12cpu-60+0 tournament

Hugo · Post by **Hugo** » Wed Dec 28, 2016 2:00 pm

My new tournament is on the way.

All engines use 12cpu and 4GB hash.
52 out of 396 games played

60 Minutes/Game  2016

                                   1 2 3 4 5 6 7 8 9 0 1 2 
1   Komodo 10.3 64-bit C0x12       * ½ 1   ½ 1 1 ½ 1 1   ½   7.0/9
2   Stockfish 261216 64 POPCNT x1  ½ *   ½   ½ 1 ½ ½ 1 1     5.5/8  22.50
3   Houdini 5.01 Pro x64-popc x12  0   * 1 ½ ½ 1 1 1   ½     5.5/8  21.75
4   Fire 5 x64 x12                   ½ 0 * ½ ½   ½ ½ ½ 1 1   5.0/9
5   Booot 6.1_x64_popcnt x12       ½   ½ ½ * 0 ½   ½ ½ ½ ½   4.0/9  17.50
6   Deep Shredder 161006 x64 x12   0 ½ ½ ½ 1 *     0 ½ ½ ½   4.0/9  17.00
7   Nirvanachess 2.3 x12           0 0 0   ½   * 1 ½ ½ 1 ½   4.0/9  14.00
8   Andscacs 0.89 x12              ½ ½ 0 ½     0 * ½ ½ ½ ½   3.5/9  15.50
9   Jonny 8.00 x12                 0 ½ 0 ½ ½ 1 ½ ½ *         3.5/8  15.00
10  Ginkgo 1.9m x12                0 0   ½ ½ ½ ½ ½   * ½ ½   3.5/9  13.50
11  Fizbo 1.8 x12                    0 ½ 0 ½ ½ 0 ½   ½ * 1   3.5/9  13.25
12  Hannibal 1.7 x64 x12           ½     0 ½ ½ ½ ½   ½ 0 *   3.0/8

pity, Houdini 5.1 lost again on time in won game

what shall I do

wonderfull game btw., from +1.44 up to -5.21. A bug for Komodo? no tbs used.

[pgn][Event "60 Minutes/Game"]
[Site "EndOf2016, DESKTOP-8B4SES4"]
[Date "2016.12.28"]
[Round "1.7"]
[White "Komodo 10.3 64-bit C0x12"]
[Black "Houdini 5.01 Pro x64-popc x12"]
[Result "1-0"]
[PlyCount "83"]
[EventDate "2016.??.??"]

1. e4 {book 0s} e6 {book 0s} 2. d4 {book 0s} b6 {book 0s} 3. Nf3 {
+0.47/32 1:39m} Bb7 {book 0s} 4. Bd3 {+0.50/33 1:19m} d5 {book 0s (c5)} 5. exd5
{+0.65/31 51s} Qxd5 {book 0s} 6. c4 {+0.62/33 2:31m} Qd8 {+0.67/27 2:27m} 7.
O-O {+0.62/32 1:29m} Bd6 {+0.67/26 1:23m} 8. Nc3 {+0.52/33 3:08m (Re1)} Nf6 {
+0.57/25 1:43m} 9. d5 {+0.70/31 30s (Re1)} O-O {+0.50/26 3:32m} 10. dxe6 {
+0.66/31 43s} fxe6 {+0.52/24 16s} 11. Ng5 {+0.67/33 2:03m} Qd7 {+0.51/25 52s}
12. Re1 {+0.70/34 48s} Na6 {+0.64/27 2:49m} 13. Nxe6 {+0.51/35 1:18m} Rfe8 {
+0.64/25 0s} 14. Bf5 {+0.45/37 1:02m} Nc5 {+0.63/28 1:01m} 15. Nxc5 {
+0.49/39 56s} Qxf5 {+0.84/27 14s} 16. Rxe8+ {+0.49/41 1:47m} Rxe8 {+0.84/26 0s}
17. Nxb7 {+0.48/41 39s} Ng4 {+0.73/29 49s} 18. Be3 {+0.66/40 1:35m} Rxe3 {
+0.80/30 4:35m} 19. Qd5+ {+0.63/43 1:16m} Qxd5 {+0.71/32 1:09m} 20. cxd5 {
+0.69/42 51s} Bxh2+ {+0.78/35 1:16m} 21. Kh1 {+0.63/43 25s} Re8 {+0.78/34 14s}
22. f3 {+0.69/43 35s} Bf4 {+0.82/35 1:09m (Rb8)} 23. fxg4 {+0.68/41 29s} Rb8 {
+0.82/34 0s} 24. Rf1 {+0.64/43 1:00m} g5 {+0.89/36 1:34m} 25. Ne2 {+0.80/39 44s
} Rxb7 {+0.89/34 14s} 26. Nxf4 {+0.75/39 23s} gxf4 {+0.89/33 0s} 27. Rxf4 {
+0.80/41 35s} a5 {+0.90/34 1:23m} 28. Re4 {+0.81/38 21s} Rb8 {+0.96/28 11s} 29.
Kh2 {+0.81/39 27s} Rd8 {+0.83/29 42s} 30. Re5 {+0.81/37 27s} a4 {+0.84/31 1:41m
} 31. Kg3 {+0.84/35 24s} Ra8 {+0.76/29 58s} 32. Kf4 {+0.77/37 1:03m} Ra5 {
+0.95/28 13s} 33. b3 {+0.90/39 55s (b4)} axb3 {+0.74/29 32s} 34. axb3 {
+0.92/39 27s} Rb5 {+0.65/24 9s (Kf7)} 35. Kf5 {+1.80/43 1:54m (Re8+)} Kf7 {
0.00/35 36s} 36. d6 {+1.90/40 18s (Kg5)} Rxe5+ {0.00/50 53s} 37. Kxe5 {
+1.62/43 3:11m} cxd6+ {0.00/50 32s} 38. Ke4 {+1.44/40 5:33m (Kf5)} Ke6 {
-5.21/23 30s} 39. g5 {-M40/35 21s} d5+ {-6.47/21 30s} 40. Kd4 {-M37/38 3s (Kf4)
} Kd6 {-11.03/21 31s} 41. b4 {-M34/38 3s} b5 {-95.76/27 41s (Ke6)} 42. g4 {
-M26/37 3s (g3) time} 1-0[/pgn]

Guenther · Post by **Guenther** » Wed Dec 28, 2016 2:16 pm

Hugo wrote: ...

pity, Houdini 5.1 lost again on time in won game :(
what shall I do :(
wonderfull game btw., from +1.44 up to -5.21. A bug for Komodo? no tbs used.

[pgn][Event "60 Minutes/Game"]
[Site "EndOf2016, DESKTOP-8B4SES4"]
[Date "2016.12.28"]
[Round "1.7"]
[White "Komodo 10.3 64-bit C0x12"]
[Black "Houdini 5.01 Pro x64-popc x12"]
[Result "1-0"]
[PlyCount "83"]
[EventDate "2016.??.??"]

1. e4 {book 0s} e6 {book 0s} 2. d4 {book 0s} b6 {book 0s} 3. Nf3 {
+0.47/32 1:39m} Bb7 {book 0s} 4. Bd3 {+0.50/33 1:19m} d5 {book 0s (c5)} 5. exd5
{+0.65/31 51s} Qxd5 {book 0s} 6. c4 {+0.62/33 2:31m} Qd8 {+0.67/27 2:27m} 7.
O-O {+0.62/32 1:29m} Bd6 {+0.67/26 1:23m} 8. Nc3 {+0.52/33 3:08m (Re1)} Nf6 {
+0.57/25 1:43m} 9. d5 {+0.70/31 30s (Re1)} O-O {+0.50/26 3:32m} 10. dxe6 {
+0.66/31 43s} fxe6 {+0.52/24 16s} 11. Ng5 {+0.67/33 2:03m} Qd7 {+0.51/25 52s}
12. Re1 {+0.70/34 48s} Na6 {+0.64/27 2:49m} 13. Nxe6 {+0.51/35 1:18m} Rfe8 {
+0.64/25 0s} 14. Bf5 {+0.45/37 1:02m} Nc5 {+0.63/28 1:01m} 15. Nxc5 {
+0.49/39 56s} Qxf5 {+0.84/27 14s} 16. Rxe8+ {+0.49/41 1:47m} Rxe8 {+0.84/26 0s}
17. Nxb7 {+0.48/41 39s} Ng4 {+0.73/29 49s} 18. Be3 {+0.66/40 1:35m} Rxe3 {
+0.80/30 4:35m} 19. Qd5+ {+0.63/43 1:16m} Qxd5 {+0.71/32 1:09m} 20. cxd5 {
+0.69/42 51s} Bxh2+ {+0.78/35 1:16m} 21. Kh1 {+0.63/43 25s} Re8 {+0.78/34 14s}
22. f3 {+0.69/43 35s} Bf4 {+0.82/35 1:09m (Rb8)} 23. fxg4 {+0.68/41 29s} Rb8 {
+0.82/34 0s} 24. Rf1 {+0.64/43 1:00m} g5 {+0.89/36 1:34m} 25. Ne2 {+0.80/39 44s
} Rxb7 {+0.89/34 14s} 26. Nxf4 {+0.75/39 23s} gxf4 {+0.89/33 0s} 27. Rxf4 {
+0.80/41 35s} a5 {+0.90/34 1:23m} 28. Re4 {+0.81/38 21s} Rb8 {+0.96/28 11s} 29.
Kh2 {+0.81/39 27s} Rd8 {+0.83/29 42s} 30. Re5 {+0.81/37 27s} a4 {+0.84/31 1:41m
} 31. Kg3 {+0.84/35 24s} Ra8 {+0.76/29 58s} 32. Kf4 {+0.77/37 1:03m} Ra5 {
+0.95/28 13s} 33. b3 {+0.90/39 55s (b4)} axb3 {+0.74/29 32s} 34. axb3 {
+0.92/39 27s} Rb5 {+0.65/24 9s (Kf7)} 35. Kf5 {+1.80/43 1:54m (Re8+)} Kf7 {
0.00/35 36s} 36. d6 {+1.90/40 18s (Kg5)} Rxe5+ {0.00/50 53s} 37. Kxe5 {
+1.62/43 3:11m} cxd6+ {0.00/50 32s} 38. Ke4 {+1.44/40 5:33m (Kf5)} Ke6 {
-5.21/23 30s} 39. g5 {-M40/35 21s} d5+ {-6.47/21 30s} 40. Kd4 {-M37/38 3s (Kf4)
} Kd6 {-11.03/21 31s} 41. b4 {-M34/38 3s} b5 {-95.76/27 41s (Ke6)} 42. g4 {
-M26/37 3s (g3) time} 1-0[/pgn]

38. Ke4??? really is a bummer...

Carlos777 · Post by **Carlos777** » Wed Dec 28, 2016 4:11 pm

Hugo wrote:My new tournament is on the way.

All engines use 12cpu and 4GB hash.
52 out of 396 games played

Code: Select all

60 Minutes/Game  2016

                                   1 2 3 4 5 6 7 8 9 0 1 2 
1   Komodo 10.3 64-bit C0x12       * ½ 1   ½ 1 1 ½ 1 1   ½   7.0/9
2   Stockfish 261216 64 POPCNT x1  ½ *   ½   ½ 1 ½ ½ 1 1     5.5/8  22.50
3   Houdini 5.01 Pro x64-popc x12  0   * 1 ½ ½ 1 1 1   ½     5.5/8  21.75
4   Fire 5 x64 x12                   ½ 0 * ½ ½   ½ ½ ½ 1 1   5.0/9
5   Booot 6.1_x64_popcnt x12       ½   ½ ½ * 0 ½   ½ ½ ½ ½   4.0/9  17.50
6   Deep Shredder 161006 x64 x12   0 ½ ½ ½ 1 *     0 ½ ½ ½   4.0/9  17.00
7   Nirvanachess 2.3 x12           0 0 0   ½   * 1 ½ ½ 1 ½   4.0/9  14.00
8   Andscacs 0.89 x12              ½ ½ 0 ½     0 * ½ ½ ½ ½   3.5/9  15.50
9   Jonny 8.00 x12                 0 ½ 0 ½ ½ 1 ½ ½ *         3.5/8  15.00
10  Ginkgo 1.9m x12                0 0   ½ ½ ½ ½ ½   * ½ ½   3.5/9  13.50
11  Fizbo 1.8 x12                    0 ½ 0 ½ ½ 0 ½   ½ * 1   3.5/9  13.25
12  Hannibal 1.7 x64 x12           ½     0 ½ ½ ½ ½   ½ 0 *   3.0/8

pity, Houdini 5.1 lost again on time in won game

what shall I do

wonderfull game btw., from +1.44 up to -5.21. A bug for Komodo? no tbs used.

Thanks for the tournament, but this is one of the reasons I don't like sudden death time controls. Besides the occasional losing on time, it is also some bad endgame moves when the engines are in time trouble. It would be better something like 55' + 5" or 45' + 15".

Jouni · Post by **Jouni** » Wed Dec 28, 2016 6:17 pm

Something is TERRIBLY WRONG, if Houdini loses by time at move 42! How can it play 200 move game? I haven't noticed any time losses above 10s + 0,1s level (even there maybe one game in 100). Have You tried to increase Move Overhead, when default is 0ms?

carldaman · Post by **carldaman** » Wed Dec 28, 2016 8:57 pm

Carlos777 wrote:
Hugo wrote:My new tournament is on the way.

All engines use 12cpu and 4GB hash.
52 out of 396 games played
Code: Select all
60 Minutes/Game  2016

                                   1 2 3 4 5 6 7 8 9 0 1 2 
1   Komodo 10.3 64-bit C0x12       * ½ 1   ½ 1 1 ½ 1 1   ½   7.0/9
2   Stockfish 261216 64 POPCNT x1  ½ *   ½   ½ 1 ½ ½ 1 1     5.5/8  22.50
3   Houdini 5.01 Pro x64-popc x12  0   * 1 ½ ½ 1 1 1   ½     5.5/8  21.75
4   Fire 5 x64 x12                   ½ 0 * ½ ½   ½ ½ ½ 1 1   5.0/9
5   Booot 6.1_x64_popcnt x12       ½   ½ ½ * 0 ½   ½ ½ ½ ½   4.0/9  17.50
6   Deep Shredder 161006 x64 x12   0 ½ ½ ½ 1 *     0 ½ ½ ½   4.0/9  17.00
7   Nirvanachess 2.3 x12           0 0 0   ½   * 1 ½ ½ 1 ½   4.0/9  14.00
8   Andscacs 0.89 x12              ½ ½ 0 ½     0 * ½ ½ ½ ½   3.5/9  15.50
9   Jonny 8.00 x12                 0 ½ 0 ½ ½ 1 ½ ½ *         3.5/8  15.00
10  Ginkgo 1.9m x12                0 0   ½ ½ ½ ½ ½   * ½ ½   3.5/9  13.50
11  Fizbo 1.8 x12                    0 ½ 0 ½ ½ 0 ½   ½ * 1   3.5/9  13.25
12  Hannibal 1.7 x64 x12           ½     0 ½ ½ ½ ½   ½ 0 *   3.0/8
pity, Houdini 5.1 lost again on time in won game
what shall I do
wonderfull game btw., from +1.44 up to -5.21. A bug for Komodo? no tbs used.
Thanks for the tournament, but this is one of the reasons I don't like sudden death time controls. Besides the occasional losing on time, it is also some bad endgame moves when the engines are in time trouble. It would be better something like 55' + 5" or 45' + 15".

+1 Couldn't agree more.

Cheers,
CL

Hugo · Post by **Hugo** » Wed Dec 28, 2016 9:14 pm

Jouni wrote:Something is TERRIBLY WRONG, if Houdini loses by time at move 42! How can it play 200 move game? I haven't noticed any time losses above 10s + 0,1s level (even there maybe one game in 100). Have You tried to increase Move Overhead, when default is 0ms?

It is not the first time that this happend with Houdini 5.
It also has nothing to do with the time controlls. It is a bug that also happend in games with increment.
Probably something in the communication between engine and GUI.

And my oppinion about sudden death games:
Why should engine not be able to play them?
And I doubt that the endgame is that bad in 60 Min game.

So sometimes i use increment, sometimes not.

Jouni · Post by **Jouni** » Wed Dec 28, 2016 10:22 pm

Bug in Shredder GUI? Houdini has a log file, which can inform about error.

mjlef · Post by **mjlef** » Wed Dec 28, 2016 10:31 pm

Guenther wrote:
Hugo wrote: ...

pity, Houdini 5.1 lost again on time in won game
what shall I do
wonderfull game btw., from +1.44 up to -5.21. A bug for Komodo? no tbs used.

[pgn][Event "60 Minutes/Game"]
[Site "EndOf2016, DESKTOP-8B4SES4"]
[Date "2016.12.28"]
[Round "1.7"]
[White "Komodo 10.3 64-bit C0x12"]
[Black "Houdini 5.01 Pro x64-popc x12"]
[Result "1-0"]
[PlyCount "83"]
[EventDate "2016.??.??"]

1. e4 {book 0s} e6 {book 0s} 2. d4 {book 0s} b6 {book 0s} 3. Nf3 {
+0.47/32 1:39m} Bb7 {book 0s} 4. Bd3 {+0.50/33 1:19m} d5 {book 0s (c5)} 5. exd5
{+0.65/31 51s} Qxd5 {book 0s} 6. c4 {+0.62/33 2:31m} Qd8 {+0.67/27 2:27m} 7.
O-O {+0.62/32 1:29m} Bd6 {+0.67/26 1:23m} 8. Nc3 {+0.52/33 3:08m (Re1)} Nf6 {
+0.57/25 1:43m} 9. d5 {+0.70/31 30s (Re1)} O-O {+0.50/26 3:32m} 10. dxe6 {
+0.66/31 43s} fxe6 {+0.52/24 16s} 11. Ng5 {+0.67/33 2:03m} Qd7 {+0.51/25 52s}
12. Re1 {+0.70/34 48s} Na6 {+0.64/27 2:49m} 13. Nxe6 {+0.51/35 1:18m} Rfe8 {
+0.64/25 0s} 14. Bf5 {+0.45/37 1:02m} Nc5 {+0.63/28 1:01m} 15. Nxc5 {
+0.49/39 56s} Qxf5 {+0.84/27 14s} 16. Rxe8+ {+0.49/41 1:47m} Rxe8 {+0.84/26 0s}
17. Nxb7 {+0.48/41 39s} Ng4 {+0.73/29 49s} 18. Be3 {+0.66/40 1:35m} Rxe3 {
+0.80/30 4:35m} 19. Qd5+ {+0.63/43 1:16m} Qxd5 {+0.71/32 1:09m} 20. cxd5 {
+0.69/42 51s} Bxh2+ {+0.78/35 1:16m} 21. Kh1 {+0.63/43 25s} Re8 {+0.78/34 14s}
22. f3 {+0.69/43 35s} Bf4 {+0.82/35 1:09m (Rb8)} 23. fxg4 {+0.68/41 29s} Rb8 {
+0.82/34 0s} 24. Rf1 {+0.64/43 1:00m} g5 {+0.89/36 1:34m} 25. Ne2 {+0.80/39 44s
} Rxb7 {+0.89/34 14s} 26. Nxf4 {+0.75/39 23s} gxf4 {+0.89/33 0s} 27. Rxf4 {
+0.80/41 35s} a5 {+0.90/34 1:23m} 28. Re4 {+0.81/38 21s} Rb8 {+0.96/28 11s} 29.
Kh2 {+0.81/39 27s} Rd8 {+0.83/29 42s} 30. Re5 {+0.81/37 27s} a4 {+0.84/31 1:41m
} 31. Kg3 {+0.84/35 24s} Ra8 {+0.76/29 58s} 32. Kf4 {+0.77/37 1:03m} Ra5 {
+0.95/28 13s} 33. b3 {+0.90/39 55s (b4)} axb3 {+0.74/29 32s} 34. axb3 {
+0.92/39 27s} Rb5 {+0.65/24 9s (Kf7)} 35. Kf5 {+1.80/43 1:54m (Re8+)} Kf7 {
0.00/35 36s} 36. d6 {+1.90/40 18s (Kg5)} Rxe5+ {0.00/50 53s} 37. Kxe5 {
+1.62/43 3:11m} cxd6+ {0.00/50 32s} 38. Ke4 {+1.44/40 5:33m (Kf5)} Ke6 {
-5.21/23 30s} 39. g5 {-M40/35 21s} d5+ {-6.47/21 30s} 40. Kd4 {-M37/38 3s (Kf4)
} Kd6 {-11.03/21 31s} 41. b4 {-M34/38 3s} b5 {-95.76/27 41s (Ke6)} 42. g4 {
-M26/37 3s (g3) time} 1-0[/pgn]
38. Ke4??? really is a bummer...

I cannot get Komodo 10.3 to reproduce 38. Ke4 here. Next time I have the 12 core machine free I can try it on that position. Such along time for that move should not have been possible at these time controls. If I find something I will let you know.

pohl4711 · Post by **pohl4711** » Wed Dec 28, 2016 10:44 pm

Jouni wrote:Bug in Shredder GUI? Houdini has a log file, which can inform about error.

I think so, too. I use Houdini 5 with bullet-speed (70''+700ms, singlecore) in the LittleBlitzerGUI: 13000 games, no crash, no timeloss.
And I use Houdini 5 in my long thinkingtime tournament (20'+5', 7 hyperthreading threads on a Quadcore Skylake CPU Notebook, no draw, no resign by GUI) with the Fritz13 GUI: no crash, no timeloss in more than 120 games.

Stefan (SPCC)

Hugo · Post by **Hugo** » Thu Dec 29, 2016 10:12 am

pohl4711 wrote:
Jouni wrote:Bug in Shredder GUI? Houdini has a log file, which can inform about error.
I think so, too. I use Houdini 5 with bullet-speed (70''+700ms, singlecore) in the LittleBlitzerGUI: 13000 games, no crash, no timeloss.
And I use Houdini 5 in my long thinkingtime tournament (20'+5', 7 hyperthreading threads on a Quadcore Skylake CPU Notebook, no draw, no resign by GUI) with the Fritz13 GUI: no crash, no timeloss in more than 120 games.

Stefan (SPCC)

the diffenece might be that i am usig a dual socket machine and 12 cpus for the engine, while u use just 1 and a onesocket machine.
On my onesocket pcs and with the Houdini standard version I had no problems at all.

C.K.

12cpu-60+0 tournament

12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament

Re: 12cpu-60+0 tournament