Without patch we have 333198 nps, with patch 334249.
A very small +0.3%, not a lot manily becuase this is a
side path that is taken very few times.
Anyhow idea is correct becuase first 'quick' condition
has an hit rate of about 95%.
No functional change.
But still keep the same original
margin for score.
Passed both short TC test
LR: 2.95 (-2.94,2.94)
Total: 3710 W: 845 L: 726 D: 2139
And long TC
LLR: 2.95 (-2.94,2.94)
Total: 57859 W: 10939 L: 10532 D: 36388
bench: 4769737
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Partially revert previous patch and use
unlikey() just as code annotation.
Actually it is better to rely on a profiler for branch prediction:
http://blog.man7.org/2012/10/how-much-do-builtinexpect-likely-and.html
"In fact, even when only one in ten thousand values is nonzero,
we're still at only roughly the break-even point"
No functional change,
It is somewhat redundant and could make SF
name too long, so use just Version, in case
of a signature build Version will be set to
'sig-xxx' otherwise, if left empty, we fall
back on usual date stamp.
No functional change.
When compiling with:
make signature-build ARCH=xxx COMP=xxx
After binary has been roduced, it will be run to
get the signature 'stockfish bench' and this
number will be used as Version, so that it
will be easy to track the original sources
from a binary.
No functinal change.
Due to a strange issue (bug?) the ternary
operator does not return a BitCountType for
icc, so revert to the expression used
before bcbc9bfd1f
No functional change.
Here speed up is the name of the game.
Speed up is gained:
- Removing the useless enoughMaterial code
- Limiting trapped rook evaluation to where it counts
Tested at long TC:
LLR: 2.97 (-2.94,2.94)
Total: 10061 W: 1948 L: 1790 D: 6323
bench: 4558173
Thanks to Don, Miguel, Louis and the other people
of talkchess forum for the suggestion:
http://www.talkchess.com/forum/viewtopic.php?t=48612
Also sync polyglot.ini with current UCI options
No functional change.
Unfortunatly a reverse test at long TC failed:
master^ vs master
LLR: 1.37 (-2.94,2.94)
Total: 33682 W: 6294 L: 6071 D: 21317
So becuase short TC score is 50% there is a good
possibility patch is not scalable.
So revert it.
bench: 4507288
Do not use threat move to detect the condition. This
let us to retire the big allows() function.
Test at short TC was within 50% score:
LLR: -2.95 (-2.94,2.94)
Total: 38272 W: 7941 L: 7940 D: 22391
To be verified with reverse long TC
bench: 4191565
Here the main difference is that now we center
aspiration window on last returned score. This allows
to simplify handling of mate scores.
We have done a reversed SPRT tests, where we wanted to
verify if master is stronger than this patch.
Long TC: master vs this patch (reverse test)
LLR: -2.95 (-2.94,2.94)
Total: 37992 W: 7012 L: 6920 D: 24060
bench: 4507288
Temporary revert aspiration window patch
so to be visible to everybody: it will be
re-applied with next patch
No functional change (together with next one)
Here the main difference is that now we center
aspiration window on last returned score. This allows
to simplify handling of mate scores.
We have done a reversed SPRT tests, where we wanted to
verify if master is stronger than this patch.
Long TC: master vs this patch (reverse test)
LLR: -2.95 (-2.94,2.94)
Total: 37992 W: 7012 L: 6920 D: 24060
bench: 4507288
A simplification of the 'dangerous' definition.
Seems neutral at reverse test at long TC
master vs patch
LLR: -2.96 (-2.94,2.94)
Total: 16974 W: 3122 L: 3139 D: 10713
bench: 4689029
Function calloc() already initializes memory to
zero, so avoid calling clear() afterwards.
Also some renaming while there (inspired by DiscoCheck).
No functional change.
