[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] Metamachine's first game
From: |
Dave Denholm |
Subject: |
Re: [gnugo-devel] Metamachine's first game |
Date: |
13 Nov 2002 15:20:09 +0000 |
<address@hidden> writes:
> I've posted a version of the oracle patch capable of
> playing a game. Compile the patch oracle_3_12.1d with
> configure --enable-oracle and run it in CGoban with the
> --metamachine command line option.
>
> This version can play using the GMP. It will need a minor
> change to play in ascii mode or using the gtp.
>
> The first game is as follows. I think CGoban stopped the
> game after the metamachine ran out of time, using an hour
> of time to play 34 moves. Probably we can get acceptable
> speed if we only search a tree when there are two large
> moves on the board.
>
Haven't paid close attention to how the metamachine works.
Does the main process think in parallel to the metamachine,
or does gnugo ask the metamachine a question then wait for
an answer ?
If the former, then it would be a simple matter to use
a socket to run the oracle on a separate machine. It's
a simple matter to plug it into inetd to be spawned
on demand with stdin/out connected to the socket.
'netcat' is a useful utility for plumbing these sorts of
things together. Eg gnuplot could spawn a script which either execs the
metamachine directly, or netcat's to a separate machine via a socket.
The main gnugo executable wouldn't be able to tell the difference.
dd
--
address@hidden http://www.insignia.com