[gnugo-devel] Re: Regressions for 3.2 take 2x as long as for .34?

[bump]

> Dan,
> 
> can you think of any reason why all_batches of the regressions in 3.2 
> would run twice as long (6+ instead of 3+ hrs) as those in 3.1.33 and 
> .34?
> 
> Thanks,
> 
> s.

Are you sure?

There's very little changed since 3.1.34. In fact, it is
most accurate to say the engine is identical. It is almost
true that nothing has changed since 3.1.33. Here's the
Changelog:

3.2 released April 25, 2002

- revision of copyright statements and THANKS
- handtalk games moved to their own directory
- correct move number reported in influence for illegal move error

3.1.34 released April 20, 2002

- "choose best anchor"-platform dependency patch in mkpat.c
- bogus (int) conversion removed in get_next_move_from_list
- 3.0 replaced by 3.2 in WINDOWS

I'm cc'ing my reply to the list.

Dan