[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] owl:11
From: |
kevin yong |
Subject: |
Re: [gnugo-devel] owl:11 |
Date: |
Thu, 10 Apr 2003 14:06:40 -0400 (EDT) |
--- Evan Berggren Daniel <address@hidden>
wrote: > I was looking through the owl test suites,
and
> noticed that owl:11
> performs very poorly. Specifically, it uses 848943
> reading nodes, and
> only one owl node. The reading nodes are used in
> evaluating the O6 eye
> (GTP command eval_eye O6 uses most of the reading
> nodes).
>
> Is there a convenient way to trace what is going on
> in the optics code, eg
> an sgf trace?
Hi, Evan:
i have been trying to figure out a way to trace
what's going in the reading.c process for while. it
has not been completed, but following are what i have
done:
in general, i have been keep modify source code to
acheive what i need, i m not sure this is the right
direction at all.
1. for each specific test case, ussually we have a
specific target (dragon/worm), and a specific move
position to exam. so, at the beginning of
make_worms(), i hard-coded something like:
mainPos = zStrToLoc(board_size,"P5");
mainPos1 = zStrToLoc(board_size,"P6");
in a long term, these could be passed in as
parameters.
2. inside worm.c find_worm_attacks_and_defenses(), i
insert:
gprintf("worm.c find_worm_attacks_and_defenses,
considering defense of %1m\n", str);
mainPos2 = str;
the usage of these is:
3. inside board.c popgo(), i insert:
popgo()
{
if (ZlastAct == TRY_MOVE) {
if (mainPos2 == mainPos) {
dump_stack();
}
}
what these do is: it only dump_stack() when gnugo is
currently exam the target dragon we are interested.
and only print a complete variation list (no printout
at middle of series popgo() to retreat from a complete
variation). it cut the printout file size by 99%.
4. inside reading.c, for each functions of
defend1,defend2,defend3,defend4,attack1,attack2,attack3,attack4,
at each of their return point:
if (str == mainPos) {
gprintf("defend2 for %1m, xpos=%1m,acode=%d\n",
str,xpos,acode);
}
then we know which function examed which pos, (it
can be found from the dump_stack), and what's the
result.
i like to hear your comments about these.
btw: i just installed cygwin, and i m runing a full
test suite before/after remove pattern CD106. i l get
the result back to you very soon.
best regards.
kevin.
______________________________________________________________________
Post your free ad now! http://personals.yahoo.ca