Quark test and a gentle hint....

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 25 July 2004 12:59:49:

Hello all,
I just tested Quark 2.35 at 1 minute and 1 second.
Winboard gui, WBTM was the tm. 4 man egtbs loaded and used.
800 mhz and 512 mb of ram. Hey, it is my slow machine, leave it alone.
Not one single game lost on time despite the hardware.
I am thinking that before concluding that the engine has a bug, wouldn't
it be good to test it under its native gui to see if the problem can be duplicated?
Good as in good scientific method I mean.
I have used Quark as a sparring partner for a number of engines I test
regularly including the Baron at 5 minutes and 5 seconds. I have never had a
loss on time at that timecontrol in Winboard.
I wonder .....
Later.


Quark 2.35 Paderborn : 2405 94 (+ 37,= 22,- 35), 51.1 %
Baron 1.4.0 : 5 (+ 4,= 0,- 1), 80.0 %
Baron 1.3.1 : 5 (+ 4,= 1,- 0), 90.0 %
Baron 1.2.1 : 5 (+ 3,= 1,- 1), 70.0 %
Baron 1.0.6 : 5 (+ 2,= 1,- 2), 50.0 %
Baron 0.99.5a : 5 (+ 2,= 0,- 3), 40.0 %
Amateur 2.70 : 5 (+ 3,= 1,- 1), 70.0 %
Amyan 1.592 : 5 (+ 1,= 3,- 1), 50.0 %
AnMon 5.30 : 5 (+ 1,= 1,- 3), 30.0 %
Arasan 7.4 : 5 (+ 0,= 3,- 2), 30.0 %
Aristarch 4.50 : 4 (+ 2,= 2,- 0), 75.0 %
Comet B.62-3 : 5 (+ 3,= 2,- 0), 80.0 %
Crafty 19.15 : 5 (+ 2,= 0,- 3), 40.0 %
Green Light Chess 3.0 : 5 (+ 1,= 0,- 4), 20.0 %
Gothmog 0.4.8 : 5 (+ 2,= 1,- 2), 50.0 %
Little Goliath 2000 v3.9 : 5 (+ 1,= 1,- 3), 30.0 %
Pepito 1.59 : 5 (+ 0,= 2,- 3), 20.0 %
Phalanx XXII : 5 (+ 2,= 0,- 3), 40.0 %
Pharaon 2.62 : 5 (+ 2,= 1,- 2), 50.0 %
Quark 2.05 beta : 5 (+ 2,= 2,- 1), 60.0 %

[Thomas Mayer]

Geschrieben von:/Posted by: Thomas Mayer at 25 July 2004 13:32:03:
Als Antwort auf:/In reply to: Quark test and a gentle hint.... geschrieben von:/posted by: Roger Brown at 25 July 2004 12:59:49:

Hi Roger,
first, thanks for testing... For this fast time control I think the result is really good for Quark... OTOH it seems that some of the changes made especially before WCCC 2003 did help it a bit to improve it's blitz strength...
By the way:

800 mhz and 512 mb of ram. Hey, it is my slow machine, leave it alone.

this is not what I had in mind thinking about a slow machine... I thought more about something like Pentium 75 MHz or slower... -> Because Quarks time management does adjust itself according to the speed of the PC, e.g. it checks every second if time for the move is over when it has still more then a second to think about the move... This is reduced when time is slow and the smallest number is, that it checks every 1000 positions... On slow hardware which might do below 10.000 positions a second this could mean that it checks only every 10th of a second... Combined with the time the GUI consumes this could cause some problems... (On my Dual 1800+ which plays sometimes on playchess in the human room Quark can theoretically make around 500 moves a second, but I have seen that not more then 150-200 moves are possible in a 1 0 game because the communication needs to much time...)
And of course you are right, Quark is built with WinBoard in mine, so at least in WinBoard it should work at it's best...
Greets, Thomas

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 25 July 2004 16:15:58:
Als Antwort auf:/In reply to: Re: Quark test and a gentle hint.... geschrieben von:/posted by: Thomas Mayer at 25 July 2004 13:32:03:

Hi Roger,
first, thanks for testing... For this fast time control I think the result is really good for Quark... OTOH it seems that some of the changes made especially before WCCC 2003 did help it a bit to improve it's blitz strength...
By the way:

800 mhz and 512 mb of ram. Hey, it is my slow machine, leave it alone.

this is not what I had in mind thinking about a slow machine... I thought more about something like Pentium 75 MHz or slower... -> Because Quarks time management does adjust itself according to the speed of the PC, e.g. it checks every second if time for the move is over when it has still more then a second to think about the move... This is reduced when time is slow and the smallest number is, that it checks every 1000 positions... On slow hardware which might do below 10.000 positions a second this could mean that it checks only every 10th of a second... Combined with the time the GUI consumes this could cause some problems... (On my Dual 1800+ which plays sometimes on playchess in the human room Quark can theoretically make around 500 moves a second, but I have seen that not more then 150-200 moves are possible in a 1 0 game because the communication needs to much time...)
And of course you are right, Quark is built with WinBoard in mine, so at least in WinBoard it should work at it's best...
Greets, Thomas

Hi Thomas,
This is the game (in expanded pgn form) that I mentioned in another thread. Quark 2.35 (white) loses on time to Crafty 19.15 (black) in a drawn position with Quark ahead in material. Time control was 4min+2sec. The gui was the fritz8 gui and Quark was using wb2uci. It was on an amd 2000+xp with 128M for each engine. Crafty was also using wb2uci. Although the pgn did not record losing on time, I noticed the time expiration from watching the game.
[Event "Q1-3"]
[Site "MyTown"]
[Date "2004.07.23"]
[Round "6.1"]
[White "Quark"]
[Black "Crafty 19.15(wb)"]
[Result "0-1"]
[ECO "E32"]
[PlyCount "254"]
[EventDate "2004.07.23"]
[Source "Pollock"]
[TimeControl "240+2"]
1. d4 {[%emt 0:00:00]} Nf6 {[%emt 0:00:00]} 2. c4 {[%emt 0:00:00]} e6 {
[%emt 0:00:00]} 3. Nc3 {[%emt 0:00:00]} Bb4 {[%emt 0:00:00]} 4. Qc2 {
[%emt 0:00:00]} O-O {[%emt 0:00:00]} 5. a3 {[%emt 0:00:00]} Bxc3+ {
[%emt 0:00:00]} 6. Qxc3 {[%emt 0:00:00]} b6 {[%emt 0:00:00]} 7. Nf3 {
[%emt 0:00:00]} Bb7 {[%emt 0:00:00]} 8. Bg5 {[%emt 0:00:00]} d6 {[%emt 0:00:00]
} 9. Nd2 {[%emt 0:00:00]} Nbd7 {[%emt 0:00:00]} 10. e3 {[%emt 0:00:00]} c5 {
[%emt 0:00:00]} 11. f3 {[%emt 0:00:00]} cxd4 {[%eval -27,10] [%emt 0:00:08]}
12. exd4 {(Qxd4) [%eval -16,11] [%emt 0:00:12]} e5 {
(h6) [%eval -31,10] [%emt 0:00:07]} 13. dxe5 {[%eval -12,10] [%emt 0:00:08]}
dxe5 {[%eval -14,11] [%emt 0:00:07]} 14. Bd3 {[%eval 5,10] [%emt 0:00:10]} Nc5
{(h6) [%eval -7,10] [%emt 0:00:08]} 15. Bf5 {[%eval 3,10] [%emt 0:00:10]} Bc8 {
(h6) [%eval 3,10] [%emt 0:00:07]} 16. Bc2 {(Bxc8) [%eval 19,10] [%emt 0:00:08]}
Ne6 {[%eval -12,10] [%emt 0:00:07]} 17. Bxf6 {(Be3) [%eval 9,10] [%emt 0:00:08]
} Qxf6 {[%eval -25,11] [%emt 0:00:07]} 18. Ne4 {
(Be4) [%eval 3,10] [%emt 0:00:09]} Qh4+ {[%eval -44,11] [%emt 0:00:07]} 19. Ng3
{(g3) [%eval -18,10] [%emt 0:00:11]} Ba6 {(Bb7) [%eval -60,10] [%emt 0:00:06]}
20. Qxe5 {(b3) [%eval -18,9] [%emt 0:00:08]} Bxc4 {[%eval -98,9] [%emt 0:00:06]
} 21. O-O-O {[%eval -25,9] [%emt 0:00:11]} Rfe8 {
(Rfd8) [%eval -22,8] [%emt 0:00:06]} 22. Qe4 {
(Rhe1) [%eval 23,8] [%emt 0:00:07]} Qf4+ {(Qxe4) [%eval 1,10] [%emt 0:00:06]}
23. Kb1 {[%eval 38,10] [%emt 0:00:04]} Qxe4 {[%eval 31,11] [%emt 0:00:06]} 24.
Nxe4 {[%eval 41,10] [%emt 0:00:07]} Be2 {(Rec8) [%eval 33,11] [%emt 0:00:06]}
25. Rd2 {[%eval 47,10] [%emt 0:00:06]} Nf4 {(Ba6) [%eval 51,12] [%emt 0:00:06]}
26. Re1 {(Nc3) [%eval 76,9] [%emt 0:00:04]} Bb5 {[%eval 36,11] [%emt 0:00:06]}
27. g3 {[%eval 55,10] [%emt 0:00:08]} Nh3 {(Ng6) [%eval 46,11] [%emt 0:00:06]}
28. Red1 {[%eval 63,9] [%emt 0:00:07]} Bc6 {[%eval 82,10] [%emt 0:00:06]} 29.
Bb3 {[%eval 80,8] [%emt 0:00:06]} Rf8 {(Re7) [%eval 85,11] [%emt 0:00:06]} 30.
Rc1 {(Bd5) [%eval 106,10] [%emt 0:00:07]} Bxe4+ {[%eval 85,11] [%emt 0:00:05]}
31. fxe4 {[%eval 106,11] [%emt 0:00:06]} Ng5 {[%eval 88,12] [%emt 0:00:05]} 32.
Rd7 {[%eval 86,10] [%emt 0:00:07]} a5 {(Rad8) [%eval 90,12] [%emt 0:00:05]} 33.
Bd5 {(Rcc7) [%eval 99,10] [%emt 0:00:07]} Rad8 {
(Rae8) [%eval 106,12] [%emt 0:00:05]} 34. Ra7 {[%eval 131,10] [%emt 0:00:09]}
Rd6 {(Kh8) [%eval 90,11] [%emt 0:00:08]} 35. Rf1 {
(Rcc7) [%eval 116,10] [%emt 0:00:09]} h6 {[%eval 91,11] [%emt 0:00:05]} 36. b4
{[%eval 133,9] [%emt 0:00:04]} axb4 {[%eval 95,11] [%emt 0:00:05]} 37. axb4 {
[%eval 146,10] [%emt 0:00:05]} Ne6 {(g6) [%eval 105,11] [%emt 0:00:08]} 38. Kb2
{[%eval 138,10] [%emt 0:00:06]} Nd8 {(Ng5) [%eval 111,11] [%emt 0:00:05]} 39.
h4 {(Re7) [%eval 150,9] [%emt 0:00:05]} Kh7 {[%eval 104,10] [%emt 0:00:05]} 40.
b5 {(Rc1) [%eval 150,9] [%emt 0:00:06]} h5 {(Re8) [%eval 108,10] [%emt 0:00:05]
} 41. Kc3 {(Bxf7) [%eval 148,9] [%emt 0:00:08]} Re8 {
(Kh6) [%eval 73,11] [%emt 0:00:05]} 42. Rf2 {(Rd1) [%eval 127,9] [%emt 0:00:05]
} f6 {(Kh6) [%eval 49,11] [%emt 0:00:05]} 43. Ra8 {
(Rf3) [%eval 120,9] [%emt 0:00:05]} Re7 {(Kh6) [%eval 32,11] [%emt 0:00:05]}
44. Rc8 {[%eval 107,8] [%emt 0:00:05]} Nb7 {(Ra7) [%eval 40,10] [%emt 0:00:05]}
45. Rf5 {(Rc6) [%eval 86,9] [%emt 0:00:07]} g6 {
(Nc5) [%eval 44,10] [%emt 0:00:06]} 46. Rf3 {[%eval 80,9] [%emt 0:00:04]} Kg7 {
(Kh6) [%eval 39,10] [%emt 0:00:04]} 47. Rc6 {[%eval 77,10] [%emt 0:00:05]} Red7
{[%eval 43,11] [%emt 0:00:05]} 48. Re3 {(Kb4) [%eval 76,10] [%emt 0:00:09]} f5
{[%eval 30,11] [%emt 0:00:04]} 49. Bc4 {(Kb4) [%eval 41,11] [%emt 0:00:10]}
fxe4 {(Re7) [%eval 19,11] [%emt 0:00:04]} 50. Rxe4 {
[%eval 46,10] [%emt 0:00:04]} Nc5 {[%eval 17,10] [%emt 0:00:04]} 51. Re1 {
[%eval 50,9] [%emt 0:00:06]} Rd2 {[%eval 7,10] [%emt 0:00:04]} 52. Re3 {
[%eval 26,8] [%emt 0:00:04]} Rd1 {[%eval 25,10] [%emt 0:00:04]} 53. Be2 {
(Kb4) [%eval 33,8] [%emt 0:00:02]} R1d4 {(Ra1) [%eval 0,10] [%emt 0:00:04]} 54.
Kc2 {[%eval 36,9] [%emt 0:00:06]} Ra4 {(Rd2+) [%eval 0,11] [%emt 0:00:04]} 55.
Re8 {(Rxb6) [%eval 7,9] [%emt 0:00:08]} Ra2+ {[%eval 0,9] [%emt 0:00:04]} 56.
Kb1 {(Kc3) [%eval 10,10] [%emt 0:00:02]} Rdd2 {[%eval -8,11] [%emt 0:00:04]}
57. Rg8+ {(Rc7+) [%eval 0,10] [%emt 0:00:03]} Kh7 {
[%eval -36,11] [%emt 0:00:04]} 58. Rc7+ {[%eval 0,11] [%emt 0:00:03]} Kxg8 {
[%eval -14,13] [%emt 0:00:04]} 59. Bc4+ {[%eval 0,11] [%emt 0:00:02]} Kf8 {
(Kh8) [%eval -11,13] [%emt 0:00:04]} 60. Bxa2 {[%eval 22,11] [%emt 0:00:07]}
Ne4 {[%eval -27,12] [%emt 0:00:04]} 61. Rf7+ {
(Rc6) [%eval 13,10] [%emt 0:00:03]} Ke8 {[%eval -34,12] [%emt 0:00:04]} 62. Rf3
{[%eval 13,10] [%emt 0:00:02]} Nd6 {(Re2) [%eval -43,11] [%emt 0:00:04]} 63.
Bb3 {(Bg8) [%eval 10,10] [%emt 0:00:02]} Kd7 {
(Nf5) [%eval -20,11] [%emt 0:00:06]} 64. Bc2 {
(Kc1) [%eval 41,10] [%emt 0:00:02]} Rd5 {[%eval -30,11] [%emt 0:00:04]} 65.
Bxg6 {[%eval 37,10] [%emt 0:00:04]} Rxb5+ {[%eval -32,10] [%emt 0:00:03]} 66.
Kc2 {[%eval 36,10] [%emt 0:00:03]} Ke6 {[%eval -17,10] [%emt 0:00:05]} 67. Rf8
{[%eval 38,9] [%emt 0:00:02]} Rc5+ {[%eval -13,10] [%emt 0:00:03]} 68. Kb3 {
(Kd3) [%eval 34,10] [%emt 0:00:04]} Re5 {[%eval -5,10] [%emt 0:00:03]} 69. Rb8
{(Kb4) [%eval 35,10] [%emt 0:00:03]} b5 {(Re3+) [%eval 0,11] [%emt 0:00:03]}
70. Kb4 {(Rh8) [%eval 50,10] [%emt 0:00:03]} Kf6 {
(Re3) [%eval 0,11] [%emt 0:00:03]} 71. Rg8 {(Bd3) [%eval 48,10] [%emt 0:00:03]}
Rd5 {[%eval 0,10] [%emt 0:00:03]} 72. Bh7 {(Bc2) [%eval 87,10] [%emt 0:00:02]}
Nf7 {(Ke7) [%eval 0,11] [%emt 0:00:03]} 73. Rg6+ {[%eval 113,11] [%emt 0:00:02]
} Ke7 {[%eval 0,12] [%emt 0:00:03]} 74. Bg8 {
(Rg7) [%eval 124,12] [%emt 0:00:05]} Rf5 {(Kf8) [%eval 1,11] [%emt 0:00:04]}
75. Ra6 {(Rg7) [%eval 54,12] [%emt 0:00:11]} Nd6 {[%eval 0,11] [%emt 0:00:04]}
76. Ra7+ {[%eval 55,11] [%emt 0:00:04]} Kf6 {[%eval 0,12] [%emt 0:00:03]} 77.
Rh7 {(Rd7) [%eval 55,11] [%emt 0:00:03]} Re5 {(Ne4) [%eval 0,10] [%emt 0:00:03]
} 78. Bb3 {(Rh6+) [%eval 56,10] [%emt 0:00:02]} Kg6 {
[%eval 0,11] [%emt 0:00:06]} 79. Bc2+ {[%eval 66,10] [%emt 0:00:02]} Nf5 {
(Kf6) [%eval 1,11] [%emt 0:00:03]} 80. Rh8 {(Rb7) [%eval 137,10] [%emt 0:00:01]
} Rd5 {[%eval 1,11] [%emt 0:00:03]} 81. Rf8 {
(Rg8+) [%eval 137,9] [%emt 0:00:03]} Re5 {[%eval 1,12] [%emt 0:00:04]} 82. Bd3
{(Bxf5+) [%eval 137,11] [%emt 0:00:02]} Rd5 {[%eval 1,10] [%emt 0:00:01]} 83.
Bb1 {(Bxf5+) [%eval 144,11] [%emt 0:00:03]} Re5 {[%eval 1,12] [%emt 0:00:03]}
84. Bc2 {(Bxf5+) [%eval 137,12] [%emt 0:00:02]} Rd5 {
[%eval 18,12] [%emt 0:00:03]} 85. Rh8 {[%eval 149,11] [%emt 0:00:02]} Kf6 {
(Re5) [%eval 20,11] [%emt 0:00:06]} 86. Rxh5 {[%eval 147,10] [%emt 0:00:01]}
Rd4+ {[%eval 19,10] [%emt 0:00:03]} 87. Kxb5 {[%eval 147,10] [%emt 0:00:01]}
Nxg3 {[%eval 19,11] [%emt 0:00:04]} 88. Rh6+ {[%eval 145,9] [%emt 0:00:03]} Kg7
{(Ke5) [%eval 19,10] [%emt 0:00:03]} 89. Rh7+ {[%eval 144,10] [%emt 0:00:05]}
Kf6 {[%eval 16,10] [%emt 0:00:03]} 90. Kc5 {[%eval 142,10] [%emt 0:00:02]} Rd8
{(Ke5) [%eval 13,9] [%emt 0:00:03]} 91. Kb4 {(Kb6) [%eval 143,8] [%emt 0:00:02]
} Nf5 {(Rd4+) [%eval 1,9] [%emt 0:00:03]} 92. h5 {[%eval 59,10] [%emt 0:00:04]}
Kg5 {[%eval 1,9] [%emt 0:00:03]} 93. Ka5 {(Bxf5) [%eval 33,9] [%emt 0:00:04]}
Ne3 {(Ng3) [%eval 0,10] [%emt 0:00:03]} 94. Be4 {
(Rg7+) [%eval 69,10] [%emt 0:00:02]} Nc4+ {(Nd5) [%eval 0,10] [%emt 0:00:03]}
95. Kb4 {(Kb5) [%eval 85,9] [%emt 0:00:02]} Rd4 {
(Ne5) [%eval 1,9] [%emt 0:00:03]} 96. Bb1 {(Bc6) [%eval 61,8] [%emt 0:00:02]}
Ne5+ {(Nd2+) [%eval 0,9] [%emt 0:00:03]} 97. Kc3 {
(Kc5) [%eval 35,9] [%emt 0:00:03]} Rc4+ {[%eval 0,10] [%emt 0:00:03]} 98. Kb3 {
(Kd2) [%eval 25,9] [%emt 0:00:02]} Rh4 {[%eval 0,9] [%emt 0:00:03]} 99. Rc7 {
(Be4) [%eval 11,9] [%emt 0:00:02]} Rh3+ {(Kf4) [%eval 0,9] [%emt 0:00:03]} 100.
Ka2 {(Kb4) [%eval 9,9] [%emt 0:00:02]} Rh2+ {(Rxh5) [%eval 0,10] [%emt 0:00:03]
} 101. Ka1 {(Kb3) [%eval 1,10] [%emt 0:00:02]} Nf3 {
(Rxh5) [%eval 0,10] [%emt 0:00:03]} 102. Rc5+ {
(h6) [%eval 74,10] [%emt 0:00:03]} Kf4 {(Kh6) [%eval 0,10] [%emt 0:00:03]} 103.
Bg6 {(h6) [%eval 148,7] [%emt 0:00:00]} Kg4 {(Ne5) [%eval 1,9] [%emt 0:00:03]}
104. Ra5 {(Rc2) [%eval 152,9] [%emt 0:00:02]} Ng1 {
(Kf4) [%eval 0,9] [%emt 0:00:03]} 105. Be8 {(Rd5) [%eval 62,9] [%emt 0:00:03]}
Nf3 {(Ne2) [%eval 0,10] [%emt 0:00:03]} 106. Bf7 {
(h6) [%eval 146,9] [%emt 0:00:02]} Kf4 {(Ng5) [%eval 1,9] [%emt 0:00:03]} 107.
Rb5 {(Kb1) [%eval 142,9] [%emt 0:00:01]} Ne5 {[%eval 1,9] [%emt 0:00:03]} 108.
Be8 {(Rb2) [%eval 139,9] [%emt 0:00:02]} Ng4 {(Kf5) [%eval 1,9] [%emt 0:00:03]}
109. Rb4+ {(Rb2) [%eval 35,9] [%emt 0:00:03]} Kg5 {[%eval 0,11] [%emt 0:00:02]}
110. Rb8 {(h6) [%eval 33,9] [%emt 0:00:02]} Ne3 {
(Nf6) [%eval 0,10] [%emt 0:00:02]} 111. Rb5+ {(h6) [%eval 119,8] [%emt 0:00:01]
} Kh6 {(Kf4) [%eval 0,9] [%emt 0:00:02]} 112. Rb6+ {
[%eval 140,9] [%emt 0:00:02]} Kg5 {[%eval 0,10] [%emt 0:00:02]} 113. Rg6+ {
(Rb5+) [%eval 120,9] [%emt 0:00:02]} Kf4 {[%eval 0,10] [%emt 0:00:02]} 114. Ra6
{(h6) [%eval 59,10] [%emt 0:00:04]} Nc2+ {(Nf5) [%eval 1,9] [%emt 0:00:02]}
115. Kb1 {[%eval 145,10] [%emt 0:00:03]} Nd4 {(Ne3) [%eval 1,9] [%emt 0:00:02]}
116. Ra4 {[%eval 146,10] [%emt 0:00:02]} Ke5 {[%eval 0,9] [%emt 0:00:02]} 117.
Ra5+ {[%eval 123,9] [%emt 0:00:02]} Kf6 {(Kf4) [%eval 0,9] [%emt 0:00:02]} 118.
Bg6 {(Bd7) [%eval 135,9] [%emt 0:00:02]} Kg7 {(Rh1+) [%eval 0,9] [%emt 0:00:02]
} 119. Ra7+ {(Ra6) [%eval 90,8] [%emt 0:00:02]} Kh6 {[%eval 1,9] [%emt 0:00:02]
} 120. Ra6 {[%eval 63,9] [%emt 0:00:03]} Kg5 {(Ne2) [%eval 0,10] [%emt 0:00:02]
} 121. Ra5+ {(Bd3) [%eval 63,9] [%emt 0:00:01]} Kh6 {
[%eval 0,10] [%emt 0:00:02]} 122. Be8 {(Rd5) [%eval 46,10] [%emt 0:00:02]} Re2
{(Ne2) [%eval 0,10] [%emt 0:00:02]} 123. Bd7 {(Rd5) [%eval 57,9] [%emt 0:00:02]
} Re1+ {(Ne6) [%eval 0,10] [%emt 0:00:02]} 124. Kb2 {
[%eval 39,10] [%emt 0:00:02]} Nf3 {[%eval 0,9] [%emt 0:00:02]} 125. Bc6 {
(Bg4) [%eval 29,9] [%emt 0:00:02]} Nd4 {(Re5) [%eval 0,10] [%emt 0:00:02]} 126.
Ba8 {(Ra6) [%eval 53,9] [%emt 0:00:01]} Re3 {[%eval 0,10] [%emt 0:00:02]} 127.
Bh1 {(Bd5) [%eval 31,9] [%emt 0:00:02]} Rb3+ {Time [%eval 0,10] [%emt 0:00:02]}
0-1

[Olivier Deville]

Geschrieben von:/Posted by: Olivier Deville at 25 July 2004 17:30:42:
Als Antwort auf:/In reply to: Quark test and a gentle hint.... geschrieben von:/posted by: Roger Brown at 25 July 2004 12:59:49:

Hello Roger, Thomas, Norm and any people who care to read this...
After reading Roger's post, I ran a match between Quark 2.35 and Zarkov 4.67 on my PIV 3Ghz : 50 games, 1mn +1s, all 5 men... in the Fritz GUI.
Result : Zarkov wins 33.5 to 16.5
No loss on time.
Roger, your experience proves nothing, and you can say the same about mine
Olivier


ChessWar

[Thomas Mayer]

Geschrieben von:/Posted by: Thomas Mayer at 25 July 2004 18:35:27:
Als Antwort auf:/In reply to: Re: Quark test and a gentle hint.... geschrieben von:/posted by: Norm Pollock at 25 July 2004 16:15:58:

Hi Norm,

This is the game (in expanded pgn form) that I mentioned in another thread.
Quark 2.35 (white) loses on time to Crafty 19.15 (black) in a drawn position
with Quark ahead in material. Time control was 4min+2sec. The gui was the
fritz8 gui and Quark was using wb2uci. It was on an amd 2000+xp with 128M
for each engine. Crafty was also using wb2uci. Although the pgn did not
record losing on time, I noticed the time expiration from watching the game.

well, at the end of the pgn it is pointed out that Quark did lose on time, but at the moment I have no idea why it did use it's full 8 seconds there... at least this is not reproduceable... Did Quark lose more then one game on your PC on time ? If yes, can you turn "log on" or better "log app" in your Quark.qbt so that I have some hints what could cause this out of the log file ?! Would be quite interesting... When I test most with WinBoard I anyway want Quark to be able to play without problems in as many other GUIs as possible...
Greets, Thomas

[Jose Carlos]

Geschrieben von:/Posted by: Jose Carlos at 26 July 2004 02:15:21:
Als Antwort auf:/In reply to: Re: Quark test and a gentle hint.... geschrieben von:/posted by: Thomas Mayer at 25 July 2004 18:35:27:

Hi Norm,

This is the game (in expanded pgn form) that I mentioned in another thread.
Quark 2.35 (white) loses on time to Crafty 19.15 (black) in a drawn position
with Quark ahead in material. Time control was 4min+2sec. The gui was the
fritz8 gui and Quark was using wb2uci. It was on an amd 2000+xp with 128M
for each engine. Crafty was also using wb2uci. Although the pgn did not
record losing on time, I noticed the time expiration from watching the game.

well, at the end of the pgn it is pointed out that Quark did lose on time, but at the moment I have no idea why it did use it's full 8 seconds there... at least this is not reproduceable... Did Quark lose more then one game on your PC on time ? If yes, can you turn "log on" or better "log app" in your Quark.qbt so that I have some hints what could cause this out of the log file ?! Would be quite interesting... When I test most with WinBoard I anyway want Quark to be able to play without problems in as many other GUIs as possible...
Greets, Thomas

Hi Thomas. I don't know if this is the case, but I remember some time ago I tested Averno and Anubis under Fritz with wb2uci. Averno seemed to work fine but Anubis didn't. What I observed was that Anubis' thinking wasn't printed until several seconds after its clock started to run. Then, if it used for example 1 second, only 1 second was added to its clock, but if it lost on time during that delay, the game was considered lost.
Both programs did setbuf(stdout,NULL) (or whatever the sysntax, I don't remember now) at boot time, both used printf to write to console and both did fflush() after the printf. The only difference is that Anubis uses two threads (one to listen to input and one to think) and Averno only one. In Anubis, both threads can print, but they can't do it at the same time.
I never discovered if multithreading had something to do with the delay, but I also didn't investigate much because I don't like Fritz at all.
José C.

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 26 July 2004 04:56:32:
Als Antwort auf:/In reply to: Re: Quark test and a gentle hint.... geschrieben von:/posted by: Thomas Mayer at 25 July 2004 18:35:27:

Hi Norm,

This is the game (in expanded pgn form) that I mentioned in another thread.
Quark 2.35 (white) loses on time to Crafty 19.15 (black) in a drawn position
with Quark ahead in material. Time control was 4min+2sec. The gui was the
fritz8 gui and Quark was using wb2uci. It was on an amd 2000+xp with 128M
for each engine. Crafty was also using wb2uci. Although the pgn did not
record losing on time, I noticed the time expiration from watching the game.

well, at the end of the pgn it is pointed out that Quark did lose on time, but at the moment I have no idea why it did use it's full 8 seconds there... at least this is not reproduceable... Did Quark lose more then one game on your PC on time ? If yes, can you turn "log on" or better "log app" in your Quark.qbt so that I have some hints what could cause this out of the log file ?! Would be quite interesting... When I test most with WinBoard I anyway want Quark to be able to play without problems in as many other GUIs as possible...
Greets, Thomas

Hi Thomas,
I am not able to reproduce it. It might turn up again in which case I will contact you. In the meantime, I took 2 precautions to decrease the likelihood that it re-occurs. First thing, I defragmented the drives where all my chess programs are located. This was long overdue and may have slowed down tablebase access. Secondly, I increased the tablebase cache in the fritz8 gui from 8M to 16M. That should also improve tablebase access speed.
-Norm

[Thomas Mayer]

Geschrieben von:/Posted by: Thomas Mayer at 26 July 2004 19:33:39:
Als Antwort auf:/In reply to: Re: Quark test and a gentle hint.... geschrieben von:/posted by: Jose Carlos at 26 July 2004 02:15:21:

Hi José,
thanks for the hint about the differences in single threaded and multi-threaded software, I will keep that in mind. At the moment Quark is only single threaded, but this might change in the feature, so it is definitely a good idea to have some problems in mind which might occur...
Greets, Thomas