Stockfish 2.2.2 JA update available

[Jouni]

Wow I got 8% speed-up for free Now Rybka 4.1 is sure beaten!

Jouni

[daylen]

Yup, same here.

2.4 GHz Intel Core 2 Duo
Windows 7

SF 2.2.1 JA w32 - 2 threads
SF 2.2.1 JA x64 - 2 threads
SF 2.2.2 JA w32 - 2 threads
SF 2.2.2 JA x64 - 1 thread

Of course, can't run the SSE4.2 compile.

[mcostalba]

Win7. My own compiles all default to 4 with msvc. I'm guessing something didn't translate correctly with GCC.

This is interesting. Now the new super builds of Jim use a different code to detect cpu counts (they must because gcc libraries are different from MSVC)

In particular this is the code:

Code: Select all

/// cpu_count() tries to detect the number of CPU cores

int cpu_count() {

#if defined(_MSC_VER)
  SYSTEM_INFO s;
  GetSystemInfo(&s);
  return std::min(int(s.dwNumberOfProcessors), MAX_THREADS);
#else

#  if defined(_SC_NPROCESSORS_ONLN)
  return std::min((int)sysconf(_SC_NPROCESSORS_ONLN), MAX_THREADS);
#  elif defined(__hpux)
  struct pst_dynamic psd;
  if (pstat_getdynamic(&psd, sizeof(psd), (size_t)1, 0) == -1)
      return 1;
  return std::min((int)psd.psd_proc_cnt, MAX_THREADS);
#  else
  return 1;
#  endif

#endif
}

I suspect _SC_NPROCESSORS_ONLN is not defined in the compile environment and so cpu_count() defaults to "return 1;"

[mcostalba]

For testing purposes, can 2.2.1 and 2.2.2 be combined?
Are the 32-bit compiles changed as well?

For testing purposes the expected improvement of 2.2.2 is far below the error bar of most of the testers, so I'd recommend to do not bother retesting....at least if you are not really getting bored today

[mcostalba]

the offending insruction is

Code: Select all

000000000041C3F0  pabsd       xmm2,xmm2

It is a SSSE3 (note the triple S) introduced in 2006

http://en.wikipedia.org/wiki/SSSE3

On AMD the architectures supporting this extension are: Bobcat and Bulldozer. Other chips are out of luck

[lucasart]

Win7. My own compiles all default to 4 with msvc. I'm guessing something didn't translate correctly with GCC.

This is interesting. Now the new super builds of Jim use a different code to detect cpu counts (they must because gcc libraries are different from MSVC)

In particular this is the code:

Code: Select all

/// cpu_count() tries to detect the number of CPU cores

int cpu_count() {

#if defined(_MSC_VER)
  SYSTEM_INFO s;
  GetSystemInfo(&s);
  return std::min(int(s.dwNumberOfProcessors), MAX_THREADS);
#else

#  if defined(_SC_NPROCESSORS_ONLN)
  return std::min((int)sysconf(_SC_NPROCESSORS_ONLN), MAX_THREADS);
#  elif defined(__hpux)
  struct pst_dynamic psd;
  if (pstat_getdynamic(&psd, sizeof(psd), (size_t)1, 0) == -1)
      return 1;
  return std::min((int)psd.psd_proc_cnt, MAX_THREADS);
#  else
  return 1;
#  endif

#endif
}

I suspect _SC_NPROCESSORS_ONLN is not defined in the compile environment and so cpu_count() defaults to "return 1;"

but it works fine with GCC on Linux. is Jim using mingw ? if so, I don't see why it shouldn't work on Windows

[mcostalba]

but it works fine with GCC on Linux. is Jim using mingw ? if so, I don't see why it shouldn't work on Windows

As I suggested, perhaps he needs to manually define _SC_NPROCESSORS_ONLN in his 64 bit mingw environment while in the 32 bits perhaps it is already defined.

[lucasart]

[quote="mcostalba"]

Code: Select all

#if defined(_MSC_VER)
  SYSTEM_INFO s;
  GetSystemInfo(&s);
  return std::min(int(s.dwNumberOfProcessors), MAX_THREADS);
#else
[/quote]
Why _MSC_VER ? Shouldn't it apply to GCC the same way ? It's just a windows API call that returns the number of CPU, as I understand

[Christian Goralski]

Thank you very much for the new release!

Stockfish 2.2.2 had still the "time bug"!?

In my third 60+15 online game on chessbase server engine lost on time after move 109!

[Event "Wertungspartie, 60 m + 15s"]
[Site "Maschinenraum"]
[Date "2012.01.15"]
[Round "?"]
[White "Kasparito_2, Houdini 2.0b Pro w3"]
[Black "Christian Gorals, Stockfish 2.2.2 JA "]
[Result "1-0"]
[ECO "C65"]
[WhiteElo "2400"]
[BlackElo "2568"]
[Annotator "0.17;0.08"]
[PlyCount "217"]
[EventDate "2012.01.15"]
[TimeControl "3600+15"]

{Houdini 2.0b Pro w32 (4 cores): 25.6 plies; 6.051kN/s Intel(R) Core(TM)
i5-2400S CPU @ 2.50GHz 2494MHz, Hiarcs13hBook.ctg, 8192 MB} 1. e4 {[%eval 0,0]
[%emt 0:00:00]} e5 {[%eval 0,0] [%emt 0:00:00]} 2. Nf3 {[%eval 0,0] [%emt 0:00:
00]} Nc6 {[%eval 0,0] [%emt 0:00:00]} 3. Bb5 {[%eval 0,0] [%emt 0:00:00]} Nf6 {
[%eval 0,0] [%emt 0:00:00]} 4. d3 {[%eval 0,0] [%emt 0:00:00]} Bc5 {[%eval 0,0]
[%emt 0:00:00]} 5. O-O {[%eval 0,0] [%emt 0:00:00]} Nd4 {[%eval 0,0] [%emt 0:
00:00]} 6. Nxd4 {[%eval 0,0] [%emt 0:00:00]} Bxd4 {[%eval 0,0] [%emt 0:00:00]}
7. Nd2 {[%eval 0,0] [%emt 0:00:00]} c6 {[%eval 0,0] [%emt 0:00:00]} 8. Ba4 {
[%eval 0,0] [%emt 0:00:00]} Bb6 {[%eval 0,0] [%emt 0:00:00]} 9. Bb3 {[%eval 17,
24] [%emt 0:02:33]} d6 {[%eval 0,0] [%emt 0:00:00] (0-0)} 10. Nc4 {[%eval 16,
23] [%emt 0:01:57]} Bc7 {[%eval 0,0] [%emt 0:00:00] (Ae6)} 11. Bg5 {[%eval 20,
23] [%emt 0:01:24]} h6 {[%eval 8,28] [%emt 0:01:39]} 12. Bh4 {[%eval 15,24]
[%emt 0:00:19]} Be6 {[%eval 16,29] [%emt 0:03:38]} 13. Qe2 {[%eval 15,24]
[%emt 0:00:12] (Qd2)} Qe7 {[%eval 0,28] [%emt 0:03:39] (g5)} 14. f3 {[%eval 7,
22] [%emt 0:01:54] (a4)} a5 {[%eval 8,27] [%emt 0:01:32] (Dd7)} 15. Ne3 {
[%eval 16,23] [%emt 0:01:19]} a4 {[%eval 16,28] [%emt 0:00:09] (0-0)} 16. Bxe6
{[%eval 19,23] [%emt 0:02:14]} fxe6 {[%eval 12,28] [%emt 0:01:55]} 17. Ng4 {
[%eval 20,24] [%emt 0:00:00] (Rad1)} Bb6+ {[%eval 0,29] [%emt 0:01:32]} 18. Kh1
{[%eval 18,25] [%emt 0:01:25] (Bf2)} g5 {[%eval 0,29] [%emt 0:02:24]} 19. Nxf6+
{[%eval 16,25] [%emt 0:00:38] (Bf2)} Qxf6 {[%eval -48,12] [%emt 0:00:00]} 20.
Bf2 {[%eval 16,25] [%emt 0:02:06]} Bxf2 {[%eval 0,31] [%emt 0:00:02]} 21. Qxf2
{[%eval 16,24] [%emt 0:05:02]} Kd7 {[%eval -8,33] [%emt 0:00:36] (0-0)} 22. c4
{[%eval 16,22] [%emt 0:01:10]} Kc7 {[%eval -8,30] [%emt 0:01:22] (Rc8)} 23.
Rab1 {[%eval 19,24] [%emt 0:01:12] (Rad1)} b6 {[%eval -8,35] [%emt 0:05:42]}
24. b4 {[%eval 16,24] [%emt 0:00:15] (b3)} h5 {[%eval 0,34] [%emt 0:01:43]
(Thf8)} 25. Rbd1 {[%eval 16,22] [%emt 0:01:12] (Qb2)} h4 {[%eval 0,34] [%emt 0:
02:00] (Thd8)} 26. h3 {[%eval 28,23] [%emt 0:02:11] (Kg1)} Rhd8 {[%eval 28,29]
[%emt 0:02:14]} 27. Qc2 {[%eval 25,24] [%emt 0:00:29] (Rfe1)} Qg7 {[%eval 16,
30] [%emt 0:02:06] (Df4)} 28. Rd2 {[%eval 34,24] [%emt 0:01:04]} Qf6 {[%eval
44,31] [%emt 0:02:15] (Rb7)} 29. a3 {[%eval 36,24] [%emt 0:01:08] (Re1)} Qg7 {
[%eval 36,34] [%emt 0:03:10] (d5)} 30. Rfd1 {[%eval 45,26] [%emt 0:01:56]} Qh8
{[%eval 36,34] [%emt 0:01:50]} 31. Qa2 {[%eval 45,27] [%emt 0:00:00] (d4)} Qf8
{[%eval 36,35] [%emt 0:01:17] (Df6)} 32. d4 {[%eval 45,27] [%emt 0:01:44]} exd4
{[%eval 36,36] [%emt 0:00:00]} 33. Rxd4 {[%eval 45,25] [%emt 0:00:01]} Qg7 {
[%eval 36,39] [%emt 0:02:14] (Df6)} 34. Qd2 {[%eval 45,27] [%emt 0:01:17]
(R4d3)} Qe5 {[%eval 36,39] [%emt 0:02:22]} 35. Qc1 {[%eval 45,28] [%emt 0:00:
00] (Rd3)} Rac8 {[%eval 36,35] [%emt 0:01:11] (Dg7)} 36. Qc2 {[%eval 45,27]
[%emt 0:01:03]} Ra8 {[%eval 36,38] [%emt 0:00:23]} 37. R4d2 {[%eval 45,27]
[%emt 0:00:45]} Qg7 {[%eval 36,37] [%emt 0:00:44]} 38. Rd3 {[%eval 45,27]
[%emt 0:00:39] (Rb1)} Qf6 {[%eval 36,38] [%emt 0:01:29] (Dh8)} 39. Qd2 {[%eval
45,26] [%emt 0:00:54] (Qf2)} Qe5 {[%eval 36,38] [%emt 0:01:23] (Df4)} 40. Kg1 {
[%eval 45,25] [%emt 0:01:22] (Qc2)} Rab8 {[%eval 36,34] [%emt 0:01:31] (Td7)}
41. Qc2 {[%eval 45,26] [%emt 0:05:13] (Rb1)} Ra8 {[%eval 36,36] [%emt 0:01:23]}
42. Qb1 {[%eval 45,26] [%emt 0:00:00] (Qa2)} Rac8 {[%eval 36,34] [%emt 0:01:16]
(Td7)} 43. Qc1 {[%eval 45,27] [%emt 0:02:27] (Qa2)} Ra8 {[%eval 36,38] [%emt 0:
01:55]} 44. Qc2 {[%eval 45,27] [%emt 0:00:00]} Qf6 {[%eval 36,38] [%emt 0:02:
09] (Td7)} 45. Qf2 {[%eval 45,26] [%emt 0:00:44] (Qd2)} Rab8 {[%eval 36,36]
[%emt 0:01:20] (De5)} 46. Qe3 {[%eval 45,26] [%emt 0:00:51] (Qc2)} Qe5 {[%eval
36,36] [%emt 0:01:06] (Df4)} 47. Qd2 {[%eval 45,27] [%emt 0:01:11]} Ra8 {
[%eval 36,37] [%emt 0:00:06]} 48. Kh1 {[%eval 45,28] [%emt 0:00:48]} Rab8 {
[%eval 36,36] [%emt 0:00:12] (Df4)} 49. Qc2 {[%eval 45,27] [%emt 0:00:40] (Rb1)
} Ra8 {[%eval 36,34] [%emt 0:01:06]} 50. Qf2 {[%eval 45,29] [%emt 0:02:30]
(Qa2)} Rac8 {[%eval 36,35] [%emt 0:01:21] (Df6)} 51. Kg1 {[%eval 43,23] [%emt
0:02:03] (Rc1)} Qf6 {[%eval 36,33] [%emt 0:02:17] (d5)} 52. Qd2 {[%eval 45,24]
[%emt 0:00:36]} Qe5 {[%eval 36,36] [%emt 0:00:17]} 53. Rc1 {[%eval 45,25]
[%emt 0:00:25] (Qa2)} Qf4 {[%eval 36,35] [%emt 0:01:06]} 54. Qc2 {[%eval 45,26]
[%emt 0:00:00]} Ra8 {[%eval 36,35] [%emt 0:00:41]} 55. Kh1 {[%eval 45,26]
[%emt 0:00:00]} Qf6 {[%eval 36,36] [%emt 0:00:54] (De5)} 56. Qf2 {[%eval 45,26]
[%emt 0:00:48] (Rdd1)} Qe5 {[%eval 36,37] [%emt 0:01:04] (Df4)} 57. Qe2 {
[%eval 45,26] [%emt 0:00:51] (Rdd1)} Qf4 {[%eval 36,35] [%emt 0:00:33] (Rb7)}
58. Qb2 {[%eval 45,27] [%emt 0:00:52] (Rdd1)} Qe5 {[%eval 36,35] [%emt 0:00:28]
} 59. Qa2 {[%eval 45,26] [%emt 0:00:36] (Qf2)} Qf6 {[%eval 36,34] [%emt 0:00:
36] (Rb7)} 60. Rcd1 {[%eval 43,26] [%emt 0:00:36] (Qd2)} Qe5 {[%eval 36,38]
[%emt 0:01:28]} 61. Qc2 {[%eval -7,25] [%emt 0:01:34]} Qf6 {[%eval 0,47] [%emt
0:00:00] (Dg7)} 62. Qb1 {[%eval 39,24] [%emt 0:00:30] (Qf2)} Qh8 {[%eval 36,33]
[%emt 0:00:25] (Df4)} 63. Rc1 {[%eval 44,24] [%emt 0:00:55] (Re1)} Qe5 {[%eval
36,35] [%emt 0:00:37] (Dg7)} 64. Rf1 {[%eval 40,23] [%emt 0:00:30] (Re1)} Rf8 {
[%eval 36,35] [%emt 0:00:30] (Tac8)} 65. Rfd1 {[%eval 44,25] [%emt 0:01:00]}
Rfd8 {[%eval 36,39] [%emt 0:00:00]} 66. Qc1 {[%eval 41,24] [%emt 0:00:37] (Qa2)
} Qf4 {[%eval 36,36] [%emt 0:00:24] (Tac8)} 67. R1d2 {[%eval 45,25] [%emt 0:00:
41] (Qa1)} Qe5 {[%eval 36,32] [%emt 0:00:23] (Df6)} 68. b5 {[%eval 39,25]
[%emt 0:01:02] (Qe1)} c5 {[%eval 0,35] [%emt 0:00:46]} 69. Qd1 {[%eval 39,29]
[%emt 0:00:00] (Rd1)} Rd7 {[%eval 0,36] [%emt 0:00:28]} 70. Kg1 {[%eval 39,28]
[%emt 0:00:49]} Rdd8 {[%eval 0,39] [%emt 0:00:00] (Ta7)} 71. Qc1 {[%eval 39,27]
[%emt 0:00:37] (Kh1)} Rd7 {[%eval 0,40] [%emt 0:00:25] (Rb7)} 72. Qc2 {[%eval
39,26] [%emt 0:00:25] (Qd1)} Rdd8 {[%eval 0,35] [%emt 0:00:26] (Rb7)} 73. Rd1 {
[%eval 36,27] [%emt 0:00:53] (Qb1) Christian Gorals,Stockfish 2.2.2 JA S
bietet Remis} Rd7 {[%eval 0,36] [%emt 0:00:25]} 74. Rb1 {[%eval 36,26] [%emt 0:
00:11] (Kf2)} Rdd8 {[%eval 0,36] [%emt 0:00:21]} 75. Qd2 {[%eval 36,26] [%emt
0:00:08] (Rbd1)} Rd7 {[%eval 0,36] [%emt 0:00:13]} 76. Rc1 {[%eval 36,26]
[%emt 0:00:32] (Rd1)} Rb8 {[%eval 0,34] [%emt 0:00:12] (Df4)} 77. Qc2 {[%eval
36,26] [%emt 0:00:35]} Ra8 {[%eval 0,40] [%emt 0:00:00]} 78. Rcd1 {[%eval 36,
26] [%emt 0:00:31]} Qf6 {[%eval 0,37] [%emt 0:00:00] (Df4)} 79. Qd2 {[%eval 36,
27] [%emt 0:00:24]} Qe5 {[%eval 0,42] [%emt 0:00:03] (Df4)} 80. Rb1 {[%eval 36,
24] [%emt 0:01:05] (Qc2)} Rg8 {[%eval 0,36] [%emt 0:00:20] (Rb7)} 81. Qc2 {
[%eval 36,24] [%emt 0:00:28] (Rd1)} Ra8 {[%eval 0,37] [%emt 0:00:18]} 82. Rf1 {
[%eval 36,26] [%emt 0:00:14] (Rbd1)} Rdd8 {[%eval 0,35] [%emt 0:00:21]} 83. Qd2
{[%eval 36,27] [%emt 0:00:06] (Rfd1)} Rd7 {[%eval 0,36] [%emt 0:00:38]} 84. Re1
{[%eval 36,27] [%emt 0:00:01] (Rb1)} Rdd8 {[%eval 0,39] [%emt 0:00:18] (Tad8)}
85. Kh1 {[%eval 36,28] [%emt 0:00:29] (Rb1)} Rf8 {[%eval 0,37] [%emt 0:00:18]
(Rb8)} 86. Rd1 {[%eval 36,30] [%emt 0:01:28]} Rfd8 {[%eval 0,42] [%emt 0:00:00]
} 87. Qc2 {[%eval 36,29] [%emt 0:00:47] (Re1)} Rd7 {[%eval 0,42] [%emt 0:00:33]
} 88. R3d2 {[%eval 36,28] [%emt 0:01:15] (Kg1)} Kb7 {[%eval 0,36] [%emt 0:00:
14] (Rb8)} 89. Kg1 {[%eval 36,26] [%emt 0:00:23]} Kc8 {[%eval 0,36] [%emt 0:00:
01] (Rb8)} 90. Rf2 {[%eval 36,26] [%emt 0:00:26] (Kf2)} Kb7 {[%eval 0,39]
[%emt 0:00:15]} 91. Rd3 {[%eval 36,26] [%emt 0:00:19] (Rfd2)} Qa1+ {[%eval 0,
31] [%emt 0:00:15] (Rb8)} 92. Rf1 {[%eval 36,26] [%emt 0:00:23]} Qe5 {[%eval 0,
36] [%emt 0:00:00] (Dg7)} 93. Qd1 {[%eval 36,26] [%emt 0:00:53] (Rdd1)} Kc7 {
[%eval 0,37] [%emt 0:00:18] (Df4)} 94. Rf2 {[%eval 36,25] [%emt 0:00:19] (Qc2)}
Kb7 {[%eval 0,38] [%emt 0:00:30] (Rd8)} 95. Rfd2 {[%eval 29,25] [%emt 0:00:17]
(Qc2)} Kc7 {[%eval 0,39] [%emt 0:00:15]} 96. Kh1 {[%eval 29,26] [%emt 0:00:11]
(Rf2)} Rdd8 {[%eval 0,38] [%emt 0:00:19] (Ta5)} 97. Qc2 {[%eval 29,25] [%emt 0:
00:17] (Kg1)} Kb7 {[%eval 0,37] [%emt 0:00:24] (Td7)} 98. Kg1 {[%eval -7,27]
[%emt 0:01:04] (Qd1)} Kc7 {[%eval 0,39] [%emt 0:00:15]} 99. Qb1 {[%eval 26,25]
[%emt 0:00:07] (Qd1)} Rd7 {[%eval 0,40] [%emt 0:00:16]} 100. Rd1 {[%eval 26,25]
[%emt 0:00:05] (Qd1)} Rdd8 {[%eval 0,35] [%emt 0:00:10] (Rb8)} 101. Qa2 {
[%eval 26,26] [%emt 0:00:19] (Qc2)} Qf6 {[%eval 0,36] [%emt 0:00:16] (Td7)}
102. Qd2 {[%eval 26,26] [%emt 0:00:16]} Qe5 {[%eval 0,36] [%emt 0:00:03]} 103.
Qe2 {[%eval 26,25] [%emt 0:00:30] (Qc2)} Kb7 {[%eval 0,35] [%emt 0:00:20] (Td7)
} 104. Qc2 {[%eval 26,26] [%emt 0:00:21] (Qd2)} Kc7 {[%eval 0,39] [%emt 0:00:
18] (Df4)} 105. Kf2 {[%eval 21,23] [%emt 0:00:16] (Qf2)} Qg3+ {[%eval 0,37]
[%emt 0:00:17] (Td7)} 106. Kf1 {[%eval 21,26] [%emt 0:00:15]} Qe5 {[%eval 0,39]
[%emt 0:00:00] (Td7)} 107. Qd2 {[%eval 17,24] [%emt 0:00:19] (R3d2)} Rd7 {
[%eval 0,37] [%emt 0:00:17]} 108. Re1 {[%eval 17,24] [%emt 0:00:23] (Qc2)} Rf7
{[%eval 0,37] [%emt 0:00:23] (Df4)} 109. Kg1 {[%eval 16,25] [%emt 0:00:14]
(Rd1) Zeit} 1-0

Sincerely Christian Goralski

[Har]

Yes,very bad time management also with increase.