Re: [gnugo-devel] GNU Go 3.3.23

[bump]

> i think Marco meant he ran both versions on 3.2 regressions (the same
> tests).

You are right, this is what he says he did.

> > One thing is that in the regressions, the persistent caches
> > are less used than in real games. I can't quite imagine that
> > that accounts for this.
> 
> i'd guess that it's mostly break-in code, but i'm not sure.  have you
> tried level 9 against 3.2?  and btw, let's not forget that we also have
> a much slower connection code and somewhat slower semeai in this version.
> however, those should be (somewhat) compensated with all the speedups
> we had.

Yes, I reported some level 9 tests yesterday. See:

http://mail.gnu.org/archive/html/gnugo-devel/2003-07/msg00132.html

There are 10 games comparing 3.3.23 level 9 with 3.2 level 10.
In those tests 3.3.23 level 8 was about 4% slower. W won 7/10
of those 2 stone games.

White: 64:32.5
Black: 61:52.9

I am wondering if we should interpolate another level between 8 and 9.
Currently level 9 is a little slow but level 8 is very fast.

Dan