Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Chess program improvement project (copy at Winboard::Programming)

Author: Robert Hyatt

Date: 18:49:28 03/07/06

Go up one level in this thread


On March 07, 2006 at 19:16:31, Stuart Cracraft wrote:

>On March 07, 2006 at 15:47:26, Robert Hyatt wrote:
>
>>On March 07, 2006 at 14:13:48, Uri Blass wrote:
>>
>>>On March 07, 2006 at 10:39:40, Will Singleton wrote:
>>>
>>>>On March 07, 2006 at 03:33:12, Richard Pijl wrote:
>>>>
>>>>>On March 06, 2006 at 22:53:05, Dann Corbit wrote:
>>>>>
>>>>>>Lot's of programs have a problem with this one.  It's a sacrifice leading to a
>>>>>>forced mate.  I would not worry about this one too much:
>>>>>>[D] 4r1k1/p1qr1p2/2pb1Bp1/1p5p/3P1n1R/1B3P2/PP3PK1/2Q4R w - - bm Qxf4
>>>>>
>>>>>Single reply extensions, mate threat extensions and checks in quiescense search
>>>>>all help in finding this one quickly. Down-side: When used unrestricted,
>>>>>searches will explode. I agree here with Dann, don't worry about this one.
>>>>>The Baron finds the move in 7.6 seconds on a Xeon 2.8, reporting the mate in 48
>>>>>seconds.
>>>>>
>>>>> ply       time        nodes   score  pv
>>>>>  2( 3)-   0:00.00       285   -1.23  Kg2-g1 (Re8-e2)
>>>>>  2( 3)&   0:00.00       383   -1.70  Kg2-f1 Kg8-f8
>>>>>  2( 3).   0:00.00       414   -1.70  Kg2-f1 Kg8-f8
>>>>>  3( 5)&   0:00.00       757   -1.44  Kg2-f1 Kg8-f8 Qc1-d2
>>>>>  3( 5).   0:00.00       903   -1.44  Kg2-f1 Kg8-f8 Qc1-d2
>>>>>  4( 7)&   0:00.02      2352   -1.58  Kg2-f1 Kg8-f8 Bf6-g5 Nf4-e2
>>>>>  4( 7).   0:00.02      2880   -1.58  Kg2-f1 Kg8-f8 Bf6-g5 Nf4-e2
>>>>>  5( 9)&   0:00.09     14038   -1.70  Kg2-f1 Re8-e2 Kf1-g1 Re2xf2 Qc1-e1
>>>>>  5( 9).   0:00.09     15667   -1.70  Kg2-f1 Re8-e2 Kf1-g1 Re2xf2 Qc1-e1
>>>>>  6(12)&   0:00.28     44250   -1.81  Kg2-f1 Re8-e2 Kf1-g1 Re2xf2 Qc1-e3
>>>>>                                      Rf2-g2 Kg1-f1 Kg8-h7
>>>>>  6(12).   0:00.30     48368   -1.81  Kg2-f1 Re8-e2 Kf1-g1 Re2xf2 Qc1-e3
>>>>>                                      Rf2-g2 Kg1-f1 Kg8-h7
>>>>>  7(13)&   0:00.97    153287   -1.93  Kg2-f1 Re8-e2 Rh1-h2 Kg8-f8 a2-a4
>>>>>                                      b5xa4 Bb3xa4 Nf4-d3
>>>>>  7(13).   0:01.45    232925   -1.93  Kg2-f1 Re8-e2 Rh1-h2 Kg8-f8 a2-a4
>>>>>                                      b5xa4 Bb3xa4 Nf4-d3
>>>>>  8(16)+   0:04.75    784071   -1.58  Kg2-f1 (Re8-e2) (Rh4xf4) (Bd6xf4) (Qc1-b1)
>>>>>
>>>>>                                      (Rd7-e7) (Qb1xg6) (Kg8-f8) (Bf6xe7)
>>>>>(Qc7xe7) (Qg6xf7)
>>>>>  8(16)&   0:07.05   1186288   -1.03  Kg2-f1 Rd7-e7 Bf6xe7 Re8xe7 Rh1-g1
>>>>>                                      Kg8-f8 Rg1-g5 f7-f6 Rg5-g3
>>>>>  8(16)+   0:07.61   1302604   -0.33  Qc1xf4 (Bd6xf4) (Rh4xh5) (g6xh5) (Rh1xh5)
>>>>>                                      (Bf4-h6) (Rh5xh6) (Kg8-f8) (Rh6-h8)
>>>>>  8(17)&   0:47.89  10652406  MATE6  Qc1xf4 Bd6xf4 Rh4xh5 g6xh5 Rh1xh5
>>>>>                                      Bf4-h6 Rh5xh6 Qc7-h2 Kg2xh2 b5-b4 Rh6-h8
>>>>>  8(17).   0:48.12  10697945  MATE6  Qc1xf4 Bd6xf4 Rh4xh5 g6xh5 Rh1xh5
>>>>>                                      Bf4-h6 Rh5xh6 Qc7-h2 Kg2xh2 b5-b4 Rh6-h8
>>>>
>>>>Based on the performance of several progs, it's possible that this is a position
>>>>you should worry about.  Most find the mate in 5 secs or less.
>>>>
>>>>Will
>>>
>>>I think that finding the mate fast is not very important.
>>>It is more important to find the move fast but it is also not the most important
>>>thing.
>>>
>>>public Movei is even slower than the baron in finding the move.
>>>Movei can find it relatively faster if I remove pruning based on evaluation but
>>>removing pruning based on evaluation is going to make it play weaker.
>>>
>>>Latest movei fail high on the move faster because I improved my condition of
>>>pruning based on evaluation but it still cannot find the mate fast and only find
>>>high score with no mate first.
>>>
>>>Uri
>>
>>
>>Latest Crafty looks like this:
>>
>>starting thread 1
>>                8     0.19     +1   1. Kf1!!
>>                8     0.80  -1.51   1. Kf1 Rc8 2. Rg1 a5 3. Bg5 Nd5 4.
>>                                    Kg2 Re8 5. Bxd5 cxd5
>>                8->   0.82  -1.51   1. Kf1 Rc8 2. Rg1 a5 3. Bg5 Nd5 4.
>>                                    Kg2 Re8 5. Bxd5 cxd5
>>                9     1.55  -1.22   1. Kf1 Ra8 2. Bc2 Rc8 3. Be4 Nd5 4.
>>                                    Bxd5 cxd5 5. Qh6
>>                9     1.83     +1   1. Qxf4!!
>>                9     1.89     +3   1. Qxf4!!
>>                9     2.02     +M   1. Qxf4!!
>>                9     9.87  Mat06   1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6
>>                                    4. Rxh6 Qh2+ 5. Kxh2 Kf8 6. Rh8#
>>                9->   9.88  Mat06   1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6
>>                                    4. Rxh6 Qh2+ 5. Kxh2 Kf8 6. Rh8#
>>               10    10.18  Mat06   1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6
>>                                    4. Rxh6 Qh2+ 5. Kxh2 Kf8 6. Rh8#
>>               10->  10.20  Mat06   1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6
>>                                    4. Rxh6 Qh2+ 5. Kxh2 Kf8 6. Rh8#
>>
>>It sometimes does not find the mate at depth 9 due to late move reductions
>>interfering somewhere along the way.  but it always finds the move Qxf4 very
>>quickly, and always has a score of +10 or better even when it fails to find the
>>mate, which is enough for me...
>>
>>The above on my dual xeon 2.8ghz hardware...
>
>Bob - you're slacking. I always remember you talking about your
>quad xeon in the past.


I used to have a quad 700mhz xeon.  A few years back I replaced it with a dual
2.8, which is at least 2x faster...



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.