Re: [gnugo-devel] Owl persistent cache

[Daniel Bump]

Inge wrote:

> After computing the result, it recalculated it again using the
> contents of the hash table.  Thus many blind alleys were never
> searched, and the active area turned out much smaller, sometimes only
> half as big as before.

This seems drastic since then the code is run twice.

There is some question as to whether the code is running
correctly. If you run gnugo -d0x200000 at move 60 on the
jagm game, you get this active area for the C6 dragon:

 Stored result in cache (entry 13):
movenum         = 59
tactical_nodes  = 276101
routine         = OWL_DEFEND
(apos)               = C6
(bpos)                    = PASS
(cpos)                         = PASS
result          = 3
(move)          = D7
(move2)         = PASS
   A B C D E F G H J K L M N O P Q R S T
19 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 19
18 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 18
17 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 17
16 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 16
15 ? ? ? . ? ? ? ? ? ? . ? ? ? ? ? ? ? ? 15
14 ? ? . x . ? ? ? . . x . ? ? ? X X ? ? 14
13 ? ? . . . o . . x . . . . . . . . . . 13
12 ? . x . . . . O O . o . o . . . o . . 12
11 . . . . . . . . . . . . . . . . . . . 11
10 . . . . . . . x . x . . . . . . . . . 10
 9 . . o . . . . . . . . . . . . . o . . 9
 8 . . . . . x . . . . . . x . . . . . . 8
 7 . . . . x . . . . . . . . . . . . . . 7
 6 . .[O]. o x . . . . . . . . . . o . . 6
 5 . . O . o . . . . . x . . . . . . x . 5
 4 . o . x . x . . . . . . . x . x . . . 4
 3 . X X . . . . . . . x . . . . . . . . 3
 2 . . . . . . . . . . . . . . . . . . . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

First, the active area should exclude the right half of
the board.

Second, it was around this time that Gunnar (watching the
traces with DEBUG_OWL_PERFORMANCE on, I believe) said 
that the C6 group was being recomputed every move.

Since I didn't replay the game I can't swear that the
owl status was recomputed on this particular move but
the last move played, at M17 was outside this active area.

Dan