in message www.icdchess.com/forums/1/message.shtml?156952 QUOTE The ply depth was 16. UNQUOTE This is what I cannot combine because with a hash size of 128MB it takes 18 plys to find f4 with me, and surely with you as well. I think that accidentally he was using a larger hash size when he reported his findings above, like in message www.icdchess.com/forums/1/message.shtml?156957 Ernst.
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.