Crafty 25.0 Release

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

User avatar
Sylwy
Posts: 4812
Joined: Fri Apr 21, 2006 4:19 pm
Location: IAȘI - the historical capital of MOLDOVA
Full name: Silvian Rucsandescu

Re: Updated

Post by Sylwy »

cdani wrote:
Dann Corbit wrote:http://www.andscacs.com/crafty/crafty25.zip
does not work for me.
What happens when you click it?
Because I was updating with the latest sources :-)

Done!

http://www.andscacs.com/crafty/crafty25.zip
Works !

THANK YOU, Daniel !

What can I do without your binaries ????

La patrie reconnaissante !
Top
Charly
Posts: 1091
Joined: Wed Jul 23, 2014 4:30 pm
Location: Bretagne

Re: Crafty 25.0 Release

Post by Charly »

Many thanks to Dr. Robert Hyatt for the long term work since years on Crafty which is always a must have ! :D

Many Thanks to Daniel José for the compils also ! 8-)
Top
User avatar
Werner
Posts: 2967
Joined: Wed Mar 08, 2006 10:09 pm
Location: Germany
Full name: Werner Schüle

Re: Updated

Post by Werner »

Thank you Daniel,
there is quite a difference between you compile from yesterday and today:

Crafty v25.0 (1 cpus) crafty-25.0-x64.exe 1.457.664 27.12.2015 11:33

White(1): bench
Running benchmark. . .
......
Total nodes: 394143343
Raw nodes per second: 5889769
Total elapsed time: 66.92
time used = 1:07

and

Crafty v25.0 (1 cpus) crafty25_64.exe 1.227.264 28.12.2015 08:20

White(1): bench
Running benchmark. . .
......
Total nodes: 143372555
Raw nodes per second: 4828984
Total elapsed time: 29.69
time used = 30.27

looks like another file set for the benchmark in the latest sources?

Or here from a test position:
FEN: r2qrbk1/5ppp/pn1p4/np2P1P1/3p4/5N2/PPB2PP1/R1BQR1K1 w - - 0 1

Crafty-25.0: new version 12-28

26 03:35 1.040.192.134 4.831.137 +0,76 0. ... () 1. e6 Rxe6 2. Rxe6 fxe6 3. Qe1 Qc8 4. Qxa5 Qxc2 5. Qxb6 e5 6. Qb7 Rc8 7. Ne1 Qc6 8. Qxc6 Rxc6 9. Nd3 Rc8 10. f3 Rc4 11. b3 Rc3 12. Nb4 a5 13. Nd5 Rc2
26 13:35 4.143.915.662 5.084.247 +3,48 0. ... () 1. Bxh7+ Kxh7 2. g6+ fxg6 3. Ng5+ Kg8 4. Qf3 Qxg5 5. Bxg5 dxe5 6. Qe4 Re6 7. Rec1 Nac4 8. a4 Rae8 9. axb5 axb5 10. b3 Ba3 11. Rc2 Nd6 12. Qg4 b4 13. Rd1 Nd5 14. f3

Crafty-25.0: compile from yesterday 12-27
16 00:12 70.887.122 6.002.296 +1,01 0. ... () 1. e6 fxe6 2. Bxh7+ Kxh7 3. g6+ Kxg6 4. Qd3+ Kf7 5. Ng5+ Ke7 6. Nxe6 Kd7 7. Nxd8 Rxe1+ 8. Kh2 Rxd8 9. Qd2 Rxc1 10. Rxc1 Nc6 11. Kg3
16 00:32 189.976.860 5.972.237 +3,24 0. ... () 1. Bxh7+ Kxh7 2. g6+ fxg6 3. Ng5+ Kg8 4. Qf3 Qf6 5. exf6 Rxe1+ 6. Kh2 gxf6 7. Qxf6 Ra7 8. Ne6 Rf7 9. Qxg6+ Bg7 10. Rb1 Nd5 11. Bd2
16 00:32 189.979.141 5.972.308 +3,24 0. ... () 1. Bxh7+ Kxh7 2. g6+ fxg6 3. Ng5+ Kg8 4. Qf3 Qf6 5. exf6 Rxe1+ 6. Kh2 gxf6 7. Qxf6 Ra7 8. Ne6 Rf7 9. Qxg6+ Bg7 10. Rb1 Nd5 11. Bd2

So what happened here and which compile should I use for my tests?
Werner
Top
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: Updated

Post by cdani »

Werner wrote:Thank you Daniel,
there is quite a difference between you compile from yesterday and today:

Crafty v25.0 (1 cpus) crafty-25.0-x64.exe 1.457.664 27.12.2015 11:33

White(1): bench
Running benchmark. . .
......
Total nodes: 394143343
Raw nodes per second: 5889769
Total elapsed time: 66.92
time used = 1:07

and

Crafty v25.0 (1 cpus) crafty25_64.exe 1.227.264 28.12.2015 08:20

White(1): bench
Running benchmark. . .
......
Total nodes: 143372555
Raw nodes per second: 4828984
Total elapsed time: 29.69
time used = 30.27

looks like another file set for the benchmark in the latest sources?

Or here from a test position:
FEN: r2qrbk1/5ppp/pn1p4/np2P1P1/3p4/5N2/PPB2PP1/R1BQR1K1 w - - 0 1

Crafty-25.0: new version 12-28

26 03:35 1.040.192.134 4.831.137 +0,76 0. ... () 1. e6 Rxe6 2. Rxe6 fxe6 3. Qe1 Qc8 4. Qxa5 Qxc2 5. Qxb6 e5 6. Qb7 Rc8 7. Ne1 Qc6 8. Qxc6 Rxc6 9. Nd3 Rc8 10. f3 Rc4 11. b3 Rc3 12. Nb4 a5 13. Nd5 Rc2
26 13:35 4.143.915.662 5.084.247 +3,48 0. ... () 1. Bxh7+ Kxh7 2. g6+ fxg6 3. Ng5+ Kg8 4. Qf3 Qxg5 5. Bxg5 dxe5 6. Qe4 Re6 7. Rec1 Nac4 8. a4 Rae8 9. axb5 axb5 10. b3 Ba3 11. Rc2 Nd6 12. Qg4 b4 13. Rd1 Nd5 14. f3

Crafty-25.0: compile from yesterday 12-27
16 00:12 70.887.122 6.002.296 +1,01 0. ... () 1. e6 fxe6 2. Bxh7+ Kxh7 3. g6+ Kxg6 4. Qd3+ Kf7 5. Ng5+ Ke7 6. Nxe6 Kd7 7. Nxd8 Rxe1+ 8. Kh2 Rxd8 9. Qd2 Rxc1 10. Rxc1 Nc6 11. Kg3
16 00:32 189.976.860 5.972.237 +3,24 0. ... () 1. Bxh7+ Kxh7 2. g6+ fxg6 3. Ng5+ Kg8 4. Qf3 Qf6 5. exf6 Rxe1+ 6. Kh2 gxf6 7. Qxf6 Ra7 8. Ne6 Rf7 9. Qxg6+ Bg7 10. Rb1 Nd5 11. Bd2
16 00:32 189.979.141 5.972.308 +3,24 0. ... () 1. Bxh7+ Kxh7 2. g6+ fxg6 3. Ng5+ Kg8 4. Qf3 Qf6 5. exf6 Rxe1+ 6. Kh2 gxf6 7. Qxf6 Ra7 8. Ne6 Rf7 9. Qxg6+ Bg7 10. Rb1 Nd5 11. Bd2

So what happened here and which compile should I use for my tests?
I have done a pgo compile with the latest sources. I really don't know from where comes the difference.
Top
User avatar
Werner
Posts: 2967
Joined: Wed Mar 08, 2006 10:09 pm
Location: Germany
Full name: Werner Schüle

Re: Updated

Post by Werner »

cdani wrote:I have done a pgo compile with the latest sources. I really don't know from where comes the difference.
Of course - if we have changes in the source only Bob can give us an answer.
And the fist compile has been from the Dec 25th sources which was a test version. So I should not compare - and use the last compile for the tests.
Werner
Top
User avatar
F.Huber
Posts: 862
Joined: Thu Mar 09, 2006 4:50 pm
Location: Austria
Full name: Franz Huber

Re: Crafty 25.0 Release

Post by F.Huber »

cdani wrote:All the problems solved! At least I was able to run well on cutechess and Arena.

http://www.andscacs.com/crafty/crafty25.zip

crafty25_32.exe: 32 bit compile able to run on only 1 cpu.
Many thanks Daniel, this one is indeed working (i.e. accepts the hash settings)!

But why did you disable the 'core' setting, i.e. restricting it to a single-CPU version?

Franz
Top
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: Crafty 25.0 Release

Post by cdani »

F.Huber wrote:
cdani wrote:All the problems solved! At least I was able to run well on cutechess and Arena.

http://www.andscacs.com/crafty/crafty25.zip

crafty25_32.exe: 32 bit compile able to run on only 1 cpu.
Many thanks Daniel, this one is indeed working (i.e. accepts the hash settings)!

But why did you disable the 'core' setting, i.e. restricting it to a single-CPU version?

Franz
Because it was crashing, dunno why. Maybe later I will try to find what happens, or maybe someone wants to try.
Top
BBauer
Posts: 658
Joined: Wed Mar 08, 2006 8:58 pm

Re: Crafty 25.0 Release

Post by BBauer »

After a long time I came back to Crafty
It has beautifull output is the fastest in nps and reaches depth 124 in no time. Since I do no longer have the outdatet nalimov bases I tried without.
Look!

Code: Select all

Initializing multiple threads.
System is SMP, not NUMA.
Warning--  xboard 'cores' option disabled
max threads set to 4.
SMP terminate extra threads when idle.
Warning--  xboard 'memory' option disabled
hash table memory = 1G bytes (64M entries).
Warning--  xboard 'memory' option disabled
pawn hash table memory = 384M bytes (16M entries).
EGTB cache memory = 32M bytes.
null move:  R = 3 + depth / 6
search time set to 300.00.
pondering disabled.
book learning disabled
book file disabled.


Crafty v25.0 (4 cpus)

White(1): setboard 8/2p5/6K1/8/8/5k2/8/8 w - - 0 1
White(1): d

       +---+---+---+---+---+---+---+---+
    8  |   | . |   | . |   | . |   | . |
       +---+---+---+---+---+---+---+---+
    7  | . |   |<P>|   | . |   | . |   |
       +---+---+---+---+---+---+---+---+
    6  |   | . |   | . |   | . |-K-| . |
       +---+---+---+---+---+---+---+---+
    5  | . |   | . |   | . |   | . |   |
       +---+---+---+---+---+---+---+---+
    4  |   | . |   | . |   | . |   | . |
       +---+---+---+---+---+---+---+---+
    3  | . |   | . |   | . |<K>| . |   |
       +---+---+---+---+---+---+---+---+
    2  |   | . |   | . |   | . |   | . |
       +---+---+---+---+---+---+---+---+
    1  | . |   | . |   | . |   | . |   |
       +---+---+---+---+---+---+---+---+
         a   b   c   d   e   f   g   h

White(1): go
        time surplus   0.00  time limit 5:00 (5:00)
        depth     time       score   variation (1)
starting thread 1 2 3 <done>
         124->   0.86/5:00     0.00   1. Kf5 Kf2 2. Ke5 Ke2 3. Kd5 Kf3 4. Kc6
                                     Kf4 5. Kxc7 Kg5 6. Kd7 Kh4 7. Kd6 Kg5
                                     8. Kc7 Kh4 9. Kd7 Kg3 10. Kd6 Kh4 11. Kc5
                                     Kg5 12. Kd4 Kg6 13. Ke3 Kg5 14. Kf3 Kh4
                                     15. Kf4 Kh3 16. Kf5 Kh4 17. Kf4 <3-fold>
        time=0.86(57%)  nodes=4066868(4.1M)  fh1=81%  pred=0  nps=4.7M
        chk=7.3K  qchk=0  fp=45.7K  mcp=0  50move=0
        LMReductions:  1/455.0K  2/682.4K  3/500.6K  4/76.6K
        null-move (R):  3/139.7K  4/125.8K  5/89.8K  6/67.9K  7/50.0K
              8/39.3K  9/29.7K  10/22.1K  11/17.9K  12/12.1K  13/10.2K
              14/7.7K  15/17.1K
        splits=267.7K(246.5K)  aborts=3.0K  joins=38.7K  data=18%(18%)
White(1): Kf5
              time used:   0.88
Black(1):
Everything is great :) :) or not? :o
Kind regards
Bernhard
Top
BBauer
Posts: 658
Joined: Wed Mar 08, 2006 8:58 pm

Re: Crafty 25.0 Release

Post by BBauer »

Next try, adding a pawn on a2
[d]8/2p5/6K1/8/8/5k2/P7/8 w - - 0 1

Code: Select all

         38->   3.07/5:00    12.68   1. Kf5 Ke3 2. Ke5 c6 3. a4 Kd3 4. a5 c5
                                     5. a6 c4 6. a7 c3 7. a8=Q c2 8. Qd5+ Ke3
                                     9. Qg2 Kd3 10. Qg5 Kc3 11. Qc1 Kb3 12. Qa1
                                     Kc4 13. Ke4 Kb5 14. Qb2+ Kc6 15. Qxc2+ Kd6
                                     16. Qd2+ Kc6 17. Qd5+ Kb6 18. Kd3 Kc7
                                     19. Kc4 Kc8 20. Kb5 Kc7 21. Kc5 Kb8
                                     22. Qd7 Ka8
         39     4.27/5:00    12.68   1. Kf5 Ke3 2. Ke5 c6 3. a4 Kd3 4. a5 c5
                                     5. a6 c4 6. a7 c3 7. a8=Q c2 8. Qd5+ Ke3
                                     9. Qg2 Kd3 10. Qg5 Kc3 11. Qc1 Kb3 12. Qa1
                                     Kc4 13. Ke4 Kb5 14. Qb2+ Kc6 15. Qxc2+ Kd6
                                     16. Qd2+ Kc6 17. Qd5+ Kb6 18. Kd3 Kc7
                                     19. Kc4 Kc8 20. Kb5 Kc7 21. Kc5 Kb8
                                     22. Qd7 Ka8
         39->   4.28/5:00    12.68   1. Kf5 Ke3 2. Ke5 c6 3. a4 Kd3 4. a5 c5
                                     5. a6 c4 6. a7 c3 7. a8=Q c2 8. Qd5+ Ke3
                                     9. Qg2 Kd3 10. Qg5 Kc3 11. Qc1 Kb3 12. Qa1
                                     Kc4 13. Ke4 Kb5 14. Qb2+ Kc6 15. Qxc2+ Kd6
                                     16. Qd2+ Kc6 17. Qd5+ Kb6 18. Kd3 Kc7
                                     19. Kc4 Kc8 20. Kb5 Kc7 21. Kc5 Kb8
                                     22. Qd7 Ka8
        time=5:00(100%)  nodes=4319111096(4.3B)  fh1=99%  pred=0  nps=14.4M
        chk=787.4M  qchk=304.3M  fp=2.2B  mcp=28.5K  50move=0
        LMReductions:  1/25.2M  2/49.1M  3/72.6M  4/62.0M  5/1.4M  6/4.5K
        null-move (R):  3/116.4M  4/35.2M  5/913.8K  6/98.8K  7/16.9K
              8/2.7K  9/64
        splits=1.0M(993.0K)  aborts=40.3K  joins=50.2K  data=29%(29%)
White(1): Kf5
4 sec to finish ply 39. After 5:00 min nothing has changed.
Kind regards
Bernhard
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Crafty 25.0 Release

Post by bob »

cdani wrote:All the problems solved! At least I was able to run well on cutechess and Arena.

http://www.andscacs.com/crafty/crafty25.zip

4 versions this time:

crafty25_32.exe: 32 bit compile able to run on only 1 cpu.
crafty25_32n.exe: 32 bit compile able to run on only 1 cpu, without popcnt for old cpus.
crafty25_64.exe: 64 bit compile able to run on up to 64 cpu.
crafty25_64n.exe: 64 bit compile able to run on up to 64 cpu, without popcnt for old cpus.

The 32 bit ones are compatible with Windows XP

I had to change size_t by unsigned long long, nothing that I was not aware but I forget :-( I didn't notice previously because as memory command was not working, it didn't crash. Not it works well.

Is compiled with Visual Studio 2013.
Just for fun, which compiler doesn't get "size_t"???
Top

Return to “Computer Chess Club: General Topics”