GC slower via XBoard than via glChess

[dae3]

Hi all,

I'm experiencing slow performance by GNU Chess (5.07) when the program
is called by XBoard (4.2.7). The engine typically takes 10 seconds to
make its first move (as black), whereas the same move is instantaneous
when called by glChess (1.0.2).

I'm issuing the following command to run XBoard:

"xboard -xclock -size 45 -flash -depth 3 -xponder &".

Omitting the -xponder option doesn't seem to improve responsiveness.

Does anyone know what might cause the slowdown?

Thanks in advance,
dae3

--
~> cat /etc/redhat-release && uname -a
Scientific Linux SL release 5.1 (Boron)
Linux acerrimo.local 2.6.18-92.1.6.el5 #1 SMP Wed Jun 25 12:46:39 EDT
2008 i686 athlon i386 GNU/Linux

[dae3]

Is this list still alive or has discussion moved elsewhere?

Pointers are welcome.

TIA

--
~> cat /etc/redhat-release && uname -a
Scientific Linux SL release 5.1 (Boron)

Linux acerrimo.local 2.6.18-92.1.10.el5 #1 SMP Mon Aug 4 14:38:32 EDT

[dae3]

Simon Waters <si...@technocool.net> wrote:

> Add "-debug" and show us the output.

Ok, I'm attaching XBoard/GnuChess' output. The first move by the
computer (black) took about 10 seconds, the second move about 9 seconds.

Thank you,
dae3

-----BEGIN ATTACHED TEXT-----
~> xboard -xclock -size 45 -flash -depth 3 -xponder -debug
recognized 'normal' (-1) as variant normal
resolved -*-helvetica-bold-r-normal--*-*-*-*-*-*-*-* at pixel size 20
to -adobe-helvetica-bold-r-normal--20-140-100-100-p-105-iso8859-1
resolved -*-helvetica-bold-r-normal--*-*-*-*-*-*-*-* at pixel size 10
to -adobe-helvetica-bold-r-normal--10-100-75-75-p-60-iso8859-1
resolved -*-helvetica-medium-r-normal--*-*-*-*-*-*-*-* at pixel size 12
to -adobe-helvetica-medium-r-normal--12-120-75-75-p-67-iso8859-1
xboard 4.2.7 + gnuchessx
Reset(1, 0) from gameMode 0
recognized 'normal' (-1) as variant normal
GameEnds(0, (null), 2)
StartChildProcess (dir="") gnuchessx
754 >first : xboard
protover 2
1043 <first : Chess
1043 <first : Adjusting HashSize to 1024 slots
1043 <first : feature setboard=1 analyze=1 ping=1 draw=0 sigint=0 variants="normal" myname="GNU Chess 5.07" done=1
1043 >first : accepted setboard
1043 >first : accepted analyze
1043 >first : accepted ping
1043 >first : accepted draw
1043 >first : accepted sigint
1043 >first : accepted variants
1043 >first : accepted myname
1043 >first : accepted done
1044 >first : new
random
1044 >first : level 40 5 0
1044 >first : sd 3
1044 >first : hard
1044 >first : easy
1044 >first : ping 1
1357 <first : Illegal move: sd 3
1392 <first : pong 1
23342 >first : time 30000
otim 30000
23342 >first : d2d4
23980 <first : 1. d2d4
33989 <first : Search time bonus near book
33989 <first : 1. ... d7d5
AnimateMove: piece 6 slides from 3,6 to 3,4
34676 <first : My move is: d7d5
46689 >first : time 30000
otim 30000
46689 >first : c2c4
47303 <first : 2. c2c4
55148 <first : Search time bonus near book
55150 <first : 2. ... d5c4
AnimateMove: piece 6 slides from 3,4 to 2,3
55838 <first : My move is: d5c4
GameEnds(0, (null), 2)
68388 >first : force
68388 >first : ping 2
68388 >first : quit
~>
------END ATTACHED TEXT------

[Simon Waters]

dae3 wrote:
>
> 33989 <first : Search time bonus near book

I'm not sure this should appear in the debug output, but it is useful as
it suggests there is no opening book available.

With 5 minutes for the rest of the game, and no opening book, I'd expect
it to take a few seconds to make the first few moves. The "bonus" is
because the program would often go wrong on leaving book due to
inadequate search.

Does the other program (glChess) manage to start gnuchess with an
opening book perhaps?

[dae3]

Il 26/10/2008 alle 19:16, Simon Waters ha scritto:

> I'm not sure this should appear in the debug output, but it is useful as
> it suggests there is no opening book available.
>
> With 5 minutes for the rest of the game, and no opening book, I'd expect
> it to take a few seconds to make the first few moves.

Actually, I'm using the "-xclock" option to avoid time limits.

> The "bonus" is because the program would often go wrong on leaving
> book due to inadequate search.
>
> Does the other program (glChess) manage to start gnuchess with an
> opening book perhaps?

I think I haven't got any opening books installed: there's nothing for
GnuChess in /usr/share/games, and a 'locate -i gnuchess' only finds the
tarball's directory and 3 binaries:

/usr/local/bin/gnuchess
/usr/local/bin/gnuchessx
/usr/local/bin/gnuchessx.org

Unfortunately, glChess doesn't have a -debug option, but a 'lsof'
reports that the GnuChess launched by glChess has the following files
open, both at game start and after the first 2 moves:

gnuchess 3676 dae3 cwd DIR 253,0 20480 2717946 /home/dae3
gnuchess 3676 dae3 rtd DIR 253,0 4096 2 /
gnuchess 3676 dae3 txt REG 253,0 202868 8020892 /usr/local/bin/gnuchess
gnuchess 3676 dae3 mem REG 253,0 125736 2685795 /lib/ld-2.5.so
gnuchess 3676 dae3 mem REG 253,0 1585788 2685809 /lib/libc-2.5.so
gnuchess 3676 dae3 mem REG 253,0 16528 2685810 /lib/libdl-2.5.so
gnuchess 3676 dae3 mem REG 253,0 125668 2685811 /lib/libpthread-2.5.so
gnuchess 3676 dae3 mem REG 253,0 210256 6567418 /usr/lib/libreadline.so.5.1
gnuchess 3676 dae3 mem REG 253,0 297464 6583562 /usr/lib/libncurses.so.5.5
gnuchess 3676 dae3 mem REG 253,0 56413600 6551462 /usr/lib/locale/locale-archive
gnuchess 3676 dae3 mem REG 253,0 25462 6645668 /usr/lib/gconv/gconv-modules.cache
gnuchess 3676 dae3 0u CHR 136,6 8 /dev/pts/6
gnuchess 3676 dae3 1u CHR 136,6 8 /dev/pts/6
gnuchess 3676 dae3 2u CHR 136,6 8 /dev/pts/6
gnuchess 3676 dae3 10r FIFO 0,6 75129 pipe
gnuchess 3676 dae3 11w FIFO 0,6 75129 pipe
gnuchess 3676 dae3 12r FIFO 0,6 75130 pipe
gnuchess 3676 dae3 13w FIFO 0,6 75130 pipe
gnuchess 3676 dae3 14r FIFO 0,6 75131 pipe
gnuchess 3676 dae3 15w FIFO 0,6 75131 pipe

Not knowing GnuChess' internals, I'm not sure that the above is
relevant, but it looks like there is no book file open at the moment.

Cheers,
dae3