[Computer-go] effectiveness of transposition tables for go

Mark Boon tesujisoftware at gmail.com
Mon May 10 19:38:31 PDT 2010


On Mon, May 10, 2010 at 3:30 PM, Petr Baudis <pasky at ucw.cz> wrote:
> It seems memory issues start to pop up when you try to support very long
> thinking times (minutes per move) or massive influx of simulations
> (cluster computation). Pachi was originally extremely wasteful about
> memory (malloc()ating each single node separately!) and it never really
> mattered (unless you wanted to play very slow on 19x19) until Jean-loup
> had to fix that in order to handle many-core machines sensibly - tens
> of cores simulating in parallel will grow the tree very quickly. ;-)

OK. Then I'm curious, was the solution something along the lines what
Don described?

We have strayed quite a bit from the original topic whether
transpositions help, of course. So far nobody came forward saying it
did for them. I did find an article from some years ago from Bouzy
that claims TT (Transposition Table I suppose) wins 70% vs. without.
The article is a bit more in detail than the one I remember reading a
long time ago and I haven't had time yet to study it in detail.

But I'm just as curious about the memory issues. And solutions people
came up with.

Mark



More information about the Computer-go mailing list