Question about Crafty

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

Moderators: hgm, Dann Corbit, Harvey Williamson

cooldalek

Question about Crafty

Post by cooldalek »

This is really a question for Dr Hyatt but others may know the answer too.

I have heard it said that the Crafty project aims to explore areas of improvements to chess engines on multiple core machines (i.e. the programming effort concentrates on efficient use of multi-core machines a bit like Zap!)

Is this true please? And does it mean that future versions of crafty for a single core 32 bit CPU user like myself will not improve greatly in strength?

It's not a complaint if it doesn't increase much in strength; I really enjoy its playing style

Kind regards

Laurence
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Question about Crafty

Post by bob »

cooldalek wrote:This is really a question for Dr Hyatt but others may know the answer too.

I have heard it said that the Crafty project aims to explore areas of improvements to chess engines on multiple core machines (i.e. the programming effort concentrates on efficient use of multi-core machines a bit like Zap!)

Is this true please? And does it mean that future versions of crafty for a single core 32 bit CPU user like myself will not improve greatly in strength?

It's not a complaint if it doesn't increase much in strength; I really enjoy its playing style

Kind regards

Laurence

We have been busily modifying Crafty's evaluation for a year now and are about to release the source. There are parallel search changes, normal search changes, and a ton of evaluation changes. We are working on all fronts in fact...
Top
skoony
Posts: 165
Joined: Thu Mar 09, 2006 2:36 am
Location: saint paul,minnesota

Re: Question about Crafty

Post by skoony »

tanks for the info and good work prof. hyatt.

regards
mike
by the time i get there,i'll be there.
Top
BBauer
Posts: 658
Joined: Wed Mar 08, 2006 8:58 pm

Re: Question about Crafty

Post by BBauer »

And I hope to get a Makefile which will run without a lot of warnings on a 64-bit linux with gcc-4.2, including profiling and correct support for tablebases.
My build of 20.14 runs on an AMD X2 very fast, 5.6 Mnps and says it uses heavily tablebase, but in fact don't uses them, perhaps too high optimisation level.
It's time for a new version.
regards
Bernhard
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Question about Crafty

Post by bob »

BBauer wrote:And I hope to get a Makefile which will run without a lot of warnings on a 64-bit linux with gcc-4.2, including profiling and correct support for tablebases.
My build of 20.14 runs on an AMD X2 very fast, 5.6 Mnps and says it uses heavily tablebase, but in fact don't uses them, perhaps too high optimisation level.
It's time for a new version.
regards
Bernhard
If crafty says it is probing and probes are successful, they _must_ be happening. There are no known EGTB bugs in crafty that I am aware of.

The standard makefile has 64 bit options for gcc and icc, including profiling although you need to modify the Makefile to change the "profile:" script to use the right build scripts depending on which of the above two compilers you use. I'd use intel's free compiler as it is better, period...
Top
BBauer
Posts: 658
Joined: Wed Mar 08, 2006 8:58 pm

Re: Question about Crafty

Post by BBauer »

Here is a run for a simple 6-piece endgame.
I use 5-piece table bases, 290 files.

$ ./crafty< aw229
unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].

Crafty v20.14 (1 cpus)

White(1): White(1): max threads set to 2
White(1): noise level set to 500000.
White(1): pondering disabled.
White(1): book file disabled.
White(1): book learning disabled
result learning disabled
position learning disabled
White(1): null depth set to 2/3 (min/max)
White(1): White(1): one-reply extension..................0.75
in-check extension...................0.50
mate thrt extension..................0.75
White(1): hash table memory = 384M bytes.
White(1): pawn hash table memory = 128M bytes.
White(1): 5 piece tablebase files found
13994kb of RAM used for TB indices and decompression tables
White(1): EGTB cache memory = 128 bytes.
White(1): search time set to 30.00.
White(1): White(1):
+---+---+---+---+---+---+---+---+
8 | | . |<K>| . | | . | | . |
+---+---+---+---+---+---+---+---+
7 | . | | . | | . | | . | |
+---+---+---+---+---+---+---+---+
6 | | . | |-K-| | . | | . |
+---+---+---+---+---+---+---+---+
5 | . |<P>| . | | . | | . | |
+---+---+---+---+---+---+---+---+
4 |<P>| . | | . | | . | | . |
+---+---+---+---+---+---+---+---+
3 |-P-| | . | | . | | . | |
+---+---+---+---+---+---+---+---+
2 | |-B-| | . | | . | | . |
+---+---+---+---+---+---+---+---+
1 | . | | . | | . | | . | |
+---+---+---+---+---+---+---+---+
a b c d e f g h

White(1): time surplus 0.00 time limit 30.00 (+0.00) (30.00)
depth time score variation (1)
starting thread 1
13 0.30 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Bd4 Kc8 7. Be5
13-> 0.39 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Bd4 Kc8 7. Be5
(s=2)
14 0.48 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7
14-> 0.67 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7 (s=2)
15 0.76 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7 8. Kc6
15-> 0.84 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7 8. Kc6
16 0.92 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT>
16-> 1.00 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT>
17 1.07 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT>
17-> 1.28 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT> (s=5)
18 1.39 5.49 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Bc3+
Ke4 8. Kxb5 Ke3 9. Kxa4 Kd3 10. Be5
(s=4)
18-> 1.64 5.49 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Bc3+
Ke4 8. Kxb5 Ke3 9. Kxa4 Kd3 10. Be5
(s=3)
19 1.75 5.65 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Kxb5
Kd5 8. Bc5 Ke4 9. Kxa4 Kd3 10. Kb5
(s=2)
19-> 2.07 5.65 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Kxb5
Kd5 8. Bc5 Ke4 9. Kxa4 Kd3 10. Kb5
(s=5)
20 2.30 5.65 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Kxb5
Kd5 8. Bc5 Ke4 9. Kxa4 Kd3 10. Kb5
Kc3 (s=4)
20 2.97 +1 1. Bf6!!
20 3.02 +3 1. Bf6!!
20 3.07 +M 1. Bf6!!
20 4.17 6.49 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 Ka8 5. Kc6 Kb8 6. Bd8 Ka8 7. Bc7
b4 8. axb4 a3 9. Be5 a2 10. Kb6 a1=Q
11. Bxa1
20-> 4.23 6.49 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 Ka8 5. Kc6 Kb8 6. Bd8 Ka8 7. Bc7
b4 8. axb4 a3 9. Be5 a2 10. Kb6 a1=Q
11. Bxa1
21 4.24 +1 1. Bf6!!
21 4.24 +3 1. Bf6!!
21 4.24 +M 1. Bf6!!
21 4.47 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. b5 Ke6 9. Ba1 Kf5 10. b6 Ke4
11. b7 Kd3 12. b8=Q <HT>
21-> 4.99 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. b5 Ke6 9. Ba1 Kf5 10. b6 Ke4
11. b7 Kd3 12. b8=Q <HT> (s=3)
22 5.08 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=2)
22-> 5.78 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=9)
23 5.99 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=8)
23-> 6.71 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=3)
24 7.00 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 <HT> (s=2)
24-> 8.23 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 <HT> (s=3)
25 8.40 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 <HT> (s=2)
25-> 9.77 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 <HT> (s=7)
26 10.21 11.50 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 a2 10. Kc6 Kf5
11. b6 Ke4 12. Kc7 Kd3 13. b7 Kc4 (s=6)
26-> 11.39 11.50 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 a2 10. Kc6 Kf5
11. b6 Ke4 12. Kc7 Kd3 13. b7 Kc4 (s=3)
27 11.82 11.67 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 a2 9. b5 Kc7 10. Bd4 Kd8
11. b6 <HT> (s=2)
27-> 14.22 11.67 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 a2 9. b5 Kc7 10. Bd4 Kd8
11. b6 <HT> (s=6)
28 14.42 +1 1. Bf6!!
28 14.47 +3 1. Bf6!!
28 15.43 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 <HT>
(s=5)
28-> 17.17 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 <HT>
(s=2)
29 17.92 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3
29-> 20.77 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 (s=3)
30 22.07 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 16. Bh8 (s=2)
30-> 27.59 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 16. Bh8 (s=4)
31 29.71 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 16. Bh8 Kd3
(s=3)
time=30.42 mat=2 n=180489762 fh=98% nps=5.9M
ext-> check=13.3M 1rep=219K mate=10K reductions=68.7M/65.2M
predicted=0 evals=81.5M 50move=0 EGTBprobes=3.3M hits=3.3M
SMP-> splits=12750 aborts=533 data=9/128 elap=30.42


So you see I get a depth of 31, nps=5.9M, EGTBprobes=3.3M, but I get
no mate score. Tablebases are in fact _not_ used. CPU usage is 100%.

The standard Makefile doesn't work for me.
I do: make linux

linux:
$(MAKE) target=LINUX \
CC=gcc-4.2 CXX=g++-4.2 \
CFLAGS='$(CFLAGS) -Wall -pipe -D_REENTRANT -march=athlon64 -O3 \
-fomit-frame-pointer\
-fno-gcse ' \
CXFLAGS=$(CFLAGS) \
LDFLAGS='$(LDFLAGS) -lstdc++' \
opt='$(opt) -DINLINE64 -DSMP -DCPUS=2' \
crafty-make

When running the executable cpu is always at 100 %. That's why I think Crafty is not using tablebases.
Here my warnings:
gcc-4.2 -Wall -pipe -D_REENTRANT -march=athlon64 -O3 \
-fomit-frame-pointer\
-fno-gcse -DINLINE64 -DSMP -DCPUS=2 -DLINUX -c crafty.c
In file included from crafty.c:1:
searchr.c: In Funktion »SearchTrace«:
searchr.c:280: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 3 hat Typ »BITBOARD«
In file included from crafty.c:3:
thread.c: In Funktion »ThreadInit«:
thread.c:176: Warnung: Typkonvertierung von Zeiger auf Ganzzahl anderer Breite
thread.c:188: Warnung: Typkonvertierung von Zeiger auf Ganzzahl anderer Breite
In file included from crafty.c:19:
utility.c: In Funktion »Kibitz«:
utility.c:2448: Warnung: format »%6llu« erwartet Typ »long long unsigned int«, aber Argument 5 hat Typ »BITBOARD«
utility.c:2451: Warnung: format »%6llu« erwartet Typ »long long unsigned int«, aber Argument 5 hat Typ »BITBOARD«
In file included from crafty.c:36:
interupt.c: In Funktion »Interrupt«:
interupt.c:80: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 2 hat Typ »BITBOARD«
interupt.c:92: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 2 hat Typ »BITBOARD«
In file included from crafty.c:40:
option.c: In Funktion »Option«:
option.c:2480: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 2 hat Typ »BITBOARD«
crafty.c: Auf höchster Ebene:
utility.c:55: Warnung: »clk_tck« definiert, aber nicht verwendet
annotate.c: In Funktion »Annotate«:
annotate.c:91: Warnung: »player_pv.path[1]« könnte in dieser Funktion uninitialisiert verwendet werden
g++-4.2 -c -DINLINE64 -DSMP -DCPUS=2 -DLINUX egtb.cpp
egtb.cpp:4493: Warnung: »TB_CRC_CHECK« initialisiert und als »extern« deklariert
gcc-4.2 -lstdc++ -o crafty crafty.o egtb.o -lm


Intel profiling didn't work either, perhaps I use a newer version.
.
Regards
Bernhard
Top
jwes
Posts: 778
Joined: Sat Jul 01, 2006 7:11 am

Re: Question about Crafty

Post by jwes »

BBauer wrote:Here is a run for a simple 6-piece endgame.
White(1): EGTB cache memory = 128 bytes.
128 bytes seems a little small.
Top
BBauer
Posts: 658
Joined: Wed Mar 08, 2006 8:58 pm

Re: Question about Crafty

Post by BBauer »

Thank you!
I removed the M by accident. Now I get, what I expected.

Code: Select all

               24->  29.25  Mat20   1. Bf6 Kb7 2. Bd4 Kc8 3. Bb6 Kb7 4.
                                    Kc5 b4 5. axb4 <EGTB> &#40;s=4&#41;
               25    29.76  Mat20   1. Bf6 Kb7 2. Bd4 Kc8 3. Bb6 Kb7 4.
                                    Kc5 b4 5. axb4 <EGTB> &#40;s=3&#41;
              time=30.27  mat=2  n=6520932  fh=98%  nps=215K
              ext-> check=139K 1rep=932 mate=0 reductions=3.3M/3.2M
              predicted=0  evals=1.3M  50move=0  EGTBprobes=72K  hits=72K
              SMP->  splits=3142  aborts=200  data=8/128  elap=30.27

mate in 20 moves.
May be I discovered a new feature in crafty :D
Kind regards
Bernhard
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Question about Crafty

Post by bob »

jwes wrote:
BBauer wrote:Here is a run for a simple 6-piece endgame.
White(1): EGTB cache memory = 128 bytes.
128 bytes seems a little small.
It is more than small. It will result in zero reads as there are no 128 byte blocks in an egtb. :)
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Question about Crafty

Post by bob »

BBauer wrote:Here is a run for a simple 6-piece endgame.
I use 5-piece table bases, 290 files.

$ ./crafty< aw229
unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].

Crafty v20.14 (1 cpus)

White(1): White(1): max threads set to 2
White(1): noise level set to 500000.
White(1): pondering disabled.
White(1): book file disabled.
White(1): book learning disabled
result learning disabled
position learning disabled
White(1): null depth set to 2/3 (min/max)
White(1): White(1): one-reply extension..................0.75
in-check extension...................0.50
mate thrt extension..................0.75
White(1): hash table memory = 384M bytes.
White(1): pawn hash table memory = 128M bytes.
White(1): 5 piece tablebase files found
13994kb of RAM used for TB indices and decompression tables
White(1): EGTB cache memory = 128 bytes.

That looks a bit ugly. Did you forget an "M" on the end to make it 128M? EGTBs are compressed into blocks of (I believe) 8K bytes. That size will make it impossible to do any egtb reads. :)


White(1): search time set to 30.00.
White(1): White(1):
+---+---+---+---+---+---+---+---+
8 | | . |<K>| . | | . | | . |
+---+---+---+---+---+---+---+---+
7 | . | | . | | . | | . | |
+---+---+---+---+---+---+---+---+
6 | | . | |-K-| | . | | . |
+---+---+---+---+---+---+---+---+
5 | . |<P>| . | | . | | . | |
+---+---+---+---+---+---+---+---+
4 |<P>| . | | . | | . | | . |
+---+---+---+---+---+---+---+---+
3 |-P-| | . | | . | | . | |
+---+---+---+---+---+---+---+---+
2 | |-B-| | . | | . | | . |
+---+---+---+---+---+---+---+---+
1 | . | | . | | . | | . | |
+---+---+---+---+---+---+---+---+
a b c d e f g h

White(1): time surplus 0.00 time limit 30.00 (+0.00) (30.00)
depth time score variation (1)
starting thread 1
13 0.30 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Bd4 Kc8 7. Be5
13-> 0.39 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Bd4 Kc8 7. Be5
(s=2)
14 0.48 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7
14-> 0.67 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7 (s=2)
15 0.76 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7 8. Kc6
15-> 0.84 5.35 1. Kc6 Kd8 2. Bd4 Ke8 3. Kxb5 Kd7 4.
Be5 Kc8 5. Kc6 Kd8 6. Kd6 Ke8 7. Bd4
Kf7 8. Kc6
16 0.92 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT>
16-> 1.00 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT>
17 1.07 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT>
17-> 1.28 5.42 1. Kc6 Kd8 2. Bf6+ Ke8 3. Kd6 Kf7 4.
Bd4 Ke8 5. Kc7 <HT> (s=5)
18 1.39 5.49 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Bc3+
Ke4 8. Kxb5 Ke3 9. Kxa4 Kd3 10. Be5
(s=4)
18-> 1.64 5.49 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Bc3+
Ke4 8. Kxb5 Ke3 9. Kxa4 Kd3 10. Be5
(s=3)
19 1.75 5.65 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Kxb5
Kd5 8. Bc5 Ke4 9. Kxa4 Kd3 10. Kb5
(s=2)
19-> 2.07 5.65 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Kxb5
Kd5 8. Bc5 Ke4 9. Kxa4 Kd3 10. Kb5
(s=5)
20 2.30 5.65 1. Kc6 Kd8 2. Bc3 Ke7 3. Kc7 Ke6 4.
Bb4 Kf7 5. Kd7 Kf6 6. Kc6 Ke5 7. Kxb5
Kd5 8. Bc5 Ke4 9. Kxa4 Kd3 10. Kb5
Kc3 (s=4)
20 2.97 +1 1. Bf6!!
20 3.02 +3 1. Bf6!!
20 3.07 +M 1. Bf6!!
20 4.17 6.49 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 Ka8 5. Kc6 Kb8 6. Bd8 Ka8 7. Bc7
b4 8. axb4 a3 9. Be5 a2 10. Kb6 a1=Q
11. Bxa1
20-> 4.23 6.49 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 Ka8 5. Kc6 Kb8 6. Bd8 Ka8 7. Bc7
b4 8. axb4 a3 9. Be5 a2 10. Kb6 a1=Q
11. Bxa1
21 4.24 +1 1. Bf6!!
21 4.24 +3 1. Bf6!!
21 4.24 +M 1. Bf6!!
21 4.47 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. b5 Ke6 9. Ba1 Kf5 10. b6 Ke4
11. b7 Kd3 12. b8=Q <HT>
21-> 4.99 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. b5 Ke6 9. Ba1 Kf5 10. b6 Ke4
11. b7 Kd3 12. b8=Q <HT> (s=3)
22 5.08 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=2)
22-> 5.78 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=9)
23 5.99 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=8)
23-> 6.71 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 <HT> (s=3)
24 7.00 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 <HT> (s=2)
24-> 8.23 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 <HT> (s=3)
25 8.40 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 <HT> (s=2)
25-> 9.77 11.48 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 <HT> (s=7)
26 10.21 11.50 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 a2 10. Kc6 Kf5
11. b6 Ke4 12. Kc7 Kd3 13. b7 Kc4 (s=6)
26-> 11.39 11.50 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 Ke6 9. b5 a2 10. Kc6 Kf5
11. b6 Ke4 12. Kc7 Kd3 13. b7 Kc4 (s=3)
27 11.82 11.67 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 a2 9. b5 Kc7 10. Bd4 Kd8
11. b6 <HT> (s=2)
27-> 14.22 11.67 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kd7 8. Ba1 a2 9. b5 Kc7 10. Bd4 Kd8
11. b6 <HT> (s=6)
28 14.42 +1 1. Bf6!!
28 14.47 +3 1. Bf6!!
28 15.43 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 <HT>
(s=5)
28-> 17.17 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 <HT>
(s=2)
29 17.92 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3
29-> 20.77 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 (s=3)
30 22.07 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 16. Bh8 (s=2)
30-> 27.59 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 16. Bh8 (s=4)
31 29.71 13.51 1. Bf6 Kb7 2. Bd8 Kc8 3. Bb6 Kb7 4.
Kc5 b4 5. axb4 a3 6. Bd8 Kc8 7. Bf6
Kb7 8. b5 Kc7 9. b6+ Kd7 10. Bg7 Ke6
11. b7 Kf5 12. b8=Q a2 13. Ba1 Kg4
14. Qg8+ Kf3 15. Qxa2 Ke3 16. Bh8 Kd3
(s=3)
time=30.42 mat=2 n=180489762 fh=98% nps=5.9M
ext-> check=13.3M 1rep=219K mate=10K reductions=68.7M/65.2M
predicted=0 evals=81.5M 50move=0 EGTBprobes=3.3M hits=3.3M
SMP-> splits=12750 aborts=533 data=9/128 elap=30.42


So you see I get a depth of 31, nps=5.9M, EGTBprobes=3.3M, but I get
no mate score. Tablebases are in fact _not_ used. CPU usage is 100%.

The standard Makefile doesn't work for me.
I do: make linux

linux:
$(MAKE) target=LINUX \
CC=gcc-4.2 CXX=g++-4.2 \
CFLAGS='$(CFLAGS) -Wall -pipe -D_REENTRANT -march=athlon64 -O3 \
-fomit-frame-pointer\
-fno-gcse ' \
CXFLAGS=$(CFLAGS) \
LDFLAGS='$(LDFLAGS) -lstdc++' \
opt='$(opt) -DINLINE64 -DSMP -DCPUS=2' \
crafty-make

When running the executable cpu is always at 100 %. That's why I think Crafty is not using tablebases.
Here my warnings:
gcc-4.2 -Wall -pipe -D_REENTRANT -march=athlon64 -O3 \
-fomit-frame-pointer\
-fno-gcse -DINLINE64 -DSMP -DCPUS=2 -DLINUX -c crafty.c
In file included from crafty.c:1:
searchr.c: In Funktion »SearchTrace«:
searchr.c:280: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 3 hat Typ »BITBOARD«
In file included from crafty.c:3:
thread.c: In Funktion »ThreadInit«:
thread.c:176: Warnung: Typkonvertierung von Zeiger auf Ganzzahl anderer Breite
thread.c:188: Warnung: Typkonvertierung von Zeiger auf Ganzzahl anderer Breite
In file included from crafty.c:19:
utility.c: In Funktion »Kibitz«:
utility.c:2448: Warnung: format »%6llu« erwartet Typ »long long unsigned int«, aber Argument 5 hat Typ »BITBOARD«
utility.c:2451: Warnung: format »%6llu« erwartet Typ »long long unsigned int«, aber Argument 5 hat Typ »BITBOARD«
In file included from crafty.c:36:
interupt.c: In Funktion »Interrupt«:
interupt.c:80: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 2 hat Typ »BITBOARD«
interupt.c:92: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 2 hat Typ »BITBOARD«
In file included from crafty.c:40:
option.c: In Funktion »Option«:
option.c:2480: Warnung: format »%llu« erwartet Typ »long long unsigned int«, aber Argument 2 hat Typ »BITBOARD«
crafty.c: Auf höchster Ebene:
utility.c:55: Warnung: »clk_tck« definiert, aber nicht verwendet
annotate.c: In Funktion »Annotate«:
annotate.c:91: Warnung: »player_pv.path[1]« könnte in dieser Funktion uninitialisiert verwendet werden
g++-4.2 -c -DINLINE64 -DSMP -DCPUS=2 -DLINUX egtb.cpp
egtb.cpp:4493: Warnung: »TB_CRC_CHECK« initialisiert und als »extern« deklariert
gcc-4.2 -lstdc++ -o crafty crafty.o egtb.o -lm


Intel profiling didn't work either, perhaps I use a newer version.
.
Regards
Bernhard
Top

Return to “Computer Chess Club: General Topics”