Crafty 22.0 Windows x64 available

[peter]

Do you have "mt=2" in your crafty.rc file? without that, crafty will only use one CPU...[/quote]

So I have, with this in the Wb2Uci too and smpnice=0 in .rc task manager shows 100%, but I crash. Maybe I'll wait for another Windows executable but thanks again
regards
Peter.

[peter]

Do you have "mt=2" in your crafty.rc file? without that, crafty will only use one CPU...

So I have, with this in the Wb2Uci too and smpnice=0 in .rc task manager shows 100%, but I crash. Maybe I'll wait for another Windows executable but thanks again
regards
Peter.[/quote]

At the moment it seems I've mananged it somehow. There was just too much hash maybe. To use both cores I have to have mt=2 not only in crafty.rc but also in Wb2UCI at command line of engine.
Great performance now
regards
Peter.

[Werner]

Sounds strange, I should be able to stop one process when I want without the GUI thinking the whole thing stops. It works fine under xboard/winboard however...

Hi Bob,
I hope you find a solution - I had crashes under Winboard too here (not to forget: I am no specialist for Winboard).
After the fist move here Crafty SMP crashes too. It only worked fine on the Quad where it could ponder.

[AdminX]

Yeah, I have to join the list also. I tried every thing I could think of. Thanks again for all of your hard work.

[peter]

Sounds strange, I should be able to stop one process when I want without the GUI thinking the whole thing stops. It works fine under xboard/winboard however...

Hi Bob,
I hope you find a solution - I had crashes under Winboard too here (not to forget: I am no specialist for Winboard).
After the fist move here Crafty SMP crashes too. It only worked fine on the Quad where it could ponder.

My first successes still sicken too. Not until looking in task manager at different processes I saw Crafty always using one core of dual at thinking time of opponent
regards
Peter.

[bob]

Sounds strange, I should be able to stop one process when I want without the GUI thinking the whole thing stops. It works fine under xboard/winboard however...

Hi Bob,
I hope you find a solution - I had crashes under Winboard too here (not to forget: I am no specialist for Winboard).
After the fist move here Crafty SMP crashes too. It only worked fine on the Quad where it could ponder.

Try this first. Open a command window, and run crafty in text mode. Do the commands you want (mt=2, ponder=off, etc) and play like that. If it doesn't crash, and I doubt it will, then there is something odd going on with the GUI.

Crafty plays so many games, there just are not any SMP bugs in the program that will make it crash. This identical version has been playing on ICC continuously for weeks and simply does not have a problem using xboard. A log.nnn file would help to see what is happening, particularly if the GUI is terminating crafty...

[bob]

Sounds strange, I should be able to stop one process when I want without the GUI thinking the whole thing stops. It works fine under xboard/winboard however...

Hi Bob,
I hope you find a solution - I had crashes under Winboard too here (not to forget: I am no specialist for Winboard).
After the fist move here Crafty SMP crashes too. It only worked fine on the Quad where it could ponder.

My first successes still sicken too. Not until looking in task manager at different processes I saw Crafty always using one core of dual at thinking time of opponent
regards
Peter.

Are you running the fixed version from yesterday? smpnice=1 is now the default, and if crafty is not pondering, it does not use any processor time at all when waiting on the opponent's move... The earlier version had a problem to be sure, as I never test with smpnice=1 and ponder=off and it was definitely broken during the CCT10 preparation. But I tested it carefully yesterday before releasing a new version and it worked just as advertised for me. Peter Skinner runs it under windows and has not seen any crashes during our testing, and obviously under Linux all is well...

I have seen some very strange stuff in various crafty.rc/.craftyrc files, so those could be problematic.

[Werner]

Hi bob,
here are my files
.rc
egtb
tbpath=J:\tbs
# adaptive 750K 24M 192M 24M 48M
egtb
tbpath=J:\tbs
cache=32M
ponder off
mode=tournament
swindle on # (Allows Crafty to try to win drawn games (according to Endgame Tables))
learn 7
book random 1
bookw freq 0.7
bookw ratio 0.8
bookw eval 0.6
bookw learn 1
bookw cap 0.5
book width 4
book on
show book
computer
log=on
mt=2 #(Increases Crafty's MaxThreads to 2 for a dual CPU computer - disables smp)
# smpmt=2 #(Increases Crafty's MaxThreads to 2 for a dual CPU computer - using smp dual-core/2+ cpus)
timebook 80 8
#adaptive NPS a
hash=512M
hashp=16M

log.001
EGTB access enabled
using tbpath=J:\tbs
5 piece tablebase files found
11059kb of RAM used for TB indices and decompression tables
EGTB cache memory = 32M bytes.
pondering disabled.
book learning enabled
result learning enabled
choose from book moves randomly (using weights.)
choose from 4 best moves.
book file enabled.
show book statistics
playing a computer!

log.002
max threads set to 2
hash table memory = 384M bytes.
pawn hash table memory = 10M bytes.

Crafty v22.0 JA (2 cpus)

White(1): e2e4
time used: 17.46
time surplus 0.00 time limit 42.75 (+20.25) (3:30)
depth time score variation (1)
starting thread 1
9-> 0.14 0.15 1. ... Nc6 2. Nc3 Nf6 3. Nf3 d5 4.
e5 Ne4 5. Nxe4 dxe4
10 0.17 0.22 1. ... Nc6 2. Nc3 Nf6 3. Nf3 e6 4.
d4 d5 5. Qd3 Bb4 6. e5
10-> 0.24 0.22 1. ... Nc6 2. Nc3 Nf6 3. Nf3 e6 4.
d4 d5 5. Qd3 Bb4 6. e5 (s=2)
11 0.57 0.32 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. exd5
exd5 5. d4 Nf6 6. Qe2+ Be7 7. Bf4
11-> 0.75 0.32 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. exd5
exd5 5. d4 Nf6 6. Qe2+ Be7 7. Bf4
12 1.07 0.32 1. ... Nc6 2. Nf3 e6 3. d4 Nf6 4. e5
Nd5 5. Bd3 Ncb4 6. O-O Nxd3 7. Qxd3
<HT>
12-> 1.55 0.32 1. ... Nc6 2. Nf3 e6 3. d4 Nf6 4. e5
Nd5 5. Bd3 Ncb4 6. O-O Nxd3 7. Qxd3
<HT> (s=2)
13 2.02 0.32 1. ... Nc6 2. Nf3 e6 3. d4 Nf6 4. Bd3
Bb4+ 5. c3 Be7 6. O-O O-O 7. e5 Nd5
8. Qd2
13-> 2.49 0.32 1. ... Nc6 2. Nf3 e6 3. d4 Nf6 4. Bd3
Bb4+ 5. c3 Be7 6. O-O O-O 7. e5 Nd5
8. Qd2 (s=2)
14 4.50 0.28 1. ... Nc6 2. Nf3 e5 3. d4 exd4 4.
Nxd4 Qh4 5. Nb5 Qxe4+ 6. Qe2 Qxe2+
7. Bxe2 Kd8 8. N1c3 Nb4 9. Nd4 <HT>
14-> 7.75 0.28 1. ... Nc6 2. Nf3 e5 3. d4 exd4 4.
Nxd4 Qh4 5. Nb5 Qxe4+ 6. Qe2 Qxe2+
7. Bxe2 Kd8 8. N1c3 Nb4 9. Nd4 <HT>
(s=2)
15 9.91 0.44 1. ... Nc6 2. Nf3 e5 3. d4 exd4 4.
Nxd4 Nf6 5. Nxc6 bxc6 6. Nc3 Bb4 7.
Qd4 Bxc3+ 8. Qxc3 O-O 9. Qe5 <HT>
15 12.22 0.29 1. ... e5 2. Nf3 Nf6 3. Nxe5 d6 4.
Nf3 Nxe4 5. Nc3 Nxc3 6. dxc3 Nc6 7.
Bc4 Qe7+ 8. Be3 Be6
15-> 14.28 0.29 1. ... e5 2. Nf3 Nf6 3. Nxe5 d6 4.
Nf3 Nxe4 5. Nc3 Nxc3 6. dxc3 Nc6 7.
Bc4 Qe7+ 8. Be3 Be6 (s=3)
16 25.64 0.34 1. ... e5 2. Nf3 Nf6 3. Nc3 Nc6 4.
d4 exd4 5. Nxd4 Bb4 6. Nxc6 bxc6 7.
Qd4 Bxc3+ 8. Qxc3 O-O 9. Qc4 Bb7 (s=2)
16-> 33.16 0.34 1. ... e5 2. Nf3 Nf6 3. Nc3 Nc6 4.
d4 exd4 5. Nxd4 Bb4 6. Nxc6 bxc6 7.
Qd4 Bxc3+ 8. Qxc3 O-O 9. Qc4 Bb7
time=44.14 mat=0 n=118859514 fh=91% nps=2.7M
ext-> check=4.9M 1rep=279K mate=3K pp=0 reduce=70.7M/8.4M
predicted=0 evals=96.0M 50move=0 EGTBprobes=0 hits=0
SMP-> splits=513 aborts=46 data=5/512 elap=44.14
terminating SMP processes.

I only started in console mode and started with e2e4

here is console output from start:

Initializing multiple threads.
System is SMP, not NUMA.
unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].
EGTB access enabled
using tbpath=J:\tbs
5 piece tablebase files found
11059kb of RAM used for TB indices and decompression tables
EGTB cache memory = 32M bytes.
pondering disabled.
use 'settc' command if a game is restarted after crafty
has been terminated for any reason.
tournament mode.
book learning enabled
result learning enabled
choose from book moves randomly (using weights.)
choose from 4 best moves.
book file enabled.
show book statistics
playing a computer!
max threads set to 2
hash table memory = 384M bytes.
pawn hash table memory = 10M bytes.

Crafty v22.0 JA (2 cpus)

White(1):


PS: it does not crash when ponder is on!!
see console output:


Initializing multiple threads.
System is SMP, not NUMA.
unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].
EGTB access enabled
using tbpath=J:\tbs
5 piece tablebase files found
11059kb of RAM used for TB indices and decompression tables
EGTB cache memory = 32M bytes.
pondering enabled.
use 'settc' command if a game is restarted after crafty
has been terminated for any reason.
tournament mode.
book learning enabled
result learning enabled
choose from book moves randomly (using weights.)
choose from 4 best moves.
book file enabled.
show book statistics
playing a computer!
max threads set to 2
hash table memory = 384M bytes.
pawn hash table memory = 10M bytes.

Crafty v22.0 JA (2 cpus)

White(1): e2e4
time used: 2.28
time surplus 0.00 time limit 57.00 (+27.00) (3:30)
depth time score variation (1)
starting thread 1
9-> 0.14 0.15 1. ... Nc6 2. Nf3 Nf6 3. Nc3 d5 4.
e5 Ne4 5. Nxe4 dxe4 <HT>
10 0.19 0.22 1. ... Nc6 2. Nf3 Nf6 3. Nc3 e6 4.
d4 d5 5. Qd3 Bb4 6. e5
10-> 0.27 0.22 1. ... Nc6 2. Nf3 Nf6 3. Nc3 e6 4.
d4 d5 5. Qd3 Bb4 6. e5 (s=2)
11 0.39 0.34 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. d4
Nf6 5. e5 Ne4 6. Nxe4 dxe4 <HT>
11-> 0.78 0.34 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. d4
Nf6 5. e5 Ne4 6. Nxe4 dxe4 <HT> (s=3)
12 1.03 0.33 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. Bb5
dxe4 5. Nxe4 f5 6. Nc3 Nf6 7. O-O (s=2)
12-> 1.78 0.33 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. Bb5
dxe4 5. Nxe4 f5 6. Nc3 Nf6 7. O-O (s=3)
13 2.17 0.34 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. Bb5
dxe4 5. Nxe4 f5 6. Nc3 Nf6 7. Bxc6+
bxc6 8. O-O (s=2)
13-> 3.95 0.34 1. ... Nc6 2. Nf3 e6 3. Nc3 d5 4. Bb5
dxe4 5. Nxe4 f5 6. Nc3 Nf6 7. Bxc6+
bxc6 8. O-O (s=4)
14 5.95 0.26 1. ... Nc6 2. Nf3 e5 3. Nc3 Nf6 4.
d4 Bb4 5. Nxe5 Bxc3+ 6. bxc3 Qe7 7.
Bc4 Nxe5 8. dxe5 Qxe5 (s=3)
14-> 7.00 0.26 1. ... Nc6 2. Nf3 e5 3. Nc3 Nf6 4.
d4 Bb4 5. Nxe5 Bxc3+ 6. bxc3 Qe7 7.
Bc4 Nxe5 8. dxe5 Qxe5
15 10.00 0.19 1. ... Nc6 2. Nf3 e5 3. Nc3 Nf6 4.
d4 Bb4 5. Nxe5 Bxc3+ 6. bxc3 Qe7 7.
Bf4 d6 8. Nxc6 Qxe4+ 9. Qe2 bxc6
15-> 11.66 0.19 1. ... Nc6 2. Nf3 e5 3. Nc3 Nf6 4.
d4 Bb4 5. Nxe5 Bxc3+ 6. bxc3 Qe7 7.
Bf4 d6 8. Nxc6 Qxe4+ 9. Qe2 bxc6
16 19.25 0.32 1. ... Nc6 2. Nf3 e5 3. Nc3 Nf6 4.
Bc4 Nxe4 5. O-O Nd6 6. Bd5 Nf5 7. Bxc6
dxc6 8. Nxe5 Bc5 9. Re1
16-> 26.75 0.32 1. ... Nc6 2. Nf3 e5 3. Nc3 Nf6 4.
Bc4 Nxe4 5. O-O Nd6 6. Bd5 Nf5 7. Bxc6
dxc6 8. Nxe5 Bc5 9. Re1
17 36.56 0.38 1. ... Nc6 2. Nf3 e5 3. Nc3 Nf6 4.
d4 exd4 5. Nxd4 Bb4 6. Nxc6 Bxc3+ 7.
bxc3 bxc6 8. e5 Nd5 9. c4 Nc3 <HT>
17 1:41 0.29 1. ... e5 2. Nf3 Nf6 3. Nxe5 d6 4.
Nf3 Nxe4 5. Nc3 Nxc3 6. dxc3 Be7 7.
Bc4 O-O 8. O-O Nc6 9. Re1 Bf5
time=1:41 mat=0 n=284005364 fh=91% nps=2.8M
ext-> check=13.0M 1rep=769K mate=11K pp=0 reduce=171.8M/19.8M
predicted=0 evals=228.6M 50move=0 EGTBprobes=0 hits=0
SMP-> splits=483 aborts=63 data=6/512 elap=1:41

Black(1): e5
time used: 1:41
time remaining 0:28 (Crafty) 0:29 (opponent).
59 moves to next time control (Crafty)
if clocks are wrong, use 'clock' command to adjust them
White(2): Nf3 [pondering]
time surplus 0.00 time limit 51.80 (+23.02) (3:21)
depth time score variation (16)
16 7.81 0.33 2. ... Nf6 3. Nxe5 d6 4. Nf3 Nxe4 5.
Nc3 Nxc3 6. dxc3 Be7 7. Bc4 Be6 8.
Bxe6 fxe6 9. O-O Nc6 10. Bf4
16 12.84 12/29? 2. ... Bb4 (2.0Mnps)

[Peter Skinner]

Peter Skinner runs it under windows and has not seen any crashes during our testing, and obviously under Linux all is well...

I have seen some very strange stuff in various crafty.rc/.craftyrc files, so those could be problematic.

I have zero problems running under Windows with ponder on or off.

This is all my rc file has:

mt=4
egtb
hash=384M
hashp=128M
cache=32M
exit

And this is what I see when I start it in console mode:

Initializing multiple threads.
System is SMP, not NUMA.
found computer opening book file [./bookc.bin].
max threads set to 4
EGTB access enabled
using tbpath=./TB
5 piece tablebase files found
14005kb of RAM used for TB indices and decompression tables
hash table memory = 384M bytes.
pawn hash table memory = 80M bytes.
EGTB cache memory = 32M bytes.


Crafty v22.0 PS (4 cpus)

White(1):

No problems here at all, and I run _many_ games a day in testing. I use 15+0 for games and use 1 cpu, 2 cpu, 4 cpu. All tests have zero crashes.

Peter

[Werner]

Yes Peter,
thanks a lot
I downloaded your compile, I used your .rc file with ponder off - and it works under Arena!
Wunderful!
So I can start next week some matches here for CEGT 40/20 with new Crafty version.
I am very happy now!

Thanks
Werner