Cfish is back!

[Ovyron]

I have now released Cfish 11a:
https://github.com/syzygy1/Cfish/releases/tag/cfish_11a

Thanks, I'm liking these more frequent updates.

[Thomas Lagershausen]

I am not so happy with latest update because Stockfish getting dumber with more selectivity.

[d]r5rk/1bq1bpp1/1p1ppn1p/pP5P/2P1P3/P1N2P2/1B1Q1P2/1K1R1BR1 w - - 0 19

bm 19.Nd5 !!

CFish 060720:
45/56+ 06:57 1.633.436.249 3.913.959 +1,26 Sc3-d5
47/86 20:49 4.508.096.454 3.607.879 +2,91 Sc3-d5

Cfish 110720:
49/75 20:49 4.457.940.753 3.567.739 +1,46 Dd2-e3 Ta8-b8 Lf1-e2 Dc7-c5 De3xc5 b6xc5 e4-e5 d6xe5 Sc3-a4 Tb8-d8 Lb2xe5 Sf6xh5 f3-f4 Sh5-f6 Le5xf6 Lb7-e4+ Kb1-c1 Le7xf6 Sa4xc5 Le4-d5 Sc5-a4 Td8-c8 Sa4-b6 Tc8-b8 Sb6xd5 e6xd5 Td1xd5 Tg8-e8 Kc1-d2 Te8-e4 c4-c5 Tb8-e8 Le2-d3 Te4xf4 Tg1-f1 Tf4-d4 Td5xd4 Lf6xd4 Tf1-c1 Ld4xf2 Tc1-c4 Lf2-e3+ Kd2-d1 Te8-b8 Kd1-e2 Le3-g1 b5-b6 g7-g6 Ld3-e4 Lg1xc5 Tc4xc5 Tb8xb6 Tc5xa5 Kh8-g7 Ta5-a8 Kg7-f6 a3-a4 Tb6-b2+ Ke2-e3

[Alayan]

When comparing two versions a few patches away, you'll have plenty of positions one version "gets" and the other doesn't. Minor seemingly unrelated changes can end up having a big impact on specific positions.

But chess strength is about finding good moves in general positions, not in one specific test position.

So there is nothing to worry about.

[syzygy]

I am not so happy with latest update because Stockfish getting dumber with more selectivity.

But the aim of Cfish is to track Stockfish.

[zullil]

I am not so happy with latest update because Stockfish getting dumber with more selectivity.

[d]r5rk/1bq1bpp1/1p1ppn1p/pP5P/2P1P3/P1N2P2/1B1Q1P2/1K1R1BR1 w - - 0 19

bm 19.Nd5 !!

CFish 060720:
45/56+ 06:57 1.633.436.249 3.913.959 +1,26 Sc3-d5
47/86 20:49 4.508.096.454 3.607.879 +2,91 Sc3-d5

Cfish 110720:
49/75 20:49 4.457.940.753 3.567.739 +1,46 Dd2-e3 Ta8-b8 Lf1-e2 Dc7-c5 De3xc5 b6xc5 e4-e5 d6xe5 Sc3-a4 Tb8-d8 Lb2xe5 Sf6xh5 f3-f4 Sh5-f6 Le5xf6 Lb7-e4+ Kb1-c1 Le7xf6 Sa4xc5 Le4-d5 Sc5-a4 Td8-c8 Sa4-b6 Tc8-b8 Sb6xd5 e6xd5 Td1xd5 Tg8-e8 Kc1-d2 Te8-e4 c4-c5 Tb8-e8 Le2-d3 Te4xf4 Tg1-f1 Tf4-d4 Td5xd4 Lf6xd4 Tf1-c1 Ld4xf2 Tc1-c4 Lf2-e3+ Kd2-d1 Te8-b8 Kd1-e2 Le3-g1 b5-b6 g7-g6 Ld3-e4 Lg1xc5 Tc4xc5 Tb8xb6 Tc5xa5 Kh8-g7 Ta5-a8 Kg7-f6 a3-a4 Tb6-b2+ Ke2-e3

Unless you've used one thread only, what your seeing in these two searches might just be "random noise".

[syzygy]

Unless you've used one thread only, what your seeing in these two searches might just be "random noise".

And even with 1 thread finding the winning move or not can easily depend on the size of the hash table and the content of the hash table at the start of the search.

If Thomas was indeed using the same settings and a single thread, then apparently two minor changes to the evaluation function have made the difference for this position (when using Thomas' settings). There was no change to the search (apart from a very small change that changed the node count of a bench at depth 20 by 1 node).

[zullil]

I am not so happy with latest update because Stockfish getting dumber with more selectivity.

[d]r5rk/1bq1bpp1/1p1ppn1p/pP5P/2P1P3/P1N2P2/1B1Q1P2/1K1R1BR1 w - - 0 19

bm 19.Nd5 !!

CFish 060720:
45/56+ 06:57 1.633.436.249 3.913.959 +1,26 Sc3-d5
47/86 20:49 4.508.096.454 3.607.879 +2,91 Sc3-d5

Cfish 110720:
49/75 20:49 4.457.940.753 3.567.739 +1,46 Dd2-e3 Ta8-b8 Lf1-e2 Dc7-c5 De3xc5 b6xc5 e4-e5 d6xe5 Sc3-a4 Tb8-d8 Lb2xe5 Sf6xh5 f3-f4 Sh5-f6 Le5xf6 Lb7-e4+ Kb1-c1 Le7xf6 Sa4xc5 Le4-d5 Sc5-a4 Td8-c8 Sa4-b6 Tc8-b8 Sb6xd5 e6xd5 Td1xd5 Tg8-e8 Kc1-d2 Te8-e4 c4-c5 Tb8-e8 Le2-d3 Te4xf4 Tg1-f1 Tf4-d4 Td5xd4 Lf6xd4 Tf1-c1 Ld4xf2 Tc1-c4 Lf2-e3+ Kd2-d1 Te8-b8 Kd1-e2 Le3-g1 b5-b6 g7-g6 Ld3-e4 Lg1xc5 Tc4xc5 Tb8xb6 Tc5xa5 Kh8-g7 Ta5-a8 Kg7-f6 a3-a4 Tb6-b2+ Ke2-e3

Unless you've used one thread only, what your seeing in these two searches might just be "random noise".

Here's Cfish-dev, with 20 threads:

info depth 48 seldepth 66 multipv 1 score cp 195 lowerbound nodes 11852528454 nps 33735401 hashfull 991 tbhits 0 time 351338 pv c3d5
info depth 47 currmove c3d5 currmovenumber 1
info depth 48 seldepth 66 multipv 1 score cp 204 lowerbound nodes 11871346981 nps 33734713 hashfull 992 tbhits 0 time 351903 pv c3d5
info depth 46 currmove c3d5 currmovenumber 1
info depth 48 seldepth 66 multipv 1 score cp 218 lowerbound nodes 11910322219 nps 33734364 hashfull 992 tbhits 0 time 353062 pv c3d5
info depth 45 currmove c3d5 currmovenumber 1
info depth 48 seldepth 66 multipv 1 score cp 237 lowerbound nodes 12013278117 nps 33728677 hashfull 993 tbhits 0 time 356174 pv c3d5
info depth 44 currmove c3d5 currmovenumber 1
info depth 48 seldepth 67 multipv 1 score cp 264 lowerbound nodes 12694511084 nps 33720743 hashfull 994 tbhits 0 time 376460 pv c3d5
info depth 43 currmove c3d5 currmovenumber 1
info depth 48 seldepth 77 multipv 1 score cp 300 lowerbound nodes 13533637692 nps 33714492 hashfull 998 tbhits 0 time 401419 pv c3d5
info depth 42 currmove c3d5 currmovenumber 1
info depth 42 currmove d2e3 currmovenumber 2
info depth 42 currmove f1d3 currmovenumber 3
info depth 42 currmove d1c1 currmovenumber 4
info depth 42 currmove b1a2 currmovenumber 5
info depth 42 currmove f1e2 currmovenumber 6
info depth 42 currmove a3a4 currmovenumber 7
info depth 42 currmove g1g3 currmovenumber 8
info depth 42 currmove d2e2 currmovenumber 9
info depth 42 currmove d2c2 currmovenumber 10
info depth 42 currmove g1h1 currmovenumber 11
info depth 42 currmove c3a4 currmovenumber 12
info depth 42 currmove c4c5 currmovenumber 13
info depth 42 currmove b1a1 currmovenumber 14
info depth 42 currmove f1h3 currmovenumber 15
info depth 42 currmove d2d3 currmovenumber 16
info depth 42 currmove d1e1 currmovenumber 17
info depth 42 currmove f3f4 currmovenumber 18
info depth 42 currmove d2f4 currmovenumber 19
info depth 42 currmove g1g6 currmovenumber 20
info depth 42 currmove e4e5 currmovenumber 21
info depth 42 currmove d2g5 currmovenumber 22
info depth 42 currmove b1c1 currmovenumber 23
info depth 42 currmove b1c2 currmovenumber 24
info depth 42 currmove g1g2 currmovenumber 25
info depth 42 currmove b2c1 currmovenumber 26
info depth 42 currmove d2d5 currmovenumber 27
info depth 42 currmove d2d4 currmovenumber 28
info depth 42 currmove d2c1 currmovenumber 29
info depth 42 currmove d2e1 currmovenumber 30
info depth 42 currmove c3e2 currmovenumber 31
info depth 42 currmove b2a1 currmovenumber 32
info depth 42 currmove f1g2 currmovenumber 33
info depth 42 currmove g1g5 currmovenumber 34
info depth 42 currmove g1g4 currmovenumber 35
info depth 42 currmove c3a2 currmovenumber 36
info depth 42 currmove d2h6 currmovenumber 37
info depth 42 currmove d2d6 currmovenumber 38
info depth 42 currmove g1g7 currmovenumber 39
info depth 48 seldepth 77 multipv 1 score cp 312 nodes 14534015098 nps 33723334 hashfull 998 tbhits 0 time 430978 pv c3d5 b7d5 c4d5 e6e5 d1c1 c7d8 c1c6 f6h7 d2e3 f7f5 f1h3 e7g5 e3b6 d8b6 c6b6 g8b8 b6d6 b8b5 b1a2 f5e4 f3e4 h7f6 h3f5 f6h5 g1h1 b5b2 a2b2 h5f4 d6c6 a8b8 b2c3 h8g8 d5d6 g7g6 c6c8 b8c8 f5c8 g8f7 c3c4 f7e8 c4b5 e8d8 c8a6 f4h5 b5c6 h5f6 h1g1 f6e4 c6d5 e4f6 d5e6 e5e4 a6c4 e4e3 f2e3

[zullil]

To my surprise, Lc0 (network 384x30-t60-4300) fails, even with 100M nodes searched:

Code: Select all

0712 18:13:44.224235 139690876974848 ../../src/mcts/stoppers/stoppers.cc:66] Stopped search: Reached visits limit: 100000000>=100000000
0712 18:13:44.224401 139690876974848 ../../src/chess/uciloop.cc:222] << info depth 30 seldepth 82 time 3339646 nodes 100000000 score cp 87 wdl 642 210 148 nps 30160 tbhits 0 pv f1d3 e6e5 f3f4 a8e8 c3d5 b7d5 c4d5 f6d7 f4e5 d6e5 d3f1 c7d6 f1h3 a5a4 g1g3 d7c5 d2e2 e8a8 g3f3 g8f8 f3c3 a8a7 b1a2 f8d8 c3c4 a7a5 c4c2 a5a7 e2f3 e7g5 c2c4 a7c7 b2c3 h8g8 c3b4 g5e7
0712 18:13:44.224698 139690876974848 ../../src/mcts/search.cc:467] === Move stats:
0712 18:13:44.224712 139690876974848 ../../src/mcts/search.cc:468] d2d5  (297 ) N:     924 (+ 0) (P:  0.50%) (WL: -0.94134) (D: 0.013) (M: 38.6) (Q: -0.94134) (U: 1.43515) (S:  0.49381) (V: -0.9970) 
0712 18:13:44.224719 139690876974848 ../../src/mcts/search.cc:468] d2g5  (298 ) N:     991 (+ 0) (P:  0.54%) (WL: -0.93273) (D: 0.012) (M: 39.1) (Q: -0.93273) (U: 1.42588) (S:  0.49315) (V: -0.9985) 
0712 18:13:44.224725 139690876974848 ../../src/mcts/search.cc:468] d2d6  (299 ) N:    1001 (+ 0) (P:  0.54%) (WL: -0.91714) (D: 0.005) (M: 35.4) (Q: -0.91714) (U: 1.41014) (S:  0.49300) (V: -0.9979) 
0712 18:13:44.224730 139690876974848 ../../src/mcts/search.cc:468] d2h6  (300 ) N:    1062 (+ 0) (P:  0.60%) (WL: -0.98332) (D: 0.004) (M: 33.8) (Q: -0.98332) (U: 1.47723) (S:  0.49391) (V: -0.9065) 
0712 18:13:44.224736 139690876974848 ../../src/mcts/search.cc:468] g1g5  (165 ) N:    1164 (+ 0) (P:  0.61%) (WL: -0.88418) (D: 0.032) (M: 52.9) (Q: -0.88418) (U: 1.37853) (S:  0.49435) (V: -0.9572) 
0712 18:13:44.224742 139690876974848 ../../src/mcts/search.cc:468] g1g6  (167 ) N:    1286 (+ 0) (P:  0.68%) (WL: -0.89932) (D: 0.027) (M: 51.2) (Q: -0.89932) (U: 1.39275) (S:  0.49343) (V: -0.3750) 
0712 18:13:44.224747 139690876974848 ../../src/mcts/search.cc:468] e4e5  (796 ) N:    1394 (+ 0) (P:  0.65%) (WL: -0.73728) (D: 0.092) (M: 64.8) (Q: -0.73728) (U: 1.23088) (S:  0.49360) (V:  0.1152) 
0712 18:13:44.224753 139690876974848 ../../src/mcts/search.cc:468] g1g4  (163 ) N:    2217 (+ 0) (P:  0.62%) (WL: -0.24786) (D: 0.246) (M: 83.7) (Q: -0.24786) (U: 0.74220) (S:  0.49434) (V: -0.5305) 
0712 18:13:44.224758 139690876974848 ../../src/mcts/search.cc:468] c4c5  (727 ) N:    2353 (+ 0) (P:  0.78%) (WL: -0.38416) (D: 0.217) (M: 81.2) (Q: -0.38416) (U: 0.87847) (S:  0.49432) (V:  0.0854) 
0712 18:13:44.224764 139690876974848 ../../src/mcts/search.cc:468] g1g7  (169 ) N:    3001 (+ 0) (P:  1.44%) (WL: -0.76811) (D: 0.072) (M: 56.6) (Q: -0.76811) (U: 1.26216) (S:  0.49405) (V: -0.5698) 
0712 18:13:44.224781 139690876974848 ../../src/mcts/search.cc:468] f3f4  (584 ) N:    3405 (+ 0) (P:  0.85%) (WL: -0.16466) (D: 0.268) (M: 79.6) (Q: -0.16466) (U: 0.65879) (S:  0.49413) (V: -0.0273) 
0712 18:13:44.224787 139690876974848 ../../src/mcts/search.cc:468] f1g2  (132 ) N:    3432 (+ 0) (P:  0.70%) (WL: -0.04202) (D: 0.295) (M: 80.4) (Q: -0.04202) (U: 0.53619) (S:  0.49417) (V: -0.1912) 
0712 18:13:44.224792 139690876974848 ../../src/mcts/search.cc:468] f1h3  (138 ) N:    3900 (+ 0) (P:  0.79%) (WL: -0.04102) (D: 0.294) (M: 79.6) (Q: -0.04102) (U: 0.53526) (S:  0.49424) (V:  0.0425) 
0712 18:13:44.224797 139690876974848 ../../src/mcts/search.cc:468] c3d5  (491 ) N:    4410 (+ 0) (P:  1.03%) (WL: -0.11935) (D: 0.285) (M: 58.5) (Q: -0.11935) (U: 0.61353) (S:  0.49418) (V:  0.5045) 
0712 18:13:44.224802 139690876974848 ../../src/mcts/search.cc:468] b1c1  (24  ) N:    5089 (+ 0) (P:  0.71%) (WL:  0.12556) (D: 0.276) (M: 87.4) (Q:  0.12556) (U: 0.36875) (S:  0.49431) (V:  0.1050) 
0712 18:13:44.224807 139690876974848 ../../src/mcts/search.cc:468] a3a4  (425 ) N:    5444 (+ 0) (P:  0.79%) (WL:  0.11123) (D: 0.280) (M: 83.8) (Q:  0.11123) (U: 0.38311) (S:  0.49433) (V:  0.1702) 
0712 18:13:44.224812 139690876974848 ../../src/mcts/search.cc:468] b1c2  (32  ) N:    5683 (+ 0) (P:  0.81%) (WL:  0.11957) (D: 0.276) (M: 87.6) (Q:  0.11957) (U: 0.37478) (S:  0.49435) (V:  0.1638) 
0712 18:13:44.224818 139690876974848 ../../src/mcts/search.cc:468] c3a2  (471 ) N:    6490 (+ 0) (P:  0.92%) (WL:  0.12017) (D: 0.273) (M: 88.1) (Q:  0.12017) (U: 0.37418) (S:  0.49435) (V:  0.1917) 
0712 18:13:44.224823 139690876974848 ../../src/mcts/search.cc:468] b2a1  (218 ) N:    6496 (+ 0) (P:  0.79%) (WL:  0.17367) (D: 0.266) (M: 83.5) (Q:  0.17367) (U: 0.32061) (S:  0.49428) (V:  0.1980) 
0712 18:13:44.224828 139690876974848 ../../src/mcts/search.cc:468] d2c2  (281 ) N:    8402 (+ 0) (P:  0.93%) (WL:  0.20136) (D: 0.269) (M: 84.1) (Q:  0.20136) (U: 0.29299) (S:  0.49435) (V:  0.1184) 
0712 18:13:44.224833 139690876974848 ../../src/mcts/search.cc:468] c3e2  (475 ) N:    8552 (+ 1) (P:  1.10%) (WL:  0.15679) (D: 0.282) (M: 90.0) (Q:  0.15679) (U: 0.33753) (S:  0.49432) (V:  0.2606) 
0712 18:13:44.224838 139690876974848 ../../src/mcts/search.cc:468] b2c1  (220 ) N:    8998 (+ 0) (P:  1.09%) (WL:  0.17496) (D: 0.281) (M: 86.0) (Q:  0.17496) (U: 0.31939) (S:  0.49435) (V:  0.1252) 
0712 18:13:44.224843 139690876974848 ../../src/mcts/search.cc:468] g1h1  (153 ) N:    9367 (+ 0) (P:  1.14%) (WL:  0.17295) (D: 0.289) (M: 90.8) (Q:  0.17295) (U: 0.32140) (S:  0.49435) (V:  0.2388) 
0712 18:13:44.224848 139690876974848 ../../src/mcts/search.cc:468] d2e2  (282 ) N:   10546 (+ 0) (P:  0.89%) (WL:  0.27074) (D: 0.260) (M: 84.7) (Q:  0.27074) (U: 0.22361) (S:  0.49436) (V:  0.1782) 
0712 18:13:44.224853 139690876974848 ../../src/mcts/search.cc:468] d2e1  (277 ) N:   11054 (+ 0) (P:  1.01%) (WL:  0.25400) (D: 0.261) (M: 85.4) (Q:  0.25400) (U: 0.24035) (S:  0.49435) (V:  0.1162) 
0712 18:13:44.224858 139690876974848 ../../src/mcts/search.cc:468] d2d3  (288 ) N:   13849 (+ 0) (P:  1.11%) (WL:  0.28400) (D: 0.258) (M: 84.9) (Q:  0.28400) (U: 0.21035) (S:  0.49436) (V:  0.2512) 
0712 18:13:44.224863 139690876974848 ../../src/mcts/search.cc:468] g1g3  (160 ) N:   18372 (+ 0) (P:  1.36%) (WL:  0.29873) (D: 0.239) (M: 80.3) (Q:  0.29873) (U: 0.19561) (S:  0.49434) (V:  0.2434) 
0712 18:13:44.224868 139690876974848 ../../src/mcts/search.cc:468] d2c1  (275 ) N:   20871 (+ 0) (P:  1.98%) (WL:  0.24379) (D: 0.254) (M: 84.1) (Q:  0.24379) (U: 0.25056) (S:  0.49435) (V:  0.3416) 
0712 18:13:44.224873 139690876974848 ../../src/mcts/search.cc:468] g1g2  (156 ) N:   22857 (+ 0) (P:  1.54%) (WL:  0.31716) (D: 0.241) (M: 79.9) (Q:  0.31716) (U: 0.17720) (S:  0.49436) (V:  0.3232) 
0712 18:13:44.224878 139690876974848 ../../src/mcts/search.cc:468] d1e1  (75  ) N:   23066 (+ 0) (P:  1.71%) (WL:  0.29936) (D: 0.237) (M: 79.5) (Q:  0.29936) (U: 0.19499) (S:  0.49435) (V:  0.2852) 
0712 18:13:44.224888 139690876974848 ../../src/mcts/search.cc:468] b1a1  (23  ) N:   24942 (+ 0) (P:  2.05%) (WL:  0.27816) (D: 0.253) (M: 82.5) (Q:  0.27816) (U: 0.21619) (S:  0.49435) (V:  0.2916) 
0712 18:13:44.224893 139690876974848 ../../src/mcts/search.cc:468] d2d4  (293 ) N:   28920 (+ 0) (P:  2.71%) (WL:  0.24706) (D: 0.253) (M: 84.6) (Q:  0.24706) (U: 0.24729) (S:  0.49435) (V:  0.3740) 
0712 18:13:44.224898 139690876974848 ../../src/mcts/search.cc:468] d2f4  (295 ) N:   31055 (+ 0) (P:  2.98%) (WL:  0.24167) (D: 0.252) (M: 84.5) (Q:  0.24167) (U: 0.25269) (S:  0.49435) (V:  0.3094) 
0712 18:13:44.224903 139690876974848 ../../src/mcts/search.cc:468] f1e2  (130 ) N:   52695 (+ 0) (P:  5.74%) (WL:  0.20751) (D: 0.261) (M: 84.9) (Q:  0.20751) (U: 0.28685) (S:  0.49436) (V:  0.3983) 
0712 18:13:44.224908 139690876974848 ../../src/mcts/search.cc:468] b1a2  (30  ) N:   62611 (+ 0) (P:  3.68%) (WL:  0.33964) (D: 0.232) (M: 82.2) (Q:  0.33964) (U: 0.15472) (S:  0.49435) (V:  0.3757) 
0712 18:13:44.224914 139690876974848 ../../src/mcts/search.cc:468] d1c1  (74  ) N:   85489 (+ 0) (P:  4.53%) (WL:  0.35480) (D: 0.232) (M: 81.6) (Q:  0.35480) (U: 0.13956) (S:  0.49436) (V:  0.3711) 
0712 18:13:44.224919 139690876974848 ../../src/mcts/search.cc:468] c3a4  (483 ) N:  237142 (+ 0) (P: 20.05%) (WL:  0.27154) (D: 0.238) (M: 90.2) (Q:  0.27154) (U: 0.22282) (S:  0.49436) (V:  0.4794) 
0712 18:13:44.224924 139690876974848 ../../src/mcts/search.cc:468] d2e3  (289 ) N:  317381 (+ 0) (P: 12.60%) (WL:  0.38975) (D: 0.222) (M: 74.3) (Q:  0.38975) (U: 0.10460) (S:  0.49436) (V:  0.4305) 
0712 18:13:44.224929 139690876974848 ../../src/mcts/search.cc:468] f1d3  (134 ) N: 98944088 (+89) (P: 17.46%) (WL:  0.49389) (D: 0.211) (M: 80.7) (Q:  0.49389) (U: 0.00046) (S:  0.49436) (V:  0.4756) 
0712 18:13:44.224934 139690876974848 ../../src/mcts/search.cc:468] node  (  39) N: 100000000 (+90) (P: 100.0%) (WL:  0.49172) (D: 0.211) (M: 81.6) (Q:  0.49172) (V:  0.5125) 

[Thomas Lagershausen]

Gentleman, i repeated the test only doubling the hashsize from 1028 to 2048 MB:

CFish 060720-4core 2048 MB hash:
43/67+ 08:26 1.838.627.508 3.628.108 +1,66 Sc3-d5
48/72 30:52 6.285.947.698 3.393.385 +2,88 Sc3-d5 Lb7xd5

Cfish 110720-4core 2048 MB hash:
47/72 30:52 6.207.978.073 3.351.294 +1,61 Dd2-e3

Once again, the latest version of Stockfish didn´t make me happy.

[zullil]

Gentleman, i repeated the test only doubling the hashsize from 1028 to 2048 MB:

CFish 060720-4core 2048 MB hash:
43/67+ 08:26 1.838.627.508 3.628.108 +1,66 Sc3-d5
48/72 30:52 6.285.947.698 3.393.385 +2,88 Sc3-d5 Lb7xd5

Cfish 110720-4core 2048 MB hash:
47/72 30:52 6.207.978.073 3.351.294 +1,61 Dd2-e3

Once again, the latest version of Stockfish didn´t make me happy.

Here are searches using 2048 MB hash and all other options at default. In particular, one thread. You should be able to exactly reproduce these.

Current Cfish-dev (bench 4366686):

info depth 44 seldepth 70 multipv 1 score cp 113 nodes 892764703 nps 1819436 hashfull 998 tbhits 0 time 490682 pv d2e3 e6e5 f3f4 f6d7 f4e5 d6e5 e3f3 d7c5 c3d5 b7d5 e4d5 e7d6 f1d3 c5d3 f3d3 a8e8 d1e1 d6c5 g1g2 f7f6 f2f3 e8e7 b1a2 c7d6 e1e2 d6d7 g2g4 g8c8 e2c2 d7d6 b2c1 d6c7 d3f5 e7f7 a3a4 c8f8 f5d3 f7e7 g4g6 e5e4 f3e4 c7e5 c2g2 e5e4 d3e4 e7e4
info depth 45 currmove d2e3 currmovenumber 1
info depth 45 currmove f1d3 currmovenumber 2
info depth 45 currmove f1e2 currmovenumber 3
info depth 45 currmove a3a4 currmovenumber 4
info depth 45 currmove g1h1 currmovenumber 5
info depth 45 currmove f1h3 currmovenumber 6
info depth 45 currmove g1g6 currmovenumber 7
info depth 45 currmove d2d4 currmovenumber 8
info depth 45 currmove c3a4 currmovenumber 9
info depth 45 currmove c3d5 currmovenumber 10
info depth 45 seldepth 57 multipv 1 score cp 122 lowerbound nodes 972136579 nps 1820385 hashfull 1000 tbhits 0 time 534028 pv c3d5

Cfish-937c4eb4f6c42197cbbe74c208acacce860d518e (bench 4882833) (This was Cfish-dev on 060720---July 6.)

info depth 43 seldepth 75 multipv 1 score cp 99 nodes 860871841 nps 1733175 hashfull 1000 tbhits 0 time 496702 pv d2e3 e6e5 f1d3 f6d7 f3f4 c7c5 e3c5 d7c5 f4e5 d6e5 d3c2 a8c8 f2f3 c8d8 c3d5 b7d5 e4d5 e7d6 b2c1 c5b7 c1e3 d8c8 g1g4 d6c5 e3c5 c8c5 d5d6 g8d8 d6d7 c5c7 d1g1 h8g8 c2e4 d8d7 g4g7 g8f8 g7h7 d7d6 e4c6 b7c5 h7h8 f8e7 g1e1 c5d7 f3f4 d6c6 b5c6 c7c6 f4e5 c6c4 h8h6 c4c3 h6h8 c3a3
info depth 44 currmove d2e3 currmovenumber 1
info depth 44 currmove d2f4 currmovenumber 2
info depth 44 currmove d2e2 currmovenumber 3
info depth 44 currmove c3a4 currmovenumber 4
info depth 44 currmove f1e2 currmovenumber 5
info depth 44 currmove a3a4 currmovenumber 6
info depth 44 currmove f1d3 currmovenumber 7
info depth 44 currmove g1h1 currmovenumber 8
info depth 44 currmove d2e1 currmovenumber 9
info depth 44 currmove c3d5 currmovenumber 10
info depth 44 seldepth 68 multipv 1 score cp 108 lowerbound nodes 925636149 nps 1738894 hashfull 1000 tbhits 0 time 532313 pv c3d5

So, yes, the current Cfish-dev requires an additional 46500430 nodes to find Nd5. But note that the difference in time for these searches was less than two seconds.