Again, during a game against Crafty-23.0 under xboard. The crash seems "identical" to the last one, at least according to the backtrace in the crash report:
zullil wrote:Again, during a game against Crafty-23.0 under xboard. The crash seems "identical" to the last one, at least according to the backtrace in the crash report:
max threads set to 8.
EGTB access enabled
using tbpath=/Users/louis/Documents/Chess/Crafty/TB
5 piece tablebase files found
14043kb of RAM used for TB indices and decompression tables
EGTB cache memory = 256M bytes.
hash table memory = 1024M bytes.
pawn hash table memory = 512M bytes.
choose from book moves randomly (using weights.)
choose from 5 best moves.
pondering disabled.
Audio output disabled
game/10 minutes primary time control
Crafty v23.1 (8 cpus)
White(1): xboard
White(1): protover 2
feature ping=1 setboard=1 san=1 time=1 draw=1
feature sigint=0 sigterm=0 reuse=1 analyze=1
feature myname="Crafty-23.1" name=1
feature playother=1 colors=0
feature variants="normal,nocastle"
feature done=1
White(1): accepted ping
White(1): accepted setboard
White(1): accepted san
White(1): accepted time
White(1): accepted draw
White(1): accepted sigint
White(1): accepted sigterm
White(1): accepted reuse
White(1): accepted analyze
White(1): accepted myname
White(1): accepted name
White(1): accepted playother
White(1): accepted colors
White(1): accepted variants
White(1): accepted done
White(1): new
parallel threads terminated.
White(1): smpmt=8
max threads set to 8.
White(1): learn=0
book learning disabled
White(1): egtb
White(1): cache=256M
EGTB cache memory = 256M bytes.
White(1): hash=1024M
hash table memory = 1024M bytes.
White(1): hashp=512M
pawn hash table memory = 512M bytes.
White(1): book random 1
choose from book moves randomly (using weights.)
White(1): book width 5
choose from 5 best moves.
White(1): ponder=off
pondering disabled.
White(1): xboard
White(1): level 0 3 0
game/180 seconds primary time control
White(1): hard
pondering enabled.
White(1): easy
pondering disabled.
White(1): ping 1
pong 1
White(1): force
White(1): computer
playing a computer!
White(1): name Crafty-23.0
Crafty 23.1 vs Crafty-23.0
White(1): time 18000
time remaining: 3:00 (Crafty).
White(1): otim 18000
time remaining: 3:00 (opponent).
White(1): go
book moves {e4, f4, d4}
book 0.0s 38% d4! Nf6
White(1): d4
time used: 0.00
Black(1): time 17999
time remaining: 2:59 (Crafty).
Black(1): otim 17836
time remaining: 2:58 (opponent).
Black(1): Nf6
time used: 1.63
book moves {c4, Nf3}
book 0.0s 79% c4 e6
Black(41): a5
time used: 4.12
time limit 2.06 (+0.00) (12.36)
depth time score variation (1)
starting thread 1
starting thread 2
starting thread 3
starting thread 4
starting thread 5
starting thread 6
starting thread 7
15 0.44 0.12 42. Qe1 Qb7+ 43. f3 Qa6 44. bxa5 Qxa5
45. Qxa5 bxa5 46. c5 Ke7 47. Kf2 Kd7
48. Ke3 Kc6 49. Kd4 e5+ 50. Kc4
15-> 0.83 0.12 42. Qe1 Qb7+ 43. f3 Qa6 44. bxa5 Qxa5
45. Qxa5 bxa5 46. c5 Ke7 47. Kf2 Kd7
48. Ke3 Kc6 49. Kd4 e5+ 50. Kc4
16 2.14 0.09 42. Qe1 Qb7+ 43. Kg1 a4 44. Qe3 Qc6
45. Qd3 e5 46. f3 Qe6 47. Kg2 f5 48.
h4 h6 49. Kf2 Kf6 50. Qd8+ Kf7
16-> 3.76 0.09 42. Qe1 Qb7+ 43. Kg1 a4 44. Qe3 Qc6
45. Qd3 e5 46. f3 Qe6 47. Kg2 f5 48.
h4 h6 49. Kf2 Kf6 50. Qd8+ Kf7 (s=2)
time=3.76 mat=0 n=66676617 fh=91% nps=17.7M
extensions=4.2M qchecks=4.6M reduced=6.1M pruned=9.7M
predicted=23 evals=36.5M 50move=0 EGTBprobes=0 hits=0
SMP-> splits=14408 aborts=1657 data=35/512 elap=3.76
terminating SMP processes.
White(42): Qe1
time used: 3.76
Black(42): time 5622
time remaining: 56.22 (Crafty).
Black(42): otim 11102
time remaining: 1:51 (opponent).
Black(42): Qc7
time used: 2.51
time limit 1.93 (+0.00) (11.58)
depth time score variation (1)
starting thread 1
starting thread 2
starting thread 3
starting thread 4
starting thread 5
starting thread 6
starting thread 7
I'll see if I can fire up some 8-core vs 8-core tests to see if I can reproduce this at all. Just running one instance of Crafty is working perfectly. I play with 8 cores on ICC all the time and have not seen any crashes whatsoever there this year. Even some of my cluster testing has used 8 cores although I don't play Crafty vs Crafty nor do I use opening books on the cluster. I have a couple of million games using 8 cores and again, not a single crash. Not sure what to make of this. Could be a hardware issue (I have seen this on overclocked hardware) or a compiler issue, or some program bug that only shows up rarely, although it seems pretty frequent from your posts...
bob wrote:
I'll see if I can fire up some 8-core vs 8-core tests to see if I can reproduce this at all. Just running one instance of Crafty is working perfectly. I play with 8 cores on ICC all the time and have not seen any crashes whatsoever there this year. Even some of my cluster testing has used 8 cores although I don't play Crafty vs Crafty nor do I use opening books on the cluster. I have a couple of million games using 8 cores and again, not a single crash. Not sure what to make of this. Could be a hardware issue (I have seen this on overclocked hardware) or a compiler issue, or some program bug that only shows up rarely, although it seems pretty frequent from your posts...
My hardware is all "standard." No over-clocking or other modifications. I compiled using icc. Here's the relevant portion of the makefile I used:
bob wrote:
I'll see if I can fire up some 8-core vs 8-core tests to see if I can reproduce this at all. Just running one instance of Crafty is working perfectly. I play with 8 cores on ICC all the time and have not seen any crashes whatsoever there this year. Even some of my cluster testing has used 8 cores although I don't play Crafty vs Crafty nor do I use opening books on the cluster. I have a couple of million games using 8 cores and again, not a single crash. Not sure what to make of this. Could be a hardware issue (I have seen this on overclocked hardware) or a compiler issue, or some program bug that only shows up rarely, although it seems pretty frequent from your posts...
My hardware is all "standard." No over-clocking or other modifications. I compiled using icc. Here's the relevant portion of the makefile I used:
As you can see, I used pgo. Maybe the crash is compiler-related. Thanks for investigating this as much as you have already.
I'm using ICC also and always use PGO with no problems. This still needs investigation. I have had quirks show up in older pthread libraries, whether this is related to that or not is unknown at present.