Move all in evaluation.
Simplify the code and concentrate in a single place
all the logic behind space evaluation, making it much
more clear.
Verified also at STC it does not regress due to a possible
slow down:
LLR: 3.91 (-2.94,2.94) [-3.00,1.00]
Total: 65744 W: 13285 L: 13194 D: 39265
No functional change.
Commenst are obsolete now, an updated description
would be quite obscure, so better let the code
to talk and remove them all together.
No functional change.
Use the same template of other pawns moves generation,
make the code more uniform, simplify generate_promotions
that has now been renamed.
No functional change (verified also with perft).
In some UNIX systems "rm" prompts user for confirmation.
However "rm -f" is always a guaranteed forced deletion.
Also move gcc profiling hack under the correct target
No Functional change
Resolves#168
ShelterWeakness and Stormdanger array are now indexed additionally by
file pair (a/h,b/g,c/f,d/e). The special case of king blocking a pawn
is incorporated in the StormDanger array. Finally the 93 parameters
are tuned by SPSA on LTC.
STC
ELO: 3.46 +-2.2 (95%) LOS: 99.9%
Total: 40000 W: 8275 L: 7877 D: 23848
LTC
LLR: 2.96 (-2.94,2.94) [0.00,6.00]
Total: 10311 W: 1876 L: 1721 D: 6714
Bench: 9498821
Resolves#163
The evaluation is already done by the specialized
function, don't need to add something elese later.
With this patch following positions are evaluated
correctly as draws:
8/6p1/1Pkp1p1p/2nNn2P/2P1K1P1/8/8/3B4 w - - 7
8/1k4p1/1P1p1p1p/3NnK1P/2P3P1/1n6/4B3/8 w - -
Verified it not regress with an STC test:
LLR: 3.15 (-2.94,2.94) [-3.00,1.00]
Total: 49812 W: 10095 L: 10016 D: 29701
Reported by Arjun Temurnikar.
bench: 8289983
In particular seems more natural to return
bool and TTEntry on the same line, actually
we should pass and return them as a pair,
but due to limitations of C++ and not wanting
to use std::pair this can be an acceptable
compromise.
No functional change.
Resolves#157
In some cases we want to go direcly to the moves loop
without checking for early return. The patch make this
logic more clear and consistent.
Tested for no regression, passed STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 25282 W: 5136 L: 5022 D: 15124
and LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 72007 W: 12133 L: 12095 D: 47779
bench: 9316798
This patch replaces RKISS by a simpler and faster PRNG, xorshift64* proposed
by S. Vigna (2014). It is extremely simple, has a large enough period for
Stockfish's needs (2^64), requires no warming-up (allowing such code to be
removed), and offers slightly better randomness than MT19937.
Paper: http://xorshift.di.unimi.it/
Reference source code (public domain):
http://xorshift.di.unimi.it/xorshift64star.c
The patch also simplifies how init_magics() searches for magics:
- Old logic: seed the PRNG always with the same seed,
then use optimized bit rotations to tailor the RNG sequence per rank.
- New logic: seed the PRNG with an optimized seed per rank.
This has two advantages:
1. Less code and less computation to perform during magics search (not ROTL).
2. More choices for random sequence tuning. The old logic only let us choose
from 4096 bit rotation pairs. With the new one, we can look for the best seeds
among 2^64 values. Indeed, the set of seeds[][] provided in the patch reduces
the effort needed to find the magics:
64-bit SF:
Old logic -> 5,783,789 rand64() calls needed to find the magics
New logic -> 4,420,086 calls
32-bit SF:
Old logic -> 2,175,518 calls
New logic -> 1,895,955 calls
In the 64-bit case, init_magics() take 25 ms less to complete (Intel Core i5).
Finally, when playing with strength handicap, non-determinism is achieved
by setting the seed of the static RNG only once. Afterwards, there is no need
to skip output values.
The bench only changes because the Zobrist keys are now different (since they
are random numbers straight out of the PRNG).
The RNG seed has been carefully chosen so that the
resulting Zobrist keys are particularly well-behaved:
1. All triplets of XORed keys are unique, implying that it
would take at least 7 keys to find a 64-bit collision
(test suggested by ceebo)
2. All pairs of XORed keys are unique modulo 2^32
3. The cardinality of { (key1 ^ key2) >> 48 } is as close
as possible to the maximum (65536)
Point 2 aims at ensuring a good distribution among the bits
that determine an TT entry's cluster, likewise point 3
among the bits that form the TT entry's key16 inside a
cluster.
Details:
Bitset card(key1^key2)
------ ---------------
RKISS
key16 64894 = 99.020% of theoretical maximum
low18 180117 = 99.293%
low32 305362 = 99.997%
Xorshift64*, old seed
key16 64918 = 99.057%
low18 179994 = 99.225%
low32 305350 = 99.993%
Xorshift64*, new seed
key16 65027 = 99.223%
low18 181118 = 99.845%
low32 305371 = 100.000%
Bench: 9324905
Resolves#148
Currently Search::RootMoves is accessed and even
modified by TB probing functions in a hidden
and sneaky way.
This is bad practice and makes the code tricky.
Instead explicily pass the vector as function
argument so to clarify that the vector is modified
inside the functions.
No functional change.
Simplified also some logic while there.
TBLargest needs renaming too, but itis for
a future patch because touches also syzygy
directory stuff.
No functional change.
We really don't need to uglify in this way
our nice count() API with this ad-hoc hack.
So remove the hack and use the already
existing infrastructure.
No functional change.
Resolves#134
Updating the makefile so that the clean and gcc-profile-clean targets also
remove the profiling data files in the syzygy directory.
No functional change.
Resolves#136
Just like in Physics, the ratio of 2 things in the same unit, should be
without unit.
Example use case:
- Ratio of a Depth by a Depth (eg. ONE_PLY) should be an int.
- Ratio of a Value by a Value (eg. PawnValueEg) should be an int.
Remove a bunch of useless const while there.
No functional change.
Resolves#128
Adds support for Syzygy tablebases to Stockfish. See
the Readme for information on using the tablebases.
Tablebase support can be enabled/disabled at the Makefile
level as well, by setting syzygy=yes or syzygy=no.
Big/little endian are both supported.
No functional change (if Tablebases are not used).
Resolves#6
It is possible that we won't have a ponder move if our PV
is too short. In that case, just don't print a ponder move.
No functional change
Resolves#130
When evaluating double pawns we use always
lsb() to extract the frontmost square.
This breaks evaluation color symmetry as is
possible to verify with an instrumented evaluate()
Value evaluate(const Position& pos) {
Value v = do_evaluate<false>(pos);
Position p = pos;
p.flip();
assert(v == do_evaluate<false>(p));
return v;
}
Passed no regression test:
STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 21035 W: 4244 L: 4122 D: 12669
LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 39839 W: 6662 L: 6572 D: 26605
bench: 8255966
It is a non functional change, but because
we now skip copying of pv[] in SpNode, patch
has been tested for regression with 3 threads:
STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 54668 W: 9873 L: 9809 D: 34986
No functional change.
This is a regression from 428962a
We have to cast to char here, otherwise the compiler
interprets it as an integer, and writes a number.
No functional change
Resolves#122
This is the first of a patch series to
rearrange and simplify accurate PV.
In this patch there is simple coding
style and reformatting stuff.
Verified with fishtest it does not crash
with MAX_PLY = 8
No functional change.
Previously token would keep its value from the previous line when an empty
line was input, leading to unexpected behaviour.
No functional change
Resolves#119
