test pychess against another, but modified one

[GoogleCodeExporter]

If somehow you can run pychess engine against a modified one (f.e. with a
changed search method in svn head), you can test if with that modification
you can get faster/better result.

Original issue reported on code.google.com by gbtami on 13 Jan 2007 at 8:07

[GoogleCodeExporter]

If you'd like to, I can add you as developer. Then you can create a branch of 
0.6,
and try out the algorithm.

I will try to make sure, that the new code, which I will commit to the 0.7 
branch as
soon as some of it is working, will contain nearly the same API. It might 
contain an
applyMove and undoMove on the Board class though, as I think it is faster than
cloning the object every time.

If you wan't to test two PyChess.py editions, I use three methods:
1) Start the engine like a normal cecp engine: "PYTHONPATH=lib/ python
lib/pychess/Players/PyChess.py" and type "analyze". It gives you benchmarks for 
each ply.
2) Same as 1, but turn on profiling in PyChess.py to see if certain methods has 
grown
faster.
3) Copy Players/PyChess.py to Players/PyChess_old.py and copy the lines that 
normally
adds the engine in Players/engines.py (line 43-44)

Btw. I think the mailing list would be a good place to discuss things like 
these :)

Original comment by lobais on 14 Jan 2007 at 3:53

[GoogleCodeExporter]

Yes, i'd like to, so thx for the invitation. 

Original comment by gbtami on 14 Jan 2007 at 4:25

[GoogleCodeExporter]

I don't know if people get some kind of email notation, but anyways, you're in 
:)

Conserning the bug issue, then I think it is solved.
When graphical configuration of engines gets written, it will be even easier.

Original comment by lobais on 14 Jan 2007 at 9:18

• Changed state: WontFix

[GoogleCodeExporter]

Got your comment #3 as email. And i'm in! Thx again.

Original comment by gbtami on 15 Jan 2007 at 9:36