crafty 12.6 losing on time
j...@yktvmv.watson.ibm.com
Crafty 12.6 has lost on time a couple of times while playing
blitz (40 moves in 5 minutes) games against me. Isn't this never
supposed to happen?
I enclose portions of the log for one case. Crafty has 6.88
seconds to make its 40th move but takes 7.07 seconds causing xboard
to flag it.
......................................................................
White(1): beep
White(1): random
White(1): hard
White(1): level 40 5 0
40 moves/5 minutes primary time control
40 moves/5 minutes secondary time control
White(1): time 30000
time remaining: 5:00 (crafty).
White(1): otim 30000
time remaining: 5:00 (opponent).
White(1): bogus
White(1): time 30000
time remaining: 5:00 (crafty).
White(1): otim 30000
time remaining: 5:00 (opponent).
.......................................................................
White(40): Qg7+ Ýpondering¨
time surplus 0.46 time limit 7.96
depth time score variation (6)
White(40): time 688
time remaining: 6.88 (crafty).
White(40): otim 799
time remaining: 7.99 (opponent).
White(40): f8f7
time used: 4.69
time surplus 715827:52 time limit 6.88
depth time score variation (1)
6 0.74 -3.903 Kd6 Qf6+ Kc7 Qg7+ Kd6 h4 a5 h5 a4
g4 <HT>
6-> 0.83 -3.903 Kd6 Qf6+ Kc7 Qg7+ Kd6 h4 a5 h5 a4
g4 <HT>
7 3.75 -3.903 Kd6 Qf6+ Kc7 Qg7+ Kd6 h4 a5 <HT>
7-> 4.11 -3.903 Kd6 Qf6+ Kc7 Qg7+ Kd6 h4 a5 <HT>
time: 7.05 cpu:47% mat:-2 n:122390 nps:36210
ext-> checks:12549 recaps:224 pawns:142 1rep:1611
predicted:22 nodes:122390 evals:33567
endgame tablebase-> probes done: 0 successful: 0
hashing-> trans/ref:45% pawn:99% used:w49% b83%
Black(40): Kd6
time used: 7.07
time control reached
.........................................................................
James B. Shearer
Robert Hyatt
j...@yktvmv.watson.ibm.com wrote:
: Crafty 12.6 has lost on time a couple of times while playing
: blitz (40 moves in 5 minutes) games against me. Isn't this never
: supposed to happen?
: I enclose portions of the log for one case. Crafty has 6.88
: seconds to make its 40th move but takes 7.07 seconds causing xboard
: to flag it.
Not much you can do about that, other than to use one of the options
to set an operator time of one minute, which will reserve one minute
that Crafty will therefore not use. Problem here is that there *is*
a delay in sending a move from Crafty to xboard/winboard. This delay
is the system's context switching time to execute crafty, which does
the write to xboard, then as soon as the scheduler chooses to, it will
schedule xboard for execution. And taking .2-.3 seconds is not
unexpected... Also, it is possible for crafty to misjudge the time
control by a fraction of a second itself... I'll see if I can't put in a
fudge factor for zero-increment games to reserve a couple of seconds, at
least..