TalkChess.com

Don wrote: Is there any way to get a linux 64 bit version? There are Pascal compilers for Linux, probably the same one you are using.

AFAIK Critter has been written in C++ since version 0.52.

Peter

alpha123 wrote:

Don wrote: Is there any way to get a linux 64 bit version? There are Pascal compilers for Linux, probably the same one you are using.

AFAIK Critter has been written in C++ since version 0.52.

Peter

Maybe I'm thinking of a different program then. But this should make it even simpler to provide a 64 bit version.

I hope Richard is willing to provide a 64 bit version of Critter. It would make a lot of Linux users happy!

Don

Will update my Ladder.

Your work is much appreciated.

Regards.

rvida wrote:Download link: http://www.mediafire.com/?2v80239zxj82yud

Book support

In addition to its own .cbk book format, Critter can now read Polyglot .bin books too. Just set the "Book File" UCI option to point to a *.bin file. Book creation from pgn file ('makebook' and 'mergebook' console commands) works only with native format.

Repetition detection

Critter now scores 2-fold repetition as draw. See this thread on rybka forum.

Enjoy

Richard

Congratulations Richard !!!

It is amazing how quickly you are able to climb on the ELO stair. I know very well how much effort it takes to get another 20-30 ELO when you reach certain levels and this is why your results greatly impress me.

...next station is Rybka 4

BTW only less than 2 years ago Rybka 3 was the undisputed king, well ahead of competitors and considered unreachable. Now we have at least four free engines stronger than R3. This is impressive !!!

Don wrote: Is there any way to get a linux 64 bit version?

I would be very happy to provide a linux build, I see there are a lot of people who would appreciate it, but I'm afraid it could take quite some time...

Not long ago I was able to compile my sources with gcc/linux, but something went wrong with my SMP code and the binary always crashed with more than 1 thread.

I made a quick test with GCC + Critter 1.0 and apart of some warnings it did compile without errors, but it does crash with the first 'go' command (even single-threaded)

Now here is my plan:
1 - make it compile with GCC under Windows [done]
2 - get rid of any warnings even with -Wall
3 - make the GCC-compiled binary work as intended (same node count in start position with 1 thread)
4 - boot into Ubuntu / compile with gcc / learn how to use gdb
5 - find out what's wrong with the SMP code

If everything fails, I will consider sending my sources to someone with more experience with linux+gcc for advice and/or a working binary...

Richard

It may also be worth checking whether the problems persist under GCC for Windows (without Cygwin) with -O0. GCC has been known to optimize aggressively based on technicalities in the C/C++ standards. The latest versions have improved in some respects.

rvida wrote: Not long ago I was able to compile my sources with gcc/linux, but something went wrong with my SMP code and the binary always crashed with more than 1 thread.

If you use CRITICAL_SECTION, then be warned that locking semantic is different with pthreads that are more similar to Windows's SRWLOCK (see SF's lock.h file).

rvida wrote:

Don wrote: Is there any way to get a linux 64 bit version?

I would be very happy to provide a linux build, I see there are a lot of people who would appreciate it, but I'm afraid it could take quite some time...

That would be great. I would be happy with a single processor version.

Not long ago I was able to compile my sources with gcc/linux, but something went wrong with my SMP code and the binary always crashed with more than 1 thread.

I made a quick test with GCC + Critter 1.0 and apart of some warnings it did compile without errors, but it does crash with the first 'go' command (even single-threaded)

Now here is my plan:
1 - make it compile with GCC under Windows [done]
2 - get rid of any warnings even with -Wall
3 - make the GCC-compiled binary work as intended (same node count in start position with 1 thread)
4 - boot into Ubuntu / compile with gcc / learn how to use gdb
5 - find out what's wrong with the SMP code

If everything fails, I will consider sending my sources to someone with more experience with linux+gcc for advice and/or a working binary...

possibly Jim Ablett might be able to make it work.

Richard

Great work Richard! It has almost reached Stockfish level - in 100 games with dual Critter lost 49,5-50,5 against SF2 But more importantly it has not lost stunning tactics/analysis speed of Critter 0.90. After about 2000 testpositions I dare to say Critter 1 is better/faster than 0.90. Even better than Houdini in same suites.

Jouni

rvida wrote:

Don wrote: Is there any way to get a linux 64 bit version?

I would be very happy to provide a linux build, I see there are a lot of people who would appreciate it, but I'm afraid it could take quite some time...

Not long ago I was able to compile my sources with gcc/linux, but something went wrong with my SMP code and the binary always crashed with more than 1 thread.

I made a quick test with GCC + Critter 1.0 and apart of some warnings it did compile without errors, but it does crash with the first 'go' command (even single-threaded)

Now here is my plan:
1 - make it compile with GCC under Windows [done]
2 - get rid of any warnings even with -Wall
3 - make the GCC-compiled binary work as intended (same node count in start position with 1 thread)
4 - boot into Ubuntu / compile with gcc / learn how to use gdb
5 - find out what's wrong with the SMP code

If everything fails, I will consider sending my sources to someone with more experience with linux+gcc for advice and/or a working binary...

Richard

I am impressed and it is looking very good. I still hope you'll consider the option for an aggressive (a la 0.90) version or on/off parameter in the UCI settings.

It might cost it some Elo when activated, but would make watching the games that much more fun.

BTW, I have passed on your new engine to a few GMs.