[gnugo-devel] evan_3_13.10

[Arend Bayer]

Here is the summary from running evan_3_13.10 with Gunnar's script:

reading:173     PASS 1 A2 [1 (A2|C7|B1)]
ld_owl:188      FAIL 0 [2 Q19]
ld_owl:302      FAIL alive [critical T19 (Q19|R18)]
optics:1201     PASS 1 2 T8 T8 [1 2 T8 T8]
connection:26   FAIL 0 [1 B4]
connection:27   FAIL 0 [1 B4]
trevora:390     FAIL J4 [A2]
arb:150         PASS J5 [J5]
lazarus:6       PASS H3 [H3]
trevorb:160     PASS 1 N8 [1 ..*]
trevorc:820     FAIL B12 [A6]
13x13:53        FAIL G10 [C6]
13x13:65        FAIL L2 [H4]
nngs2:1         PASS A2 [B1|A2]
nngs3:1010      FAIL N16 [H16|H17]
Total nodes: 1289236074 2123284 5430615

(This is with regression directory as of 3-3-13-pre-1.)
With my fixed point integer patch, which should almost not touch node
number, I had:
Total nodes: 1253184862 2130309 5431181

So this is a 2.8% increase in reading nodes.

Anyway, since the attack and defense parts of Evan's patch are
completely independant, we'll split it up, which should make evaluation
easier.

Arend