Test position: Robson vs Vocaturo

Martin Thoresen · Post by **Martin Thoresen** » Mon Jan 25, 2010 8:07 pm

[d]2r2rk1/1b2bp1p/pq2p1p1/1p1pP3/n1nN1P2/P1P5/1P2QNPP/1BB1RRK1 w - -[/d]

Engine: Stockfish 1.6.2 JA 64-bit 4CPU (1024 MB)
by Tord Romstad, Marco Costalba, Joona Kii

14.01  0:01   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg6+ Kh8 8.Qh5+ Kg8 (9.822.022) 5340 

15.01  0:02   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg6+ Kh8 8.Qh5+ Kg8 (15.426.014) 5709 

16.01  0:04   +0.20++  1.f5 gxf5 2.Nxf5 exf5 3.Bxf5 Rc6 
                       4.Qg4+ Rg6 5.Bxg6 hxg6 6.b3 Nxc3 
                       7.bxc4 bxc4 8.Bh6 Re8 9.e6 (29.809.356) 6175 

17.01  0:09   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg6+ Kh8 8.Qh5+ Kg8 (63.837.431) 6486 

18.01  0:18   +0.24    1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.Qg4+ Kh8 5.Qh3 Rg6 6.b3 Rfg8 
                       7.Bxg6 Qxg6 8.bxa4 d4 9.g3 dxc3 
                       10.Qd7 Qc6 11.Qxc6 Bxc6 12.Ne4 Bxe4 
                       13.Rxe4 Bc5+ 14.Kg2 (119.238.855) 6597 

19.01  0:35   +0.36    1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.Qg4+ Kh8 5.Qh3 Rg6 6.b3 Rfg8 
                       7.Bxg6 Qxg6 8.bxa4 d4 9.g3 dxc3 
                       10.Qd7 Qc6 11.Qxc6 Bxc6 12.Ne4 Rg4 
                       13.Nxc3 Bc5+ 14.Rf2 (236.926.937) 6703 

20.01  0:57   +0.16    1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.Qg4+ Kh8 5.Qh3 Rg6 6.b3 Rfg8 
                       7.Bxg6 Qxg6 8.bxa4 Bc5 9.g3 bxa4 
                       10.Bf4 Bc8 11.Qh4 Bxa3 12.Nd1 Bg4 
                       13.Qf6+ Qxf6 14.exf6 (390.917.593) 6761 

21.01  2:04   +0.12    1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.Qg4+ Kh8 5.Qh3 Rg6 6.b3 Rfg8 
                       7.Bxg6 Qxg6 8.bxa4 Bc5 9.g3 bxa4 
                       10.Qh4 Re8 11.Qf4 Bc6 12.Be3 Nxe3 
                       13.Rxe3 Bxe3 14.Qxe3 (840.946.303) 6775 

22.01  5:42   +0.04    1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.Qg4+ Kh8 5.Qh3 Rg6 6.b3 Rfg8 
                       7.Bxg6 Qxg6 8.bxa4 Bc5 9.g3 bxa4 
                       10.Qh4 Re8 11.Qf4 Bc6 12.Be3 Nxe3 
                       13.Rxe3 Bxe3 14.Qxe3 (2.274.130.162) 6637 

best move: f4-f5 time: 5:43.733 min  n/s: 6.636.121  nodes: 2.281.040.635 





Engine: Spark 0.3 64-bit 4CPU (1024 MB)
by AJ Siemelink

15/22  0:01   -0.13    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qe3 Nxd1 
                       4.Rxd1 Rc7 5.Ba2 Rfc8 6.Bb2 Kf8 
                       7.Ng4 h5 8.Nf6 Bxf6 9.exf6 (15.132.687) 8737 

16/28  0:03   -0.09    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qe3 Nxd1 
                       4.Rxd1 Rc7 5.Bb2 Rfc8 6.Ba2 Kf8 
                       7.Ng4 h5 8.Nf6 Bxf6 9.exf6 Qd6 (27.610.835) 9029 

17/30  0:04   -0.24    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qe3 Nxd1 
                       4.Rxd1 Rc7 5.Bb2 Rfc8 6.Ng4 b4 
                       7.axb4 Qxb4 8.Qf2 h5 9.Ne3 a5 10.Bd3 a4 (36.873.899) 9197 

18/31  0:07   -0.24    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qe3 Nxd1 
                       4.Rxd1 Rc7 5.Bb2 Rfc8 6.Ng4 b4 
                       7.axb4 Qxb4 8.Qf2 h5 9.Ne3 Bc5 
                       10.Ba2 Bb6 11.Qe2 (68.681.899) 9288 

19/32  0:13   -0.22    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qe3 Nxd1 
                       4.Rxd1 Rc7 5.Bb2 Rfc8 6.Ng4 b4 
                       7.axb4 Qxb4 8.Qf2 h5 9.Ne3 Bc5 
                       10.Ba2 Bb6 11.Qe2 a5 (128.123.433) 9322 

20/33  0:38   -0.44    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qe3 Rc4 
                       4.Bb2 b4 5.Bxc3 Rxc3 6.Rd3 Rxa3 
                       7.Rxa3 bxa3 8.Nd3 Rc8 9.Ba2 a5 
                       10.Rb1 Bb4 11.Nxb4 axb4 12.Qd3 (341.980.293) 8954 

20/38  1:14   -0.34    1.Bd3 Bc5 2.Bxc4 bxc4 3.Be3 h5 4.Rb1 Bc6 
                       5.Nh3 Rb8 6.Ng5 Rfc8 7.Qe1 Bxd4 
                       8.Bxd4 Nc5 9.Bxc5 Qxc5+ 10.Qf2 Ba4 
                       11.Qxc5 Rxc5 12.Kf2 Bc2 (654.470.400) 8808 

21/38  2:41   -0.39    1.Bd3 Kh8 2.Ng4 a5 3.Ne3 Bc5 4.Nb3 Ba6 
                       5.Rf2 Rfd8 6.Nxc5 Nxc5 7.Rd1 Nxd3 
                       8.Qxd3 b4 9.Nxc4 dxc4 10.Qe2 Rxd1+ 
                       11.Qxd1 Rd8 12.Qe2 Bb7 13.Be3 (1.388.599.865) 8585 

21/38  3:54   -0.09    1.f5 (1.954.880.089) 8333 

21/55  4:41   +0.02    1.f5 exf5 2.Nxf5 Rce8 3.Nd4 Bd6 
                       4.Nd3 f5 5.Bh6 Bxa3 6.e6 Naxb2 
                       7.Bxf8 Rxf8 8.Ne5 Nxe5 9.Qxe5 Be7 
                       10.Qg3 b4 11.Bc2 bxc3 12.Qxc3 (2.318.042.192) 8231 

22/55  5:36   -0.07    1.f5 exf5 2.Nxf5 Rce8 3.Nd4 Bd6 
                       4.Nd3 f5 5.Bh6 Bxa3 6.e6 Naxb2 
                       7.Bxf8 Rxf8 8.Ne5 Nxe5 9.Qxe5 Be7 
                       10.Qg3 b4 11.Rxf5 Rxf5 12.Bxf5 bxc3 
                       13.Qb8+ Kg7 (2.766.223.108) 8225 

best move: f4-f5 time: 5:38.179 min  n/s: 8.228.067  nodes: 2.782.559.479

Eelco de Groot · Post by **Eelco de Groot** » Mon Jan 25, 2010 8:13 pm

Martin T wrote:

Code: Select all

22.01  5:42   +0.04    1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.Qg4+ Kh8 5.Qh3 Rg6 6.b3 Rfg8 
                       7.Bxg6 Qxg6 8.bxa4 Bc5 9.g3 bxa4 
                       10.Qh4 Re8 11.Qf4 Bc6 12.Be3 Nxe3 
                       13.Rxe3 Bxe3 14.Qxe3 (2.274.130.162) 6637 

best move: f4-f5 time: 5:43.733 min  n/s: 6.636.121  nodes: 2.281.040.635

Thanks Martin, but I bet the score for f5 will go up a bit at ply 23 or 24 at the latest

Louis tested it with Stockfish on his octocore see earlier post on page 1, but only to 22 plies.

Martin Thoresen · Post by **Martin Thoresen** » Mon Jan 25, 2010 8:24 pm

Code: Select all

[d]2r2rk1/1b2bp1p/pq2p1p1/1p1pP3/n1nN1P2/P1P5/1P2QNPP/1BB1RRK1 w - -[/d]

Engine: Stockfish 5894 100124 32-bit 4CPU (1024 MB)
by Tord Romstad, Marco Costalba, Joona Kii

14.02  0:01   -0.60    1.Nf3 Bc5 2.Bd3 Rc7 3.Kh1 Be7 4.Ng4 h5 
                       5.Be3 Nxe3 6.Nxe3 Nc5 7.Nd4 Ne4 
                       8.Bxe4 dxe4 (5.649.486) 3188 

15.01  0:02   -0.76    1.Nf3 Bc5 2.Bd3 Rfd8 3.Kh1 Bf8 4.Ng4 Bg7 
                       5.Qc2 Nc5 6.Bxc4 dxc4 7.Be3 h5 
                       8.Nf6+ Bxf6 9.Bxc5 Qxc5 10.exf6 (7.995.556) 3379 

15.06  0:02   -0.44    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qd3 f6 
                       4.Ng4 Nxd1 5.Rxd1 f5 6.Nf2 Rc7 7.Bb2 Rfc8 
                       8.Ba2 g5 9.Qg3 g4 10.Qd3 Bf8 11.Qe3 (10.390.652) 3512 

16.01  0:03   -0.40    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qd3 f6 
                       4.Ng4 Nxd1 5.Rxd1 f5 6.Nf2 Rc7 7.Bb2 Rfc8 
                       8.Ba2 g5 9.Qg3 g4 10.Qd3 Bf8 11.Qe3 (13.160.871) 3632 

17.01  0:04   -0.36    1.Rd1 Nxa3 2.bxa3 Nxc3 3.Qd3 f6 
                       4.Ng4 Nxd1 5.Rxd1 f5 6.Nf2 Rc7 7.Bb2 Rfc8 
                       8.Ba2 g5 9.Qg3 g4 10.Qd3 Bf8 11.Qd2 Bc5 (17.165.536) 3754 

17.02  0:06   -0.16++  1.Ng4 Nc5 2.Nf2 Na4 3.Ng4 (23.685.554) 3861 

18.01  0:09   -0.12    1.Ng4 Nc5 2.Nf2 Nd7 3.Ng4 Bc5 4.Qd1 a5 
                       5.Bc2 a4 6.Rf2 Bxd4 7.cxd4 f5 8.Ne3 Rfe8 
                       9.Nxc4 Rxc4 10.Be3 (36.616.230) 3898 

18.04  0:12   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (51.231.243) 3949 

19.01  0:15   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (63.327.758) 3995 

20.01  0:19   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (77.813.588) 4047 

21.01  0:28   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (117.816.135) 4105 

21.02  1:07   +0.20++  1.Qg4 Ncxb2 2.f5 exf5 3.Nxf5 Qe6 
                       4.Qh3 Rxc3 5.Nxe7+ Qxe7 6.Qh6 Rxc1 
                       7.Ng4 Rxe1 8.Nf6+ Qxf6 9.exf6 (277.937.584) 4144 

22.01  2:01   -0.16    1.Qg4 Kh8 2.f5 exf5 3.Nxf5 gxf5 
                       4.Bxf5 Rc6 5.Qh3 Rg6 6.b3 Rfg8 
                       7.Bxg6 Qxg6 8.bxa4 d4 9.g4 d3 10.Bf4 Bc5 
                       11.axb5 Qxg4+ 12.Qxg4 Rxg4+ 13.Bg3 axb5 
                       14.e6 (505.961.313) 4153 

22.02  2:09   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (538.068.709) 4157 

23.01  2:32   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (638.801.919) 4185 

24.01  3:02   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (768.166.387) 4202 

25.01  3:45   0.00     1.f5 exf5 2.Nxf5 gxf5 3.Bxf5 Rc6 
                       4.e6 fxe6 5.Bxh7+ Kxh7 6.Qh5+ Kg8 
                       7.Qg4+ Kh7 8.Qh5+ (956.942.445) 4235 

26.01  6:43   +0.20++  1.f5 Kh8 2.fxg6 fxg6 3.Bh6 Naxb2 
                       4.Qg4 Bc5 5.Bxf8 Rxf8 6.Qh3 Rg8 
                       7.Ng4 Bxd4+ 8.cxd4 Qxd4+ 9.Kh1 (1.737.879.388) 4306 

best move: f4-f5 time: 6:46.071 min  n/s: 4.306.254  nodes: 1.748.640.869

Martin Thoresen · Post by **Martin Thoresen** » Mon Jan 25, 2010 8:26 pm

Eelco de Groot wrote:Thanks Martin, but I bet the score for f5 will go up a bit at ply 23 or 24 at the latest Louis tested it with Stockfish on his octocore see earlier post on page 1, but only to 22 plies.

I just posted a Stockfish evaluation with a developer version. Also see my Spark 0.3 64-bit 4CPU evaluation further up.

Note that the developer Stockfish only is 32-bit.

Also, Louis' octocore seems really slow compared to my Quad?

zullil · Post by **zullil** » Mon Jan 25, 2010 8:42 pm

Martin T wrote: Also, Louis' octocore seems really slow compared to my Quad?

Yes, and I'm trying to understand why. Did you compile your own binary? If so, what compiler? With PGO? If not, whose binary do you use?

Also, what speed is your CPU?

Martin Thoresen · Post by **Martin Thoresen** » Mon Jan 25, 2010 8:50 pm

zullil wrote: Yes, and I'm trying to understand why. Did you compile your own binary? If so, what compiler? With PGO? If not, whose binary do you use?

Also, what speed is your CPU?

No, using JA builds. I am not a programmer so my skills in that area are closer to zero.

My CPU (i920) runs at 3808 MHz, RAM latency at 7-7-7-20-60 1T.

I assume you have dual quad-Xeon's somewhere around 2 Ghz then? Else I can't really explain the small difference we're seeing.

Eelco de Groot · Post by **Eelco de Groot** » Mon Jan 25, 2010 10:20 pm

zullil wrote:
Eelco de Groot wrote:Rainbow Serpent only likes 21. Ng4 from the start. After a while f5 rises up to second place in the move list, but is only played once as twentysecond move after 21.Ng4 later not anymore so there must be something in the sacrifice that the program does not like.

What does Rainbow Serpent say about this?

[d]2r2rk1/1b2bp1p/pq2p1p1/1p1pPP2/n1nN4/P1P5/1P2QNPP/1BB1RRK1 w - -

It was White to move! I hadn't noticed that. I compared the two builds 55 and 61 again, just on single processor Athlon, I still would have to make changes to the search after splitpoints for a SMP build, it would run now, at least I hope, it could probably be made better. I do have a preference for the latest build 61 but it isn't faster in nodes per second as I had hoped. Maybe this means that the compiler is smart enough to save the result of evaluate() without having been told explicitly to do this and can use the result later in futility pruning. I'm not sure that is true or would be correct because it may not be what the programmer intended, the compiler is too smart for the programmer!

This is with the older build 55, White to move after Black has "null moved" so it is not really a game position anymore if I understand correctly:

2r2rk1/1b2bp1p/pq2p1p1/1p1pPP2/n1nN4/P1P5/1P2QNPP/1BB1RRK1 w - -

Engine: Rainbow Serpent 1.6.3s(dc) Build 55 (Athlon 2009 MHz, 256 MB)
by Tord Romstad, Marco Costalba, Joona Kiiski Modifications: Dann Corbit

1.00 0:00 -0.40 1.f6 (698) 2

2.00 0:00 -0.28 1.f6 Bc5 (769) 3

3.00 0:00 -0.44 1.f6 Bc5 2.Rd1 (1.334) 5

3.00 0:00 -0.32 1.fxe6 fxe6 2.Ng4 Kh8 (2.348) 9

4.00 0:00 +1.85 1.fxe6 fxe6 2.Ng4 Kh8 3.Nf6 (5.330) 20

5.00 0:00 +1.05 1.fxe6 fxe6 2.Ng4 Kh8 3.Nf6 Bxf6
4.exf6 (7.780) 29

6.00 0:00 +1.45 1.fxe6 fxe6 2.Ng4 Kh8 3.Nh6 Ncxb2
4.Bxb2 Nxc3 5.Rxf8+ Bxf8 (15.171) 51

7.00 0:00 +2.94 1.fxe6 fxe6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.e6 Re8 6.Nf7+ Kg8
7.Bh6 Naxb2 (88.409) 194

8.00 0:00 +2.94 1.fxe6 fxe6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.e6 Re8 6.Nf7+ Kg8
7.Bh6 Naxb2 (116.485) 218

9.00 0:01 +2.62 1.fxe6 Bc5 2.exf7+ Rxf7 3.e6 Re7
4.Bg5 Rce8 5.Bxe7 Rxe7 6.Nd3 Bxd4+
7.cxd4 Qxd4+ 8.Qf2 Qxf2+ 9.Rxf2 (344.650) 333

9.07 0:01 +3.07 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.e6 Re8 6.Bf4 Bf6
7.Nf7+ Kg8 (433.411) 350

10.01 0:01 +3.31++ 1.fxg6 hxg6 2.Ng4 Kh8 3.Qd3 (684.769) 384

11.01 0:02 +3.07 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.Qf2 Rf8 6.b3 Nxc3
7.bxc4 bxc4 (1.148.732) 401

12.01 0:22 +2.94 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Nxa3
4.Nf7+ Kg8 5.Nh6+ Kh8 6.Bd3 Bc5
7.Nf7+ Kg8 8.Qf3 Qc7 9.Nh6+ Kh8
10.Qh3 Nxb2 11.Bxb2 Nc4 12.Bxc4 (10.032.193) 448

13.01 0:27 +3.67++ 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Nxa3
4.Bd3 Bc5 5.Rf7 Bxd4+ 6.cxd4 Qxd4+
7.Be3 Qxb2 8.Qg4 (12.589.755) 449

14.01 1:16 +3.87 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Nxa3
4.Bd3 Nc4 5.Nf7+ Kg8 6.Qf2 Nc5 7.Bc2 Ne4
8.Bxe4 dxe4 9.Nh6+ Kh8 10.Nf7+ Rxf7
11.Qxf7 Bf8 12.Rf6 Nxe5 (34.707.851) 453

15.01 4:26 +3.35 1.fxg6 fxg6 2.Ng4 Ncxb2 3.Bxb2 Kh8
4.Rxf8+ Rxf8 5.Nf6 Rf7 6.Ba1 Nc5
7.Rf1 Qd8 8.Qf2 Bxf6 9.exf6 e5
10.Nf3 Nd7 11.Qa7 (123.395.142) 463

16.01 10:25 +3.87 1.fxg6 fxg6 2.Qg4 Rf7 3.Nxe6 Kh8
4.Qh3 Rg8 5.Nd4 Bc8 6.e6 Rf6 7.b3 Nxc3
8.Qxc3 Nxa3 9.Bxa3 Bxa3 10.b4 (283.940.804) 454

17.01 14:45 +4.00 1.fxg6 fxg6 2.Qg4 Rf7 3.Nxe6 Kh8
4.Qh3 Rg8 5.Nd4 Bc8 6.e6 Rf6 7.b3 Nxc3
8.Qxc3 b4 9.Qd3 Nxa3 10.Ba2 a5
11.Bh6 Bxe6 12.Nxe6 Rxe6 13.Qxd5 (395.320.957) 446

18.01 29:07 +4.40 1.fxg6 fxg6 2.Qg4 Kh8 3.Nxe6 Bc5
4.Nxf8 Rxf8 5.Qh4 Kg8 6.e6 Bxf2+
7.Rxf2 Rxf2 8.Qxf2 Qxf2+ 9.Kxf2 Kf8
10.Kg3 Nd6 11.Bf4 Nf5+ 12.Kh3 Ke8
13.g4 Ne7 14.Re2 (761.962.895) 436

best move: f5xg6 time: 37:39.031 min n/s: 444.270 nodes: 1.003.620.327

With build 61 a PV that I like a bit better, for instance earlier 2. Qg4
The two builds start to diverge only around plydepth twelve. These are my own 32 bit builds but only slightly optimized for speed, MSVC 2005, no PGO.

2r2rk1/1b2bp1p/pq2p1p1/1p1pPP2/n1nN4/P1P5/1P2QNPP/1BB1RRK1 w - -

Engine: Rainbow Serpent 1.6.3s(dc) Build 61 (Athlon 2009 MHz, 256 MB)
by Tord Romstad, Marco Costalba, Joona Kiiski Modifications: Dann Corbit

1.00 0:00 -0.40 1.f6 (698) 2

2.00 0:00 -0.28 1.f6 Bc5 (769) 3

3.00 0:00 -0.44 1.f6 Bc5 2.Rd1 (1.334) 5

3.00 0:00 -0.32 1.fxe6 fxe6 2.Ng4 Kh8 (2.348) 9

4.00 0:00 +1.85 1.fxe6 fxe6 2.Ng4 Kh8 3.Nf6 (5.330) 21

5.00 0:00 +1.05 1.fxe6 fxe6 2.Ng4 Kh8 3.Nf6 Bxf6
4.exf6 (7.780) 31

6.00 0:00 +1.45 1.fxe6 fxe6 2.Ng4 Kh8 3.Nh6 Ncxb2
4.Bxb2 Nxc3 5.Rxf8+ Bxf8 (15.171) 53

7.00 0:00 +2.94 1.fxe6 fxe6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.e6 Re8 6.Nf7+ Kg8
7.Bh6 Naxb2 (88.409) 195

8.00 0:00 +2.94 1.fxe6 fxe6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.e6 Re8 6.Nf7+ Kg8
7.Bh6 Naxb2 (116.485) 226

9.00 0:01 +2.62 1.fxe6 Bc5 2.exf7+ Rxf7 3.e6 Re7
4.Bg5 Rce8 5.Bxe7 Rxe7 6.Nd3 Bxd4+
7.cxd4 Qxd4+ 8.Qf2 Qxf2+ 9.Rxf2 (344.665) 339

9.07 0:01 +3.07 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.e6 Re8 6.Bf4 Bf6
7.Nf7+ Kg8 (433.435) 355

10.01 0:01 +3.31++ 1.fxg6 hxg6 2.Ng4 Kh8 3.Qd3 (684.851) 388

11.01 0:02 +3.07 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.Qf2 Rf8 6.b3 Nxc3
7.bxc4 bxc4 (1.098.095) 406

12.01 0:17 +3.47 1.fxg6 fxg6 2.Ng4 Kh8 3.Nh6 Rf5
4.Bxf5 exf5 5.g4 Bc5 6.Qf2 Nxa3
7.bxa3 Nxc3 8.Nf7+ Kg8 9.Be3 Ne4 (7.642.814) 447

13.01 0:57 +3.07 1.fxg6 fxg6 2.Qg4 Kh8 3.Nxe6 Rf7
4.Qh3 Rg8 5.Nd4 Bc5 6.e6 Re7 7.Qh4 Naxb2
8.Ng4 Bxd4+ 9.cxd4 Qxd4+ 10.Rf2 (26.479.122) 464

14.01 1:28 +3.83 1.fxg6 fxg6 2.Qg4 Kh8 3.Nxe6 Rg8
4.Ng5 Rg7 5.e6 Nd6 6.Be3 Qd8 7.Nxh7 Rc4
8.Bd4 Rxd4 9.cxd4 Kxh7 10.Qh5+ Kg8
11.Bxg6 Nxb2 (39.442.175) 445

15.01 3:31 +3.91 1.fxg6 fxg6 2.Qg4 Kh8 3.Nxe6 Rg8
4.Ng5 Rg7 5.e6 Rf8 6.Nf7+ Rfxf7
7.exf7 Rxf7 8.Qd7 Qc6 9.Qxc6 Bxc6
10.Nd3 Kg8 11.Rxf7 Kxf7 12.b3 Nxc3
13.bxc4 Nxb1 14.Ne5+ (95.967.967) 454

16.01 4:00 +4.00 1.fxg6 fxg6 2.Qg4 Kh8 3.Nxe6 Rg8
4.Ng5 Rg7 5.e6 Rf8 6.Nf7+ Rfxf7
7.exf7 Rxf7 8.Qd7 Qc6 9.Qxc6 Bxc6
10.b3 Nxc3 11.bxc4 Nxb1 12.cxd5 Bxd5
13.Bb2+ Kg8 14.Rxb1 (109.827.694) 456

17.01 6:42 +4.24++ 1.fxg6 fxg6 2.Qg4 Kh8 3.Nxe6 Rg8
4.Ng5 Rg7 5.e6 Rf8 6.Nf7+ Rfxf7
7.exf7 Rxf7 8.Re6 Qd8 9.Bxg6 hxg6
10.Rxg6 Bc5 11.Rh6+ Rh7 12.Rxh7+ Kxh7
13.Qh5+ Kg8 14.Qg6+ (180.291.361) 447

18.01 12:37 +4.08 1.fxg6 fxg6 2.Qg4 Kh8 3.Nxe6 Bc5
4.Nxf8 Rxf8 5.Qh4 Kg8 6.e6 Bxf2+
7.Rxf2 Rxf2 8.Qxf2 Qxf2+ 9.Kxf2 Kf8
10.Bh6+ Ke8 11.b3 Nxc3 12.bxc4 Nxb1
13.Rxb1 dxc4 14.Bg5 (338.779.707) 447

best move: f5xg6 time: 13:05.062 min n/s: 447.633 nodes: 351.420.242

Martin's overclocked(?) i7 does seem very fast! I7 does have Popcount instruction that would give it a bit of a boost even I presume in the 32 bit builds.

Eelco

zullil · Post by **zullil** » Mon Jan 25, 2010 10:40 pm

Eelco de Groot wrote:
zullil wrote:
Eelco de Groot wrote:Rainbow Serpent only likes 21. Ng4 from the start. After a while f5 rises up to second place in the move list, but is only played once as twentysecond move after 21.Ng4 later not anymore so there must be something in the sacrifice that the program does not like.

What does Rainbow Serpent say about this?

[d]2r2rk1/1b2bp1p/pq2p1p1/1p1pPP2/n1nN4/P1P5/1P2QNPP/1BB1RRK1 w - -
It was White to move!

Sorry, it was supposed to be Black to move. And I thought I could edit the fen correctly.

zullil · Post by **zullil** » Mon Jan 25, 2010 10:43 pm

Martin T wrote:
zullil wrote: Yes, and I'm trying to understand why. Did you compile your own binary? If so, what compiler? With PGO? If not, whose binary do you use?

Also, what speed is your CPU?
No, using JA builds. I am not a programmer so my skills in that area are closer to zero.

My CPU (i920) runs at 3808 MHz, RAM latency at 7-7-7-20-60 1T.

I assume you have dual quad-Xeon's somewhere around 2 Ghz then? Else I can't really explain the small difference we're seeing.

Yes, 2.26 GHz. And I'm guessing that PGO with gcc is not as good as PGO with icc (which is not available free for Mac OS)!

Eelco de Groot · Post by **Eelco de Groot** » Tue Jan 26, 2010 10:05 pm

zullil wrote: Sorry, it was supposed to be Black to move.

Look! Rainbow Serpent just broke through the postulated "1.29 but still a material imbalance draw or unclear position?" barrier, about an hour ago! Well maybe, but I have not really looked at the endpositions. If the 1.65 is still a draw we will have to do something about it.. This was with build 55, which is not better than build 61 but I just wanted to let one analysis run longer.

[d]2r2rk1/1b2bp1p/pq2p1p1/1p1pPP2/n1nN4/P1P5/1P2QNPP/1BB1RRK1 b - -

Engine: Rainbow Serpent 1.6.3s(dc) Build 55 (Athlon 2009 MHz, 128 MB)
by Tord Romstad, Marco Costalba, Joona Kiiski Modifications: Dann Corbit

1.00 0:00 +1.77 21...gxf5 (584) 2

2.00 0:00 +1.73 21...gxf5 22.Nd3 (646) 2

3.00 0:00 +1.29 21...gxf5 22.Ng4 Kh8 (1.471) 6

4.00 0:00 +1.13 21...gxf5 22.Ng4 Nc5 23.Nh6+ Kh8 (3.970) 15

5.00 0:00 -0.56 21...gxf5 22.Nxf5 exf5 23.Bxf5 Ncxb2
24.Qg4+ Kh8 (9.723) 36

5.00 0:00 -0.16 21...Bh4 22.Qg4 Bd8 23.fxe6 Naxb2
24.exf7+ Rxf7 (12.644) 47

5.00 0:00 +0.44 21...Bc5 22.fxg6 fxg6 23.Ng4 Bxd4+
24.cxd4 Qxd4+ 25.Kh1 Naxb2 (30.779) 98

6.00 0:00 +0.64 21...Bc5 22.fxg6 fxg6 23.Rd1 Kh8
24.Ba2 Be7 (39.465) 120

6.00 0:00 +0.68 21...Nxa3 22.fxg6 fxg6 23.bxa3 Nxc3
24.Qg4 Nxb1 25.Nxe6 Nxa3 (50.033) 139

7.00 0:00 -0.48-- 21...Nxa3 22.fxg6 fxg6 23.Bd3 Nc4
24.Ng4 Kh8 25.Nh6 (79.825) 182

8.00 0:00 -0.24 21...Nxa3 22.fxg6 fxg6 23.Bd3 Bc5
24.Bh6 Nc4 25.Qg4 Bxd4 26.cxd4 Naxb2 (173.633) 271

9.00 0:01 -0.76 21...Nxa3 22.fxg6 Nxb1 23.gxh7+ Kh8
24.Bh6 f5 25.Bxf8 Nbxc3 26.bxc3 Rxf8
27.Qh5 Nxc3 (611.260) 365

9.11 0:02 0.00 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.e6 fxe6 25.Bxh7+ Kxh7 26.Qh5+ Kg8
27.Qg4+ Kh8 28.Qh3+ Kg8 29.Qg4+ (1.012.410) 357

10.01 0:03 0.00 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.e6 fxe6 25.Bxh7+ Kxh7 26.Qh5+ Kg8
27.Qg4+ Kh8 28.Qh3+ Kg8 29.Qg4+ (1.208.572) 370

11.01 0:04 0.00 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.e6 fxe6 25.Bxh7+ Kxh7 26.Qh5+ Kg8
27.Qg4+ Kh8 28.Qh3+ Kg8 29.Qg4+ (1.937.800) 398

12.01 0:06 0.00 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.e6 fxe6 25.Bxh7+ Kxh7 26.Qh5+ Kg8
27.Qg4+ Kh8 28.Qh3+ Kg8 29.Qg4+ (2.949.954) 424

13.01 0:11 0.00 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.e6 fxe6 25.Bxh7+ Kxh7 26.Qh5+ Kg8
27.Qg4+ Kh8 28.Qh3+ Kg8 29.Qg4+ (5.005.980) 446

14.01 0:19 -0.20-- 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Rg6 25.Bxg6 hxg6 26.b3 Nxc3
27.bxc4 bxc4 28.Qh3 Qd4 29.Bh6 Bc8
30.Qe3 (8.726.515) 453

15.01 0:57 -0.92 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Kh8 25.Qh3 Rg6 26.b3 Nxc3
27.Bxg6 Qxg6 28.Qxc3 d4 29.Qg3 Nxa3
30.Qxg6 fxg6 31.Bh6 Rd8 32.Rd1 Kg8
33.Ng4 Bd5 34.Rxd4 Bxb3 (25.610.430) 443

16.01 1:50 -0.48 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Rg6 25.Bxg6 hxg6 26.b3 Nxc3
27.bxc4 bxc4 28.Qh3 Ne4 29.Be3 d4
30.Nxe4 Bxe4 31.Bh6 d3+ 32.Kh1 Bf5
33.Qg3 Qe6 34.Bxf8 Kxf8 (49.295.882) 446

17.01 7:53 -1.09 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Kh8 25.Qh3 Rg6 26.b3 Nxc3
27.Bxg6 Qxg6 28.Qxc3 d4 29.Qg3 Qxg3
30.hxg3 Nxa3 31.e6 Bd5 32.exf7 Rxf7
33.Rxe7 Rxe7 34.Bxa3 Re3 (213.996.584) 451

18.01 13:07 -1.09 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Kh8 25.Qh3 Rg6 26.b3 Nxc3
27.Bxg6 Qxg6 28.Qxc3 d4 29.Qg3 Qxg3
30.hxg3 Nxa3 31.e6 f6 32.Bxa3 Bxa3
33.Nd3 Be7 34.Rf4 Rd8 (352.386.765) 447

19.01 33:55 -1.29 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Kh8 25.Qh3 Rg6 26.b3 Nxc3
27.Bxg6 Qxg6 28.Qxc3 d4 29.Qg3 Qxg3
30.hxg3 Nxa3 31.e6 f6 32.Bxa3 Bxa3
33.Nd3 Be7 34.Rf4 Rd8 (908.198.678) 446

20.01 50:44 -0.88++ 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Kh8 25.Qh3 Rg6 26.b3 Rfg8
27.Kh1 d4 28.Ng4 Nxc3 29.bxc4 bxc4
30.e6 R6g7 31.Rf2 fxe6 32.Rxe6 Qb3 (1.367.547.261) 449

21.01 114:24 -0.48 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Kh8 25.Qh3 Rg6 26.e6 d4
27.Ng4 d3+ 28.Kh1 Qd8 29.Bxg6 fxg6
30.Rxf8+ Bxf8 31.Rf1 Be7 32.Bh6 Kg8
33.Rf7 d2 34.Bxd2 h5 (3.067.370.450) 446

22.01 274:04 -1.29-- 21...exf5 22.Nxf5 Bc5 23.Qg4 Rce8 (7.395.299.546) 449

23.01 1265:00-1.65 21...exf5 22.Nxf5 gxf5 23.Bxf5 Rc6
24.Qg4+ Rg6 25.Bxg6 hxg6 26.b3 Nxc3
27.bxc4 Bc5 28.Bg5 dxc4 29.Qh3 Nd5
30.Qh6 Re8 31.Kh1 Bxf2 32.Bf6 Nxf6
33.exf6 Bxg2+ 34.Kxg2 Qc6+ (34.993.866.385) 461

Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo

Re: Test position: Robson vs Vocaturo