Crafty 21.6

[bob]

Crafty 21.6 is now available at ftp://ftp.cis.uab.edu/pub/hyatt/

Thanks to Bob Hyatt!

There will be more new versions before long as well, until we stop releasing as we prepare for the next tournament event... then we will pick back up again. This has been a pretty long "dead zone" for releases, but we have tried and discarded a lot of changes. The current version is definitely better than anything we have had in the past, verified by hundreds of thousands of test games...

Bob,

Thats great!

No one change. +lots+ of eval tweaks and a _bunch_ of testing to choose the good ones has made the difference. There's no revolution in the search at all other than the LMR stuff does not use history counters which were not working for me at all... I do not LMR at the root in 21.6 but that is going to change (it was an oversight) in 21.7...

Which of the recent changes do you think have helped the most?

Mark

[bob]

The current version is definitely better than anything we have had in the past, verified by hundreds of thousands of test games...

Thanks,Bob for the release, how much elo improvement did it show in your tests?

Regards.

I don't do those kinds of tests, because I personally do not believe that the Elo claims made today are valid anyway. I simply verify that version N+1 plays better than version N against a suite of good opponents. And I repeat that from time to time as we make changes, always moving forward or not at all...

[bob]

OK, the fix is back in and will be in 21.7. But there is a trick to use it. You will need to add the command "smpnice=1" to your crafty.rc/.craftyrc file. It is definitely less efficient to terminate the processes and then restart them for each search, so that will not be the "normal behavior." But with smpnice=1, it will behave just as you want... Note that you should not play extremely fast games in that mode as time is lost in creating and terminating the processes.

[jwes]

Another small problem is that it will not compile a single cpu version. Try compiling without the /D CPUS=8 and see what happens. It should default to CPUS=1, but doesn't work.

[Jim Ablett]

Hi J.Wesley,

Another small problem is that it will not compile a single cpu version. Try compiling without the /D CPUS=8 and see what happens. It should default to CPUS=1, but doesn't work.

The code only compiles SMP.
All references to

Code: Select all

#if !defined(SMP)

are removed and
replaced with

Code: Select all

#if  (CPUS>1)

but there are some defines missing throughout the src in order to compile for single cpu.

I've fixed the source now for compiling single-cpu version >

http://www.2xupload.de/file/b636f6b6b82 ... u.zip.html

I'll post link for my single-cpu build later.

rgds,
Jim.

[Jim Ablett]

Crafty 21.6 win32 Intel Compiler 10 p.g.o (single-cpu)

http://mihd.net/frxlg3

Jim.

[bob]

Another small problem is that it will not compile a single cpu version. Try compiling without the /D CPUS=8 and see what happens. It should default to CPUS=1, but doesn't work.

Actually it does default CPUS to 1 if not defined, but there was another problem, particularly in egtb.cpp. 21.7 (out soon) will fix this... Unfortunately I don't have a single one-core machine anywhere around so this slipped through.

[Werner]

Hi,
this command does not work:


Initializing multiple threads.
System is SMP, not NUMA.
EPD Kit revision date: 1996.04.21
unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].
hash table memory = 384M bytes.
pawn hash table memory = 24M bytes.
EGTB cache memory = 8 bytes.
use 'settc' command if a game is restarted after crafty
has been terminated for any reason.
tournament mode.
ERROR "smpnice=1 " is unknown rc-file option
max threads set to 2

Crafty v21.6 (2 cpus)

White(1):

[beachknight]

Hi,

the line in crafty.rc

Code: Select all

cache=8

should be replaced with

Code: Select all

cache=8M

EGTB cache memory = 8 bytes.

[Werner]

Thanks,
I set it to 8M
....and my question before was a bit too early as this command only has crafty 21.7 - so sorry about that - I did not read carefully the mail before.