stockfish
7 months agoMerge remote-tracking branch 'upstream/master' into HEAD
Steinar H. Gunderson [Mon, 28 Dec 2020 18:20:52 +0000 (19:20 +0100)]
Merge remote-tracking branch 'upstream/master' into HEAD

7 months agoDo more LMR for captures
Vizvezdenec [Thu, 24 Dec 2020 22:11:09 +0000 (01:11 +0300)]
Do more LMR for captures

This patch enables LMR for all captures at allNodes that were not in PV.
Currently we do LMR for all captures at cutNodes so this is an expansion of this logic:
now we do LMR for all captures almost at all non-pv nodes,
excluding only allNodes that were in PV.

passed STC
https://tests.stockfishchess.org/tests/view/5fe50b9d3932f79192d3973c
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 83128 W: 7606 L: 7368 D: 68154
Ptnml(0-2): 292, 5905, 28939, 6129, 299

passed LTC
https://tests.stockfishchess.org/tests/view/5fe552e43932f79192d39744
LLR: 2.92 (-2.94,2.94) {0.25,1.25}
Total: 13968 W: 568 L: 466 D: 12934
Ptnml(0-2): 5, 418, 6043, 506, 12

closes https://github.com/official-stockfish/Stockfish/pull/3273

Bench: 4194835

7 months agoCorrect Outflanking calculations in classical eval
Moez Jellouli [Sun, 20 Dec 2020 21:28:23 +0000 (22:28 +0100)]
Correct Outflanking calculations in classical eval

Take signed value of rank difference between kings squares instead absolute value in outflanking calculation. This change correct evaluation of endgames with one king invading opponent last ranks.

Passed STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 122240 W: 24326 L: 23896 D: 74018
Ptnml(0-2): 2101, 14139, 28236, 14517, 2127
https://tests.stockfishchess.org/tests/view/5fdfc33a3932f79192d394b8

Passed LTC:
LLR: 2.97 (-2.94,2.94) {0.25,1.25}
Total: 157416 W: 20870 L: 20292 D: 116254
Ptnml(0-2): 973, 13954, 48333, 14418, 1030
https://tests.stockfishchess.org/tests/view/5fe07a453932f79192d39502

closes https://github.com/official-stockfish/Stockfish/pull/3271

Bench: 4162769

7 months agoTweak the formulas for unsafeSquares
FauziAkram [Sun, 20 Dec 2020 15:50:34 +0000 (17:50 +0200)]
Tweak the formulas for unsafeSquares

We give more bonus for a special case: If there are some enemy squares occupied
or attacked by the enemy on the passed pawn span,
but if they are all attacked by our pawn, use new intermediate factor 30.

The main credit goes to Rocky for the idea, with additional tuning and tests.

Passed STC:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 96464 W: 19233 L: 18834 D: 58397
Ptnml(0-2): 1683, 11327, 21950, 11452, 1820
https://tests.stockfishchess.org/tests/view/5fdd21ab3932f79192d39357

Passed LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 81320 W: 10784 L: 10352 D: 60184
Ptnml(0-2): 602, 7524, 24044, 7820, 670
https://tests.stockfishchess.org/tests/view/5fddec983932f79192d393a4

closes https://github.com/official-stockfish/Stockfish/pull/3268

Bench: 4338972

7 months agoMerge remote-tracking branch 'upstream/master' into HEAD
Steinar H. Gunderson [Sat, 19 Dec 2020 15:05:08 +0000 (16:05 +0100)]
Merge remote-tracking branch 'upstream/master' into HEAD

7 months agoSimplify condition for assigning static-eval based bonus
pb00067 [Mon, 14 Dec 2020 15:30:56 +0000 (16:30 +0100)]
Simplify condition for assigning static-eval based bonus

for quiet move ordering and simplify bonus formula.

Due to clamping the bonus to relative low values the impact on high
depths is minimal, thus the restriction to low depths seems not
necessary.
Also the condition of movecount in previous node seems to be not
determinant.

Passed STC:
LLR: 2.95 (-2.94,2.94) {-1.25,0.25}
Total: 14600 W: 1424 L: 1323 D: 11853
Ptnml(0-2): 55, 1033, 5020, 1140, 52
https://tests.stockfishchess.org/tests/view/5fd67b381ac16912018885ec

Passed LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 85008 W: 3218 L: 3206 D: 78584
Ptnml(0-2): 49, 2840, 36700, 2880, 35
https://tests.stockfishchess.org/tests/view/5fd6af041ac16912018885f8

closes https://github.com/official-stockfish/Stockfish/pull/3265

bench: 4524994

7 months agoAdjust definition of unsafeSquares
FauziAkram [Wed, 16 Dec 2020 13:35:39 +0000 (15:35 +0200)]
Adjust definition of unsafeSquares

and adjust related bonus values. The bonus is now not given whenever
there is an enemy piece in front of the pawn.

Passed STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 109472 W: 22097 L: 21673 D: 65702
Ptnml(0-2): 2111, 12800, 24482, 13240, 2103
https://tests.stockfishchess.org/tests/view/5fd8d3740c5870924361ffad

Passed LTC:
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 39384 W: 5334 L: 4990 D: 29060
Ptnml(0-2): 279, 3648, 11535, 3910, 320
https://tests.stockfishchess.org/tests/view/5fd971ab0c5870924361fff0

closes https://github.com/official-stockfish/Stockfish/pull/3266

Bench: 4488955

7 months agoMerge remote-tracking branch 'upstream/master' into HEAD
Steinar H. Gunderson [Mon, 14 Dec 2020 23:33:25 +0000 (00:33 +0100)]
Merge remote-tracking branch 'upstream/master' into HEAD

7 months agoIncrease reduction in case of stable best move
Vizvezdenec [Mon, 14 Dec 2020 00:49:04 +0000 (03:49 +0300)]
Increase reduction in case of stable best move

The idea of this patch is pretty simple - we already do more reductions
for non-PV and root nodes in case of stable best move for depth > 10.
This patch makes us do so if root depth if > 10 instead, which
is logical since best move changes (thus instability of it) is
counted at root, so it makes a lot of sense to use depth of the root.

passed STC
https://tests.stockfishchess.org/tests/view/5fd643271ac16912018885c5
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 13232 W: 1308 L: 1169 D: 10755
Ptnml(0-2): 39, 935, 4535, 1062, 45

passed LTC
https://tests.stockfishchess.org/tests/view/5fd68db11ac16912018885f0
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 14024 W: 565 L: 463 D: 12996
Ptnml(0-2): 3, 423, 6062, 517, 7

closes https://github.com/official-stockfish/Stockfish/pull/3263

Bench: 4050630

7 months agoMerge static history into main history,
pb00067 [Sun, 13 Dec 2020 20:23:30 +0000 (21:23 +0100)]
Merge static history into main history,

thus simplifying and reducing the memory footprint.
I believe using static diff for better move ordering is more suited for
low depths, so restrict writing to low depths.

Todo: probably the condition for writing can be simplified

LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 18752 W: 768 L: 705 D: 17279
Ptnml(0-2): 7, 635, 8034, 688, 12
https://tests.stockfishchess.org/tests/view/5fd631791ac169120188859e

STC:
LLR: 2.95 (-2.94,2.94) {-1.25,0.25}
Total: 36504 W: 3380 L: 3313 D: 29811
Ptnml(0-2): 116, 2667, 12645, 2682, 142
https://tests.stockfishchess.org/tests/view/5fd5ed861ac1691201888569

closes https://github.com/official-stockfish/Stockfish/pull/3262

bench: 4018036

7 months agoAVX512, AVX2 and SSSE3 speedups
mstembera [Sat, 12 Dec 2020 22:18:38 +0000 (14:18 -0800)]
AVX512, AVX2 and SSSE3 speedups

Improves throughput by summing 2 intermediate dot products using 16 bit addition before upconverting to 32 bit.

Potential saturation is detected and the code-path is avoided in this case.
The saturation can't happen with the current nets,
but nets can be constructed that trigger this check.

STC https://tests.stockfishchess.org/tests/view/5fd40a861ac1691201888479
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 25544 W: 2451 L: 2296 D: 20797
Ptnml(0-2): 92, 1761, 8925, 1888, 106

about 5% speedup

closes https://github.com/official-stockfish/Stockfish/pull/3261

No functional change

7 months agoNew Imbalance Tables Tweak
FauziAkram [Mon, 7 Dec 2020 17:28:47 +0000 (19:28 +0200)]
New Imbalance Tables Tweak

Imbalance tables tweaked to contain MiddleGame and Endgame values, instead of a single value.

The idea started from Fisherman, which requested my help to tune the values back in June/July,
so I tuned the values back then, and we were able to accomplish good results,
but not enough to pass both STC and LTC tests.

So after the recent changes, I decided to give it another shot, and I am glad that it was a successful attempt.

A special thanks goes also to mstembera, which notified me a simple way to let the patch perform a little better.

Passed STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 115976 W: 23124 L: 22695 D: 70157
Ptnml(0-2): 2074, 13652, 26285, 13725, 2252
https://tests.stockfishchess.org/tests/view/5fc92d2d42a050a89f02ccc8

Passed LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 156304 W: 20617 L: 20024 D: 115663
Ptnml(0-2): 1138, 14647, 46084, 15050, 1233
https://tests.stockfishchess.org/tests/view/5fc9fee142a050a89f02cd3e

closes https://github.com/official-stockfish/Stockfish/pull/3255

Bench: 4278746

7 months agoUse arithmetic right shift for sign extension in MMX and SSE2 paths
Fanael Linithien [Mon, 7 Dec 2020 13:46:29 +0000 (14:46 +0100)]
Use arithmetic right shift for sign extension in MMX and SSE2 paths

This appears to be slightly faster than using a comparison against zero
to compute the high bits, on both old (like Pentium III) and new (like
Zen 2) hardware.

closes https://github.com/official-stockfish/Stockfish/pull/3254

No functional change.

7 months agoAdd comments to uncommented parts of code
Vizvezdenec [Tue, 1 Dec 2020 12:02:35 +0000 (15:02 +0300)]
Add comments to uncommented parts of code

https://github.com/official-stockfish/Stockfish/pull/3250

No functional change

7 months agoIntroduce static history
Vizvezdenec [Sat, 5 Dec 2020 01:00:41 +0000 (04:00 +0300)]
Introduce static history

The idea of this patch can be described as following: we update static
history stats based on comparison of the static evaluations of the
position before and after the move. If the move increases static evaluation
it's assigned positive bonus, if it decreases static evaluation
it's assigned negative bonus. These stats are used in movepicker
to sort quiet moves.

passed STC
https://tests.stockfishchess.org/tests/view/5fca4c0842a050a89f02cd66
LLR: 3.00 (-2.94,2.94) {-0.25,1.25}
Total: 78152 W: 7409 L: 7171 D: 63572
Ptnml(0-2): 303, 5695, 26873, 5871, 334

passed LTC
https://tests.stockfishchess.org/tests/view/5fca6be442a050a89f02cd75
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 40240 W: 1602 L: 1441 D: 37197
Ptnml(0-2): 19, 1306, 17305, 1475, 15

closes https://github.com/official-stockfish/Stockfish/pull/3253

bench 3845156

7 months agoMerge remote-tracking branch 'upstream/master' into HEAD
Steinar H. Gunderson [Tue, 1 Dec 2020 22:03:04 +0000 (23:03 +0100)]
Merge remote-tracking branch 'upstream/master' into HEAD

8 months agoUpdate default net to nn-62ef826d1a6d.nnue
SFisGOD [Sat, 28 Nov 2020 15:14:34 +0000 (23:14 +0800)]
Update default net to nn-62ef826d1a6d.nnue

Include scaling change as suggested by Dietrich Kappe,
the one who trained net for Komodo.  According to him,
some nets may require different scaling in order to utilize its full strength.

STC:
LLR: 2.93 (-2.94,2.94) {-0.25,1.25}
Total: 99856 W: 9669 L: 9401 D: 80786
Ptnml(0-2): 374, 7468, 34037, 7614, 435
https://tests.stockfishchess.org/tests/view/5fc2697642a050a89f02c8ec

LTC:
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 29840 W: 1220 L: 1081 D: 27539
Ptnml(0-2): 10, 969, 12827, 1100, 14
https://tests.stockfishchess.org/tests/view/5fc2ea5142a050a89f02c957

Bench: 3561701

8 months agoReductions simplification
Unai Corzo [Sun, 29 Nov 2020 12:52:36 +0000 (13:52 +0100)]
Reductions simplification

Simplify increase reduction for captures/promotions if late move and at low depth.

STC https://tests.stockfishchess.org/tests/view/5fbff65067cbf42301d6b3ae
LLR: 2.97 (-2.94,2.94) {-1.25,0.25}
Total: 49088 W: 4607 L: 4555 D: 39926
Ptnml(0-2): 177, 3615, 16932, 3619, 201

LTC https://tests.stockfishchess.org/tests/view/5fc0902967cbf42301d6b3fc
LLR: 2.99 (-2.94,2.94) {-0.75,0.25}
Total: 160944 W: 6153 L: 6193 D: 148598
Ptnml(0-2): 90, 5525, 69294, 5461, 102

closes https://github.com/official-stockfish/Stockfish/pull/3248

bench: 3834568

8 months agoRemove piece lists
syzygy1 [Sun, 29 Nov 2020 11:05:26 +0000 (12:05 +0100)]
Remove piece lists

This patch removes the incrementally updated piece lists from the Position object.

This has been tried before but always failed. My reasons for trying again are:

* 32-bit systems (including phones) are now much less important than they were some years ago (and are absent from fishtest);
* NNUE may have made SF less finely tuned to the order in which moves were generated.

STC:
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 55272 W: 5260 L: 5216 D: 44796
Ptnml(0-2): 208, 4147, 18898, 4159, 224
https://tests.stockfishchess.org/tests/view/5fc2986a42a050a89f02c926

LTC:
LLR: 2.96 (-2.94,2.94) {-0.75,0.25}
Total: 16600 W: 673 L: 608 D: 15319
Ptnml(0-2): 14, 533, 7138, 604, 11
https://tests.stockfishchess.org/tests/view/5fc2f98342a050a89f02c95c

closes https://github.com/official-stockfish/Stockfish/pull/3247

Bench: 3940967

8 months agoUpdate README.md
Unai Corzo [Sun, 29 Nov 2020 08:07:31 +0000 (09:07 +0100)]
Update README.md

fix a few typos

closes https://github.com/official-stockfish/Stockfish/pull/3245

No functional change

8 months agoSearch simplification
Unai Corzo [Sat, 28 Nov 2020 15:32:52 +0000 (16:32 +0100)]
Search simplification

STC https://tests.stockfishchess.org/tests/view/5fc2083942a050a89f02c8bb
LLR: 2.93 (-2.94,2.94) {-1.25,0.25}
Total: 23200 W: 2251 L: 2160 D: 18789
Ptnml(0-2): 86, 1726, 7895, 1797, 96

LTC https://tests.stockfishchess.org/tests/view/5fc22d7b42a050a89f02c8d0
LLR: 2.92 (-2.94,2.94) {-0.75,0.25}
Total: 15832 W: 653 L: 590 D: 14589
Ptnml(0-2): 7, 521, 6795, 588, 5

closes https://github.com/official-stockfish/Stockfish/pull/3244/files

bench: 3827317

Simplify search.

8 months agoRemove bonus for killers.
lonfom169 [Thu, 26 Nov 2020 12:40:54 +0000 (09:40 -0300)]
Remove bonus for killers.

Passed non-regression STC:
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 14712 W: 1416 L: 1315 D: 11981
Ptnml(0-2): 59, 1029, 5082, 1124, 62
https://tests.stockfishchess.org/tests/view/5fbfa31f67cbf42301d6b36e

Passed non-regression LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 27536 W: 1099 L: 1044 D: 25393
Ptnml(0-2): 11, 929, 11838, 974, 16
https://tests.stockfishchess.org/tests/view/5fbfac9167cbf42301d6b371

closes https://github.com/official-stockfish/Stockfish/pull/3241

Bench: 3887644

8 months agoRefine rook penalty on closed files
Lolligerhans [Fri, 20 Nov 2020 17:09:41 +0000 (18:09 +0100)]
Refine rook penalty on closed files

+-----------------+
| . . . . . . . . | All files are closed. Some files are
| . . . . . o o . | more valuable for rooks, because
| . . . . o . . o | they might open in the future.
| . . . o x . . x |
| o . o x . x x . |
| x o x . . . . . | x  our pawns
| . x . . . . . . | o  their pawns
| . . . . . . . . | ^  rooks are scored higher on these files
+-----------------+
            ^ ^

Files containing none of our own pawns are open or half-open (otherwise
they are closed). Rooks on (half-)open files recieve a bonus for the
future potential to act along all ranks.

This commit refines the (relative) penalty of rooks on closed files.
Files that contain one of our blocked pawns are considered less likely
to open in the future; rooks on these files are now penalized stronger.

This bonus does not generally correlate with mobility. If the condition
is sufficiently refined in the future, it may be beneficial to adjust or
override mobility scores in some cases.

LTC
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 494384 W: 71565 L: 70231 D: 352588
Ptnml(0-2): 3907, 48050, 142118, 49036, 4081
https://tests.stockfishchess.org/tests/view/5fb9312e67cbf42301d6afb9

LTC (non-regression w/ book noob_3moves.epd)
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 208520 W: 27044 L: 26937 D: 154539
Ptnml(0-2): 1557, 19850, 61391, 19853, 1609
https://tests.stockfishchess.org/tests/view/5fc01ced67cbf42301d6b3df

STC
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 98392 W: 20269 L: 19868 D: 58255
Ptnml(0-2): 1804, 11297, 22589, 11706, 1800
https://tests.stockfishchess.org/tests/view/5fb7f88a67cbf42301d6af10

closes https://github.com/official-stockfish/Stockfish/pull/3242

Bench: 3682630

8 months agoCleaned up MakeIndex()
mstembera [Tue, 17 Nov 2020 23:58:25 +0000 (15:58 -0800)]
Cleaned up MakeIndex()

The index order in kpp_board_index[][] is reversed to be more optimal for the access pattern

STC https://tests.stockfishchess.org/tests/view/5fbd74f967cbf42301d6b24f
LLR: 2.93 (-2.94,2.94) {-1.25,0.25}
Total: 27504 W: 2686 L: 2607 D: 22211
Ptnml(0-2): 84, 2001, 9526, 2034, 107

closes https://github.com/official-stockfish/Stockfish/pull/3233

No functional change

8 months agouse classical for certain endgames.
Vizvezdenec [Thu, 26 Nov 2020 05:38:09 +0000 (06:38 +0100)]
use classical for certain endgames.

STC https://tests.stockfishchess.org/tests/view/5fbc64c067cbf42301d6b1d6
LLR: 2.97 (-2.94,2.94) {-0.25,1.25}
Total: 53360 W: 5223 L: 5024 D: 43113
Ptnml(0-2): 184, 3877, 18390, 4014, 215

LTC https://tests.stockfishchess.org/tests/view/5fbc97f267cbf42301d6b1ee
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 126472 W: 5111 L: 4766 D: 116595
Ptnml(0-2): 50, 4032, 54749, 4333, 72

closes https://github.com/official-stockfish/Stockfish/pull/3240

bench: 3820648

8 months agoCalculate sum from first elements
MaximMolchanov [Sat, 14 Nov 2020 00:55:29 +0000 (02:55 +0200)]
Calculate sum from first elements

in affine transform for AVX512/AVX2/SSSE3

The idea is to initialize sum with the first element instead of zero.
Reduce one add_epi32 and one set_zero SIMD instructions for each output dimension.

sum = 0; for i = 1 to n sum += a[i] ->
sum = a[1]; for i = 2 to n sum += a[i]

STC:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 69048 W: 7024 L: 6799 D: 55225
Ptnml(0-2): 260, 5175, 23458, 5342, 289
https://tests.stockfishchess.org/tests/view/5faf2cf467cbf42301d6aa06

closes https://github.com/official-stockfish/Stockfish/pull/3227

No functional change.

8 months agoAssorted search and eval parameter tune
Unai Corzo [Tue, 24 Nov 2020 16:06:30 +0000 (17:06 +0100)]
Assorted search and eval parameter tune

Search and eval parameter tune.

STC https://tests.stockfishchess.org/tests/view/5fba850a67cbf42301d6b07d
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 24312 W: 2388 L: 2228 D: 19696
Ptnml(0-2): 85, 1800, 8241, 1930, 100

LTC https://tests.stockfishchess.org/tests/view/5fbad5ea67cbf42301d6b0fa
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 88376 W: 3619 L: 3351 D: 81406
Ptnml(0-2): 56, 2977, 37849, 3255, 51

closes https://github.com/official-stockfish/Stockfish/pull/3232

bench: 3600361

8 months agoSmall cleanups 13
Stéphane Nicolet [Thu, 12 Nov 2020 13:05:28 +0000 (14:05 +0100)]
Small cleanups 13

No functional change

8 months agoRook Mobility Tweak
FauziAkram [Wed, 11 Nov 2020 18:56:29 +0000 (20:56 +0200)]
Rook Mobility Tweak

Passed STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 171152 W: 34715 L: 34202 D: 102235
Ptnml(0-2): 3278, 20155, 38228, 20606, 3309
https://tests.stockfishchess.org/tests/view/5fa861f467cbf42301d6a68e

Passed LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 149616 W: 20471 L: 19882 D: 109263
Ptnml(0-2): 1172, 14434, 43102, 14833, 1267
https://tests.stockfishchess.org/tests/view/5fa9c8ff67cbf42301d6a74f

closes https://github.com/official-stockfish/Stockfish/pull/3226

Bench: 3597730

8 months agoIncrease reduction at root
SFisGOD [Tue, 10 Nov 2020 17:28:43 +0000 (18:28 +0100)]
Increase reduction at root

when the best move does not change frequently

STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 51320 W: 5159 L: 4956 D: 41205
Ptnml(0-2): 215, 3897, 17242, 4082, 224
https://tests.stockfishchess.org/tests/view/5faa072367cbf42301d6a767

LTC:
LLR: 2.98 (-2.94,2.94) {0.25,1.25}
Total: 15952 W: 762 L: 642 D: 14548
Ptnml(0-2): 8, 561, 6725, 667, 15
https://tests.stockfishchess.org/tests/view/5faa4c3567cbf42301d6a794

closes https://github.com/official-stockfish/Stockfish/pull/3225

Bench: 3954692

8 months agoIncrease reduction based on the number of best move changes.
lonfom169 [Mon, 9 Nov 2020 02:43:32 +0000 (23:43 -0300)]
Increase reduction based on the number of best move changes.

Thanks to Vizvezdenec for the PvNode idea and also to vondele the !PvNode idea.

Passed STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 19120 W: 1998 L: 1839 D: 15283
Ptnml(0-2): 76, 1445, 6375, 1572, 92
https://tests.stockfishchess.org/tests/view/5fa8af3e67cbf42301d6a6c9

Passed LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 75584 W: 3454 L: 3205 D: 68925
Ptnml(0-2): 54, 2832, 31771, 3081, 54

closes https://github.com/official-stockfish/Stockfish/pull/3224

Bench: 3595418

8 months agoQsearch pruning: follow-up
Stéphane Nicolet [Fri, 6 Nov 2020 18:20:27 +0000 (19:20 +0100)]
Qsearch pruning: follow-up

This is a follow-up of the recent qsearch pruning patch in
https://github.com/official-stockfish/Stockfish/commit/a260c9a8a24a2630a900efc3821000c3481b0c5d

We now use the same guard condition (testing that we already have a defense with
a score better  score than a TB loss) for all pruning heuristics in qsearch().
This allows some pruning when in check, but  in a controlled way to ensure that
no wrong mate scores appear.

Tested with Elo-gaining bounds:

STC:
LLR: 2.97 (-2.94,2.94) {-0.25,1.25}
Total: 22632 W: 2433 L: 2264 D: 17935
Ptnml(0-2): 98, 1744, 7487, 1865, 122
https://tests.stockfishchess.org/tests/view/5fa59405936c54e11ec99515

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 105432 W: 4965 L: 4648 D: 95819
Ptnml(0-2): 85, 4110, 44011, 4423, 87
https://tests.stockfishchess.org/tests/view/5fa5b609936c54e11ec9952a

closes https://github.com/official-stockfish/Stockfish/pull/3221

Bench: 3578092

8 months agoUpdate default net to nn-c3ca321c51c9.nnue
SFisGOD [Sat, 7 Nov 2020 00:50:02 +0000 (08:50 +0800)]
Update default net to nn-c3ca321c51c9.nnue

Optimization of the net biases of the 32 x 32 layer and the output layer.

Tuning of 32 x 32 layer (200k games, 5 seconds TC)
https://tests.stockfishchess.org/tests/view/5f9aaf266a2c112b60691c68

STC:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 41848 W: 4665 L: 4461 D: 32722
Ptnml(0-2): 239, 3308, 13659, 3446, 272
https://tests.stockfishchess.org/tests/view/5fa5ef5a936c54e11ec9954f

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 88008 W: 4045 L: 3768 D: 80195
Ptnml(0-2): 69, 3339, 36908, 3622, 66
https://tests.stockfishchess.org/tests/view/5fa62a78936c54e11ec99577

closes https://github.com/official-stockfish/Stockfish/pull/3220

Bench: 3649288

8 months agoAVX-512 for smaller affine and feature transforms.
Tomasz Sobczyk [Tue, 3 Nov 2020 21:49:10 +0000 (22:49 +0100)]
AVX-512 for smaller affine and feature transforms.

For the feature transformer the code is analogical to AVX2 since there was room for easy adaptation of wider simd registers.

For the smaller affine transforms that have 32 byte stride we keep 2 columns in one zmm register. We also unroll more aggressively so that in the end we have to do 16 parallel horizontal additions on ymm slices each consisting of 4 32-bit integers. The slices are embedded in 8 zmm registers.

These changes provide about 1.5% speedup for AVX-512 builds.

Closes https://github.com/official-stockfish/Stockfish/pull/3218

No functional change.

8 months agoIntroducing King On File
FauziAkram [Wed, 4 Nov 2020 23:54:53 +0000 (01:54 +0200)]
Introducing King On File

this new concept calculates bonuses/penalties for the king when the king is in a semiopen or open file.

Passed STC:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 44904 W: 9365 L: 9028 D: 26511
Ptnml(0-2): 857, 5309, 9841, 5530, 915
https://tests.stockfishchess.org/tests/view/5fa343625d72639a7acef72b

Passed LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 60552 W: 8449 L: 8051 D: 44052
Ptnml(0-2): 466, 5772, 17481, 6012, 545
https://tests.stockfishchess.org/tests/view/5fa40e365d72639a7acef79e

closes https://github.com/official-stockfish/Stockfish/pull/3219

Bench: 3689484

8 months agoChange handling the special case of a single legal move.
Joost VandeVondele [Fri, 23 Oct 2020 05:39:35 +0000 (07:39 +0200)]
Change handling the special case of a single legal move.

Using no searching time in case of a single legal move is not beneficial from
a strength point of view, and this special case can be easily removed:

STC:
LLR: 2.93 (-2.94,2.94) {-1.25,0.25}
Total: 22472 W: 2458 L: 2357 D: 17657
Ptnml(0-2): 106, 1733, 7453, 1842, 102
https://tests.stockfishchess.org/tests/view/5f926cbc81eda81bd78cb6df

LTC:
LLR: 2.94 (-2.94,2.94) {-0.75,0.25}
Total: 37880 W: 1736 L: 1682 D: 34462
Ptnml(0-2): 22, 1392, 16057, 1448, 21
https://tests.stockfishchess.org/tests/view/5f92a26081eda81bd78cb6fe

The advantage of using the normal time management for a single legal move is that scores
reported for that move are reasonable, not searching leads to artifacts during games
(see e.g. https://tcec-chess.com/#div=sf&game=96&season=19)

The disadvantage of using normal time management of a single legal move is that thinking
times can be unnaturally long, making it 'painful to watch' in online tournaments.

This patch uses normal time management, but caps the used time to 500ms.
This should lead to reasonable scores, and be hardly perceptible.

closes https://github.com/official-stockfish/Stockfish/pull/3195
closes https://github.com/official-stockfish/Stockfish/pull/3183

variant of a patch suggested by SFisGOD

No functional change.

8 months agoManually align arrays on the stack
Tomasz Sobczyk [Tue, 3 Nov 2020 10:23:35 +0000 (11:23 +0100)]
Manually align arrays on the stack

as a workaround to issues with overaligned alignas() on stack variables in gcc < 9.3 on windows.

closes https://github.com/official-stockfish/Stockfish/pull/3217

fixes #3216

No functional change

8 months agoFix incorrect pruning in qsearch
J. Oster [Sun, 1 Nov 2020 17:33:17 +0000 (18:33 +0100)]
Fix incorrect pruning in qsearch

Only do countermove based pruning in qsearch if we already have a move with a better score than a TB loss.

This patch fixes a bug (started as 843a961) that incorrectly prunes moves if in check,
and adds an assert to make sure no wrong mate scores are given in the future.
It replaces a no-op moveCount check with a check for bestValue.

Initially discussed in #3171 and later in #3199, #3198 and #3210.
This PR effectively closes #3171
It also likely fixes #3196 where this causes user visible incorrect TB scores,
which probably result from these incorrect mate scores.

Passed STC and LTC non-regression tests.
https://tests.stockfishchess.org/tests/view/5f9ef8dabca9bf35bae7f648
LLR: 2.93 (-2.94,2.94) {-1.25,0.25}
Total: 21672 W: 2339 L: 2230 D: 17103
Ptnml(0-2): 126, 1689, 7083, 1826, 112

https://tests.stockfishchess.org/tests/view/5f9f0caebca9bf35bae7f666
LLR: 2.97 (-2.94,2.94) {-0.75,0.25}
Total: 33152 W: 1551 L: 1485 D: 30116
Ptnml(0-2): 27, 1308, 13832, 1390, 19

closes https://github.com/official-stockfish/Stockfish/pull/3214

Bench: 3625915

8 months agoElo Worth in King Danger
FauziAkram [Thu, 29 Oct 2020 14:33:18 +0000 (17:33 +0300)]
Elo Worth in King Danger

Adding the EloWorth for each term in King Danger.
Should be useful for simplifications, tuning patches, and new ideas.

closes https://github.com/official-stockfish/Stockfish/pull/3204

non-functional change

8 months agoOptimize affine transform for SSSE3 and higher targets.
Tomasz Sobczyk [Wed, 28 Oct 2020 23:14:53 +0000 (00:14 +0100)]
Optimize affine transform for SSSE3 and higher targets.

A non-functional speedup. Unroll the loops going over
the output dimensions in the affine transform layers by
a factor of 4 and perform 4 horizontal additions at a time.
Instead of doing naive horizontal additions on each vector
separately use hadd and shuffling between vectors to reduce
the number of instructions by using all lanes for all stages
of the horizontal adds.

passed STC of the initial version:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 17808 W: 1914 L: 1756 D: 14138
Ptnml(0-2): 76, 1330, 5948, 1460, 90
https://tests.stockfishchess.org/tests/view/5f9d516f6a2c112b60691da3

passed STC of the final version after cleanup:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 16296 W: 1750 L: 1595 D: 12951
Ptnml(0-2): 72, 1192, 5479, 1319, 86
https://tests.stockfishchess.org/tests/view/5f9df5776a2c112b60691de3

closes https://github.com/official-stockfish/Stockfish/pull/3203

No functional change

8 months agoUpdate default net to nn-cb26f10b1fd9.nnue
mstembera [Fri, 30 Oct 2020 20:45:40 +0000 (13:45 -0700)]
Update default net to nn-cb26f10b1fd9.nnue

Result of https://tests.stockfishchess.org/tests/view/5f9a06796a2c112b60691c0f tuning.

STC
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 53712 W: 5776 L: 5561 D: 42375
Ptnml(0-2): 253, 4282, 17604, 4431, 286
https://tests.stockfishchess.org/tests/view/5f9c7bbc6a2c112b60691d4d

LTC
LLR: 2.97 (-2.94,2.94) {0.25,1.25}
Total: 80184 W: 4007 L: 3739 D: 72438
Ptnml(0-2): 58, 3302, 33130, 3518, 84
https://tests.stockfishchess.org/tests/view/5f9d01f06a2c112b60691d87

closes https://github.com/official-stockfish/Stockfish/pull/3209

bench: 3517795

9 months agoDo not skip non-recapture ttMove when in check
syzygy1 [Tue, 27 Oct 2020 18:22:41 +0000 (19:22 +0100)]
Do not skip non-recapture ttMove when in check

The qsearch() MovePicker incorrectly skips a non-recapture ttMove
when in check (if depth <= DEPTH_QS_RECAPTURES). This is clearly not
intended and can cause qsearch() to return a mate score when there
is no mate. Introduced in cad300c and 6596f0e, as observed by
joergoster in #3171 and #3198.

This PR fixes the bug by not skipping the non-recapture ttMove when in check.

Passed non-regression STC:
https://tests.stockfishchess.org/tests/view/5f9867ea6a2c112b60691b10
LLR: 2.98 (-2.94,2.94) {-1.25,0.25}
Total: 27112 W: 2943 L: 2842 D: 21327
Ptnml(0-2): 127, 2170, 8878, 2237, 144

Passed non-regression LTC:
https://tests.stockfishchess.org/tests/view/5f9967326a2c112b60691bb0
LLR: 2.99 (-2.94,2.94) {-0.75,0.25}
Total: 18392 W: 807 L: 738 D: 16847
Ptnml(0-2): 9, 655, 7802, 718, 12

closes https://github.com/official-stockfish/Stockfish/pull/3199
closes https://github.com/official-stockfish/Stockfish/pull/3198

Bench: 3870606

9 months agoUpdate default net to nn-2eb2e0707c2b.nnue
SFisGOD [Tue, 27 Oct 2020 20:24:55 +0000 (04:24 +0800)]
Update default net to nn-2eb2e0707c2b.nnue

Optimization of the net weights of the 32 x 32 layer (1024 parameters) and net biases of the 512 x 32 layer (32 parameters) using SPSA.

Tuning of 32 x 32 Layer (800,000 games, 5 seconds time control):
https://tests.stockfishchess.org/tests/view/5f942040d3978d7e86f1aa05

Tuning of 512 x 32 Layer (80,000 games, 20 seconds time control):
https://tests.stockfishchess.org/tests/view/5f8f926d2c92c7fe3a8c608b

STC:
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 17336 W: 1918 L: 1754 D: 13664
Ptnml(0-2): 79, 1344, 5672, 1480, 93
https://tests.stockfishchess.org/tests/view/5f9882346a2c112b60691b34

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 37304 W: 1822 L: 1651 D: 33831
Ptnml(0-2): 27, 1461, 15501, 1640, 23
https://tests.stockfishchess.org/tests/view/5f98a4b36a2c112b60691b40

closes https://github.com/official-stockfish/Stockfish/pull/3201

Bench: 3403528

9 months agoBishop Pawns based on Files
FauziAkram [Fri, 23 Oct 2020 23:01:04 +0000 (02:01 +0300)]
Bishop Pawns based on Files

Passed STC:
https://tests.stockfishchess.org/tests/view/5f8cc8145a4eacb45305da3c
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 132544 W: 27795 L: 27328 D: 77421
Ptnml(0-2): 2756, 15558, 29272, 15835, 2851

Passed LTC:
https://tests.stockfishchess.org/tests/view/5f8df614bacb75a4f9a4721e
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 169608 W: 23257 L: 22622 D: 123729
Ptnml(0-2): 1408, 16316, 48758, 16877, 1445

closes https://github.com/official-stockfish/Stockfish/pull/3194

Bench: 4067106

9 months agoMore incremental accumulator updates
syzygy1 [Tue, 20 Oct 2020 19:06:06 +0000 (21:06 +0200)]
More incremental accumulator updates

This patch was inspired by c065abd which updates the accumulator,
if possible, based on the accumulator of two plies back if
the accumulator of the preceding ply is not available.

With this patch we look back even further in the position history
in an attempt to reduce the number of complete recomputations.
When we find a usable accumulator for the position N plies back,
we also update the accumulator of the position N-1 plies back
because that accumulator is most likely to be helpful later
when evaluating positions in sibling branches.
By not updating all intermediate accumulators immediately,
we avoid doing too much work that is not certain to be useful.
Overall, roughly 2-3% speedup.

This patch makes the code more specific to the net architecture,
changing input features of the net will require additional changes
to the incremental update code as discussed in the PR #3193 and #3191.

Passed STC:
https://tests.stockfishchess.org/tests/view/5f9056712c92c7fe3a8c60d0
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 10040 W: 1116 L: 968 D: 7956
Ptnml(0-2): 42, 722, 3365, 828, 63

closes https://github.com/official-stockfish/Stockfish/pull/3193

No functional change.

9 months agoAdd net as dependency of config
Joost VandeVondele [Sun, 18 Oct 2020 13:01:19 +0000 (15:01 +0200)]
Add net as dependency of config

cleaner output and error message if the server is down and the net is not available.

closes https://github.com/official-stockfish/Stockfish/pull/3188

No functional change

9 months agoReduce big time spikes by reducing PV re-searches.
xoto10 [Wed, 21 Oct 2020 13:52:13 +0000 (14:52 +0100)]
Reduce big time spikes by reducing PV re-searches.

Save time by reducing PV re-searches above original depth. Instead use 5% extra time on every move.

STC 10+0.1 th 1 :
LLR: 2.93 (-2.94,2.94) {-0.25,1.25}
Total: 90688 W: 9702 L: 9436 D: 71550
Ptnml(0-2): 408, 7252, 29792, 7450, 442
https://tests.stockfishchess.org/tests/view/5f8df807bacb75a4f9a47223

LTC 60+0.6 th 1 :
LLR: 2.97 (-2.94,2.94) {0.25,1.25}
Total: 97856 W: 4602 L: 4303 D: 88951
Ptnml(0-2): 53, 3757, 41057, 3960, 101
https://tests.stockfishchess.org/tests/view/5f8ec4872c92c7fe3a8c602d

closes https://github.com/official-stockfish/Stockfish/pull/3192

Bench 3943959

9 months agoDo more reductions for late quiet moves in case of consecutive fail highs.
Vizvezdenec [Sat, 17 Oct 2020 11:40:10 +0000 (13:40 +0200)]
Do more reductions for late quiet moves in case of consecutive fail highs.

Idea of this patch can be described as following - in case we have consecutive fail highs and we reach late enough moves at root node probability of remaining quiet moves being able to produce even bigger value than moves that produced previous cutoff (so ones that should be high in move ordering but now they fail to produce beta cutoff because we actually reached high move count) should be quiet small so we can reduce them more.

passed STC
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 53392 W: 5681 L: 5474 D: 42237
Ptnml(0-2): 214, 4104, 17894, 4229, 255
https://tests.stockfishchess.org/tests/view/5f88501adcdad978fe8c527e

passed LTC
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 59136 W: 2773 L: 2564 D: 53799
Ptnml(0-2): 30, 2117, 25078, 2300, 43
https://tests.stockfishchess.org/tests/view/5f884dbfdcdad978fe8c527a

closes https://github.com/official-stockfish/Stockfish/pull/3184

Bench: 4066972

9 months agoUpdate default net to nn-eba324f53044.nnue
mstembera [Sun, 18 Oct 2020 11:23:28 +0000 (04:23 -0700)]
Update default net to nn-eba324f53044.nnue

The new net is based on the previous net 04cf2b4ed1da but with the biases
for the 1st hidden layer tuned SPSA, see the SPSA session on fishtest there:
https://tests.stockfishchess.org/tests/view/5f875213dcdad978fe8c5211

Thanks to @vondele for writing out the net, see discussion in this thread:
https://github.com/mstembera/Stockfish/commit/432da86721647dff1d9426a7cdcfd2dbada8155e

Passed STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 15000 W: 1640 L: 1483 D: 11877
Ptnml(0-2): 50, 1183, 4908, 1278, 81
https://tests.stockfishchess.org/tests/view/5f8955e20fea1a44ec4f0a5d

Passed LTC:
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 81272 W: 3948 L: 3682 D: 73642
Ptnml(0-2): 64, 3194, 33856, 3456, 66
https://tests.stockfishchess.org/tests/view/5f89e8efeae8a6e60644d6e7

closes https://github.com/official-stockfish/Stockfish/pull/3187

Bench: 3762411

9 months agoScale factor tweak
Unai Corzo [Mon, 12 Oct 2020 07:03:49 +0000 (09:03 +0200)]
Scale factor tweak

Add !pawnsOnBothFlanks heuristic to scale factor.

STC https://tests.stockfishchess.org/tests/view/5f8080575b3847b5d41f9134
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 250960 W: 49779 L: 49168 D: 152013
Ptnml(0-2): 4224, 28822, 58802, 29383, 4249

LTC https://tests.stockfishchess.org/tests/view/5f832f498ea73fb8ddf83ddb
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 88584 W: 11827 L: 11388 D: 65369
Ptnml(0-2): 585, 8079, 26578, 8412, 638

closes https://github.com/official-stockfish/Stockfish/pull/3179

bench: 3834252

9 months agoRookOnQueenFile Removal
FauziAkram [Tue, 6 Oct 2020 19:43:48 +0000 (22:43 +0300)]
RookOnQueenFile Removal

Removing Rook On Queen File looks beneficial, and it might even bring some ELO.
I will try to reintroduce it with a different method later on.

Passed STC:
https://tests.stockfishchess.org/tests/view/5f7cea204389873867eb10cb
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 18624 W: 3800 L: 3568 D: 11256
Ptnml(0-2): 308, 2131, 4257, 2253, 363

Passed LTC:
https://tests.stockfishchess.org/tests/view/5f7d76a4e936c6892bf50598
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 117864 W: 15515 L: 15340 D: 87009
Ptnml(0-2): 926, 11127, 34671, 11262, 946

closes https://github.com/official-stockfish/Stockfish/pull/3176

Bench: 3756191

9 months agoUpdate default net to nn-04cf2b4ed1da.nnue
Joost VandeVondele [Wed, 14 Oct 2020 08:23:30 +0000 (10:23 +0200)]
Update default net to nn-04cf2b4ed1da.nnue

Further tune the net parameters, now the last but one layer (32x32).
To limit the number of parameters optimized, the network layer was
decomposed using SVD, and the singular values were treated
as parameters and tuned.

Tuning branch: https://github.com/vondele/Stockfish/tree/svdTune
Tuner: https://github.com/vondele/nevergrad4sf

passed STC:
https://tests.stockfishchess.org/tests/view/5f83e82f8ea73fb8ddf83e4e
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 8488 W: 944 L: 795 D: 6749
Ptnml(0-2): 39, 609, 2811, 734, 51

passed LTC:
https://tests.stockfishchess.org/tests/view/5f83f4118ea73fb8ddf83e66
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 169016 W: 8043 L: 7589 D: 153384
Ptnml(0-2): 133, 6623, 70538, 7085, 129

closes https://github.com/official-stockfish/Stockfish/pull/3181

Bench: 3945198

9 months agoPawn Tuning
FauziAkram [Fri, 2 Oct 2020 12:32:19 +0000 (15:32 +0300)]
Pawn Tuning

Tuning of pawns, for classical evaluation:

Passed STC:
https://tests.stockfishchess.org/tests/view/5f771f0e52560f5fc78559ec
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 252696 W: 50321 L: 49692 D: 152683
Ptnml(0-2): 4614, 29845, 57049, 29978, 4862

Passed LTC:
https://tests.stockfishchess.org/tests/view/5f77cfef090dcf9aaa16d38b
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 48184 W: 6556 L: 6193 D: 35435
Ptnml(0-2): 335, 4516, 14100, 4733, 408

closes https://github.com/official-stockfish/Stockfish/pull/3169

bench: 4016121

9 months agoSimplify away futility pruning for captures
Unai Corzo [Fri, 2 Oct 2020 20:00:55 +0000 (22:00 +0200)]
Simplify away futility pruning for captures

Remove futility pruning for captures.

STC https://tests.stockfishchess.org/tests/view/5f749bfed930428c36d34c56
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 38064 W: 4011 L: 3929 D: 30124
Ptnml(0-2): 192, 3004, 12567, 3068, 201

LTC https://tests.stockfishchess.org/tests/view/5f74d99bf18675b1ce2f7412
LLR: 2.94 (-2.94,2.94) {-0.75,0.25}
Total: 184984 W: 8567 L: 8610 D: 167807
Ptnml(0-2): 146, 7593, 77058, 7548, 147

closes https://github.com/official-stockfish/Stockfish/pull/3166

bench: 3890648

9 months agoSchedule threads fairly under valgrind
Joost VandeVondele [Wed, 30 Sep 2020 19:22:36 +0000 (21:22 +0200)]
Schedule threads fairly under valgrind

fixes a rare case that can cause CI to fail when running multithreaded under valgrind.

closes https://github.com/official-stockfish/Stockfish/pull/3165

No functional change.

10 months agoInclude pawns in NNUE scaling
Stéphane Nicolet [Mon, 28 Sep 2020 20:32:55 +0000 (22:32 +0200)]
Include pawns in NNUE scaling

We now include the total pawn count in the scaling factor for the output
of the NNUE evaluation network. This should have the effect of trying to
keep more pawns when SF has the advantage, but exchange them when she
is defending.

Thanks to Alexander Pagel (Lolligerhans) for the idea of using the
value of pawns to ease the comparison with the rest of the material
estimation.

Passed STC:
LLR: 2.93 (-2.94,2.94) {-0.25,1.25}
Total: 15072 W: 1700 L: 1539 D: 11833
Ptnml(0-2): 65, 1202, 4845, 1355, 69
https://tests.stockfishchess.org/tests/view/5f7235a63b22d6afa50699b3

Passed LTC:
LLR: 2.93 (-2.94,2.94) {0.25,1.25}
Total: 25880 W: 1270 L: 1124 D: 23486
Ptnml(0-2): 23, 980, 10788, 1126, 23
https://tests.stockfishchess.org/tests/view/5f723b483b22d6afa5069a99

closes https://github.com/official-stockfish/Stockfish/pull/3164

Bench: 3776081

10 months agoTweak reduction formula.
Stefan Geschwentner [Mon, 28 Sep 2020 20:21:14 +0000 (22:21 +0200)]
Tweak reduction formula.

Replace log(i) with log(i + 0.25 * log(i)). This increases especially for low values the reductions. But for bigger values there are nearly no changes.

STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 49640 W: 5505 L: 5289 D: 38846
Ptnml(0-2): 270, 4074, 15924, 4274, 278
https://tests.stockfishchess.org/tests/view/5f71f04d3b22d6afa5069478

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 43856 W: 2209 L: 2021 D: 39626
Ptnml(0-2): 32, 1776, 18128, 1956, 36
https://tests.stockfishchess.org/tests/view/5f7232ee3b22d6afa50699a2

closes https://github.com/official-stockfish/Stockfish/pull/3163

Bench: 3555769

10 months agoUpdate default net to nn-baeb9ef2d183.nnue
SFisGOD [Mon, 28 Sep 2020 18:24:26 +0000 (02:24 +0800)]
Update default net to nn-baeb9ef2d183.nnue

Further optimization of Sergio's nn-03744f8d56d8.nnue
This patch is the result of collaboration with Joost VandeVondele.

STC:
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 37000 W: 4145 L: 3947 D: 28908
Ptnml(0-2): 191, 3016, 11912, 3166, 215
https://tests.stockfishchess.org/tests/view/5f71e7983b22d6afa5069475

LTC:
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 60224 W: 2992 L: 2769 D: 54463
Ptnml(0-2): 48, 2420, 24956, 2637, 51
https://tests.stockfishchess.org/tests/view/5f722bb83b22d6afa506998f

closes https://github.com/official-stockfish/Stockfish/pull/3161

Bench: 3720073

10 months agoTweaking Mobility and Safe Check
FauziAkram [Mon, 28 Sep 2020 19:09:43 +0000 (22:09 +0300)]
Tweaking Mobility and Safe Check

Passed STC:
https://tests.stockfishchess.org/tests/view/5f70d86d3b22d6afa50693b9
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 100368 W: 20323 L: 19914 D: 60131
Ptnml(0-2): 1927, 11641, 22605, 12118, 1893

Passed LTC:
https://tests.stockfishchess.org/tests/view/5f71bb553b22d6afa5069457
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 77648 W: 10613 L: 10181 D: 56854
Ptnml(0-2): 634, 7280, 22594, 7652, 664

closes https://github.com/official-stockfish/Stockfish/pull/3160

Bench: 3861984

10 months agoAdjust null move pruning constants
Vizvezdenec [Mon, 28 Sep 2020 17:20:06 +0000 (20:20 +0300)]
Adjust null move pruning constants

Idea is that division by fraction of 2 is slightly faster than by other numbers so parameters are adjusted in a way that division in null move pruning depth reduction features dividing by 256 instead of dividing by 213.
Other than this patch is almost non-functional - difference starts to exist by depth 133.

passed STC
https://tests.stockfishchess.org/tests/view/5f70dd943b22d6afa50693c5
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 57048 W: 6616 L: 6392 D: 44040
Ptnml(0-2): 304, 4583, 18531, 4797, 309

passed LTC
https://tests.stockfishchess.org/tests/view/5f7180db3b22d6afa506941f
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 45960 W: 2419 L: 2229 D: 41312
Ptnml(0-2): 43, 1779, 19137, 1987, 34

closes https://github.com/official-stockfish/Stockfish/pull/3159

bench 3789924

10 months agoUpdate default net to nn-04a843f8932e.nnue
Joost VandeVondele [Sat, 26 Sep 2020 15:47:52 +0000 (17:47 +0200)]
Update default net to nn-04a843f8932e.nnue

an optimization of Sergio's nn-03744f8d56d8.nnue tuning the output layer (33 parameters) on game play.

WIP code to make layer parameters tunable is https://github.com/vondele/Stockfish/tree/optionOutput
Optimization itself is using https://github.com/vondele/nevergrad4sf
Writing of the modified net using WIP code based on the learner code https://github.com/vondele/Stockfish/tree/evalWrite

Most parameters in the output layer are changed only little (~5 for int8_t).

passed STC:
https://tests.stockfishchess.org/tests/view/5f716f6b3b22d6afa506941a
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 15488 W: 1859 L: 1689 D: 11940
Ptnml(0-2): 79, 1260, 4917, 1388, 100

passed LTC:
https://tests.stockfishchess.org/tests/view/5f71908e3b22d6afa506942e
LLR: 2.93 (-2.94,2.94) {0.25,1.25}
Total: 8728 W: 518 L: 400 D: 7810
Ptnml(0-2): 7, 338, 3556, 456, 7

closes https://github.com/official-stockfish/Stockfish/pull/3158

Bench: 3789924

10 months agoUse incremental updates more often
noobpwnftw [Sun, 27 Sep 2020 18:29:21 +0000 (02:29 +0800)]
Use incremental updates more often

Use incremental updates for accumulators for up to 2 plies.
Do not copy accumulator. About 2% speedup.

Passed STC:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 21752 W: 2583 L: 2403 D: 16766
Ptnml(0-2): 128, 1761, 6923, 1931, 133
https://tests.stockfishchess.org/tests/view/5f7150cf3b22d6afa5069412

closes https://github.com/official-stockfish/Stockfish/pull/3157

No functional change

10 months agoTweak nnue scaling to keep more material
Stéphane Nicolet [Sat, 26 Sep 2020 21:19:53 +0000 (23:19 +0200)]
Tweak nnue scaling to keep more material

Current master uses a constant scale factor of 5/4 = 1.25 for the output
of the NNUE network, for compatibility with search and classical evaluation.
We modify this scale factor to make it dependent on the phase of the game,
going from about 1.5 in the opening to 1.0 for pure pawn endgames.

This helps Stockfish to avoid exchanges of pieces (heavy pieces in particular)
when she has the advantage, keeping more material on the board when attacking.

Passed STC:
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 14744 W: 1771 L: 1599 D: 11374
Ptnml(0-2): 87, 1184, 4664, 1344, 93
https://tests.stockfishchess.org/tests/view/5f6fb0a63b22d6afa506904f

Passed LTC:
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 8912 W: 512 L: 393 D: 8007
Ptnml(0-2): 7, 344, 3637, 459, 9
https://tests.stockfishchess.org/tests/view/5f6fcf533b22d6afa5069066

closes https://github.com/official-stockfish/Stockfish/pull/3154

Bench: 3943952

10 months agoSwitch to NNUE eval probabilistically for OCB
SFisGOD [Wed, 23 Sep 2020 06:00:42 +0000 (14:00 +0800)]
Switch to NNUE eval probabilistically for OCB

Introduce a small chance of switching to NNUE if PSQ imbalance is large but we have opposite colored bishops and the classical eval is struggling to win.

STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 25304 W: 3179 L: 2983 D: 19142
Ptnml(0-2): 172, 2171, 7781, 2345, 183
https://tests.stockfishchess.org/tests/view/5f6b14dec7759d4ee307cfe3

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 84680 W: 4846 L: 4556 D: 75278
Ptnml(0-2): 89, 3933, 34011, 4213, 94
https://tests.stockfishchess.org/tests/view/5f6b3fb6c7759d4ee307cff9

closes https://github.com/official-stockfish/Stockfish/pull/3146

Bench: 3865413

10 months agoSuppress info strings before 'uci'
Stéphane Nicolet [Thu, 24 Sep 2020 09:38:35 +0000 (11:38 +0200)]
Suppress info strings before 'uci'

On Windows, Stockfish wouldn't launch in some GUI because we output some
info strings (about the use of large pages) before sending the 'uci'
command. It seems more robust to suppress these info strings, and instead
to add a proper section section in the Readme about large pages use.

fixes https://github.com/official-stockfish/Stockfish/issues/3052
closes https://github.com/official-stockfish/Stockfish/pull/3147

No functional change

10 months agoIncrease reductions with the number of threads
Stéphane Nicolet [Tue, 22 Sep 2020 20:43:41 +0000 (22:43 +0200)]
Increase reductions with the number of threads

Passed STC with 8 threads:
LLR: 2.92 (-2.94,2.94) {-0.25,1.25}
Total: 13520 W: 1135 L: 1012 D: 11373
Ptnml(0-2): 39, 815, 4929, 938, 39
https://tests.stockfishchess.org/tests/view/5f68e274ded68c240be73f41

Passed LTC with 8 threads:
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 48384 W: 2183 L: 1994 D: 44207
Ptnml(0-2): 28, 1777, 20402, 1948, 37
https://tests.stockfishchess.org/tests/view/5f68f068ded68c240be747e9

closes https://github.com/official-stockfish/Stockfish/pull/3142

No functional change (for one thread)

10 months agoSmall cleanups 12
Stéphane Nicolet [Wed, 9 Sep 2020 08:49:31 +0000 (10:49 +0200)]
Small cleanups 12

- Clean signature of functions in namespace NNUE
- Add comment for countermove based pruning
- Remove bestMoveCount variable
- Add const qualifier to kpp_board_index array
- Fix spaces in get_best_thread()
- Fix indention in capture LMR code in search.cpp
- Rename TtmemDeleter to LargePageDeleter

Closes https://github.com/official-stockfish/Stockfish/pull/3063

No functional change

10 months agoAdd large page support for NNUE weights and simplify TT mem management
Sami Kiminki [Sun, 30 Aug 2020 16:41:30 +0000 (19:41 +0300)]
Add large page support for NNUE weights and simplify TT mem management

Use TT memory functions to allocate memory for the NNUE weights. This
should provide a small speed-up on systems where large pages are not
automatically used, including Windows and some Linux distributions.

Further, since we now have a wrapper for std::aligned_alloc(), we can
simplify the TT memory management a bit:

- We no longer need to store separate pointers to the hash table and
  its underlying memory allocation.
- We also get to merge the Linux-specific and default implementations
  of aligned_ttmem_alloc().

Finally, we'll enable the VirtualAlloc code path with large page
support also for Win32.

STC: https://tests.stockfishchess.org/tests/view/5f66595823a84a47b9036fba
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 14896 W: 1854 L: 1686 D: 11356
Ptnml(0-2): 65, 1224, 4742, 1312, 105

closes https://github.com/official-stockfish/Stockfish/pull/3081

No functional change.

10 months agoTweak hybrid treshold.
Stefan Geschwentner [Sun, 20 Sep 2020 20:25:19 +0000 (22:25 +0200)]
Tweak hybrid treshold.

Increase the first hybrid threshold with more material.
Rewrite the hybrid rules for clarity.

STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 24416 W: 3039 L: 2848 D: 18529
Ptnml(0-2): 135, 2136, 7503, 2271, 163
https://tests.stockfishchess.org/tests/view/5f6451efbb0cae038ca8f4dc

LTC;
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 65016 W: 3702 L: 3455 D: 57859
Ptnml(0-2): 66, 2991, 26157, 3218, 76
https://tests.stockfishchess.org/tests/view/5f64b143bb0cae038ca8f51f

closes https://github.com/official-stockfish/Stockfish/pull/3140

Bench: 3973739

10 months agoSimplify reduced depth search
Unai Corzo [Sun, 20 Sep 2020 07:03:37 +0000 (09:03 +0200)]
Simplify reduced depth search

Simplification in reduced depth search.

STC https://tests.stockfishchess.org/tests/view/5f64c72fbb0cae038ca8f531
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 28320 W: 3475 L: 3359 D: 21486
Ptnml(0-2): 170, 2485, 8773, 2523, 209

LTC https://tests.stockfishchess.org/tests/view/5f650cfabb0cae038ca8f585
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 58392 W: 3354 L: 3285 D: 51753
Ptnml(0-2): 74, 2826, 23336, 2877, 83

closes https://github.com/official-stockfish/Stockfish/pull/3139

bench: 4201295

10 months agoUse tiling to speed up accumulator refreshes and updates
syzygy1 [Wed, 16 Sep 2020 15:39:11 +0000 (17:39 +0200)]
Use tiling to speed up accumulator refreshes and updates

Perform the update and refresh operations tile by tile in a local
array of vectors. By selecting the array size carefully, we
achieve that the compiler keeps the whole array in vector registers.

Idea and original implementation by @sf-x.

STC: https://tests.stockfishchess.org/tests/view/5f623eec912c15f19854b855
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 4872 W: 623 L: 477 D: 3772
Ptnml(0-2): 14, 350, 1585, 450, 37

LTC: https://tests.stockfishchess.org/tests/view/5f62434e912c15f19854b860
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 25808 W: 1565 L: 1401 D: 22842
Ptnml(0-2): 23, 1186, 10332, 1330, 33

closes https://github.com/official-stockfish/Stockfish/pull/3130

No functional change

10 months agoSimplify futility pruning for captures
Unai Corzo [Wed, 16 Sep 2020 18:42:38 +0000 (20:42 +0200)]
Simplify futility pruning for captures

STC https://tests.stockfishchess.org/tests/view/5f61f0e4b91f2ec371e429c2
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 75512 W: 8747 L: 8704 D: 58061
Ptnml(0-2): 440, 6589, 23683, 6576, 468

LTC https://tests.stockfishchess.org/tests/view/5f6215d3912c15f19854b801
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 92912 W: 5030 L: 4992 D: 82890
Ptnml(0-2): 88, 4363, 37532, 4369, 104

closes https://github.com/official-stockfish/Stockfish/pull/3129

bench: 3856086

10 months agoRemove castling extension
Unai Corzo [Wed, 16 Sep 2020 17:14:32 +0000 (19:14 +0200)]
Remove castling extension

STC https://tests.stockfishchess.org/tests/view/5f5fa5348fbc1c8a3f476eca
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 38520 W: 4713 L: 4610 D: 29197
Ptnml(0-2): 233, 3486, 11734, 3559, 248

LTC https://tests.stockfishchess.org/tests/view/5f62166a912c15f19854b806
LLR: 2.93 (-2.94,2.94) {-0.75,0.25}
Total: 48024 W: 2673 L: 2600 D: 42751
Ptnml(0-2): 64, 2247, 19316, 2322, 63

closes https://github.com/official-stockfish/Stockfish/pull/3128

bench: 3818400

10 months agoAdded FEN string to bench output
GoldenRare [Thu, 10 Sep 2020 04:24:40 +0000 (00:24 -0400)]
Added FEN string to bench output

fixes https://github.com/official-stockfish/Stockfish/pull/3117

closes https://github.com/official-stockfish/Stockfish/pull/3118

No functional change

10 months agoImprove NDK section in Makefile
syzygy1 [Sun, 13 Sep 2020 18:16:52 +0000 (20:16 +0200)]
Improve NDK section in Makefile

This PR sets the "comp" variable simply to "clang",
which seems to be more consistent and allows a small simplification.

The PR also moves the section that sets "profile_make" and "profile_use" to after the NDK section,
which ensures that these variables are now set correctly for NDK/clang.

closes https://github.com/official-stockfish/Stockfish/pull/3121

No functional change

10 months agoUse 2 * bestMoveChanges.
xoto10 [Thu, 10 Sep 2020 20:10:57 +0000 (21:10 +0100)]
Use 2 * bestMoveChanges.

NNUE appears to provide a more stable eval than the classic eval,
so the time use dependencies on bestMoveChanges, fallingEval,
etc may need to change to make the best use of available time.
This change doubles the effect of totBestMoveChanges when giving
more time because the choice of best move is unstable.

STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 101928 W: 11995 L: 11698 D: 78235 Elo +0.78
Ptnml(0-2): 592, 8707, 32103, 8936, 626
https://tests.stockfishchess.org/tests/view/5f538a462d02727c56b36cec

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 186392 W: 10383 L: 9877 D: 166132 Elo +0.81
Ptnml(0-2): 207, 8370, 75539, 8870, 210
https://tests.stockfishchess.org/tests/view/5f54a9712d02727c56b36d5a

closes https://github.com/official-stockfish/Stockfish/pull/3119

Bench 4222126

10 months agoUpdate default net to nn-03744f8d56d8.nnue
Sergio Vieri [Mon, 14 Sep 2020 09:24:05 +0000 (17:24 +0800)]
Update default net to nn-03744f8d56d8.nnue

Equivalent to 20200914-1520

closes https://github.com/official-stockfish/Stockfish/pull/3123

Bench: 4222126

10 months agoSimplify StatSCore Initialization
mckx00 [Mon, 14 Sep 2020 02:28:32 +0000 (19:28 -0700)]
Simplify StatSCore Initialization

No need to initialize StatScore at rootNode. Current Logic is redundant because at subsequent levels the grandchildren statScore is initialized to zero.

closes https://github.com/official-stockfish/Stockfish/pull/3122

Non functional change.

10 months agoMerge remote-tracking branch 'upstream/master' into HEAD
Steinar H. Gunderson [Fri, 11 Sep 2020 17:49:12 +0000 (19:49 +0200)]
Merge remote-tracking branch 'upstream/master' into HEAD

10 months agoDouble probability of using classical eval
SFisGOD [Sun, 6 Sep 2020 20:54:26 +0000 (04:54 +0800)]
Double probability of using classical eval

This patch doubles the moderate imbalance threshold and probability of using classical eval.
So now if imbalance is greater than PawnValueMg / 4 then there is a 1/8 chance of using classical eval.

STC:
LLR: 2.93 (-2.94,2.94) {-0.25,1.25}
Total: 10984 W: 1303 L: 1140 D: 8541
Ptnml(0-2): 58, 867, 3489, 1010, 68
https://tests.stockfishchess.org/tests/view/5f554c9f97da2d5437d3813e

LTC:
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 43064 W: 2476 L: 2276 D: 38312
Ptnml(0-2): 37, 1985, 17308, 2145, 57
https://tests.stockfishchess.org/tests/view/5f55690a00a0aa2ca79f0a43

closes https://github.com/official-stockfish/Stockfish/pull/3114

Bench: 4161067

10 months agoAlways re-enable NNUE after "bench".
Gian-Carlo Pascutto [Tue, 8 Sep 2020 13:37:53 +0000 (15:37 +0200)]
Always re-enable NNUE after "bench".

Restore the default NNUE setting (enabled) after a bench command.
This also makes the resulting program settings independent of the
number of FENs that are being benched.

Fixes issue #3112.

closes https://github.com/official-stockfish/Stockfish/pull/3113

No functional change.

10 months agoBug fix in do_null_move() and NNUE simplification.
syzygy1 [Sun, 6 Sep 2020 15:29:12 +0000 (17:29 +0200)]
Bug fix in do_null_move() and NNUE simplification.

This fixes #3108 and removes some NNUE code that is currently not used.

At the moment, do_null_move() copies the accumulator from the previous
state into the new state, which is correct. It then clears the "computed_score"
flag because the side to move has changed, and with the other side to move
NNUE will return a completely different evaluation (normally with changed
sign but also with different NNUE-internal tempo bonus).

The problem is that do_null_move() clears the wrong flag. It clears the
computed_score flag of the old state, not of the new state. It turns out
that this almost never affects the search. For example, fixing it does not
change the current bench (but it does change the previous bench). This is
because the search code usually avoids calling evaluate() after a null move.

This PR corrects do_null_move() by removing the computed_score flag altogether.
The flag is not needed because nnue_evaluate() is never called twice on a position.

This PR also removes some unnecessary {}s and inserts a few blank lines
in the modified NNUE files in line with SF coding style.

Resulf ot STC non-regression test:
LLR: 2.95 (-2.94,2.94) {-1.25,0.25}
Total: 26328 W: 3118 L: 3012 D: 20198
Ptnml(0-2): 126, 2208, 8397, 2300, 133
https://tests.stockfishchess.org/tests/view/5f553ccc2d02727c56b36db1

closes https://github.com/official-stockfish/Stockfish/pull/3109

bench: 4109324

10 months agoUse classical eval more often
SFisGOD [Fri, 4 Sep 2020 09:14:50 +0000 (17:14 +0800)]
Use classical eval more often

If there is a moderate imbalance, use classical eval with small probability (1/16),
as derived from the node counter.

STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 32320 W: 3562 L: 3377 D: 25381
Ptnml(0-2): 144, 2609, 10478, 2776, 153
https://tests.stockfishchess.org/tests/view/5f520615ba100690c5cc5f80

LTC:
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 21032 W: 1116 L: 974 D: 18942
Ptnml(0-2): 20, 837, 8664, 971, 24
https://tests.stockfishchess.org/tests/view/5f522eaaba100690c5cc5f8c

closes https://github.com/official-stockfish/Stockfish/pull/3107

Bench: 4109324

10 months agoAdjust penalty on refuted early quiet moves
Vizvezdenec [Fri, 4 Sep 2020 12:53:59 +0000 (15:53 +0300)]
Adjust penalty on refuted early quiet moves

This patch changes how previous early moves are penalized in case
search finds a best move. Here, the first quiet move that was not
a transposition table move is penalized.

passed STC
https://tests.stockfishchess.org/tests/view/5f51d839ba100690c5cc5f69
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 10088 W: 1150 L: 997 D: 7941
Ptnml(0-2): 41, 772, 3278, 899, 54

passed LTC
https://tests.stockfishchess.org/tests/view/5f51e435ba100690c5cc5f76
LLR: 2.93 (-2.94,2.94) {0.25,1.25}
Total: 30808 W: 1564 L: 1405 D: 27839
Ptnml(0-2): 19, 1245, 12717, 1404, 19

closes https://github.com/official-stockfish/Stockfish/pull/3106

bench 3983758

10 months agoUpdate default net to nn-308d71810dff.nnue
Sergio Vieri [Thu, 3 Sep 2020 12:22:51 +0000 (20:22 +0800)]
Update default net to nn-308d71810dff.nnue

equivalent to 20200903-1739

Net trained from scratch, so it has quite different features extracted compared to the previous net (82215d0fd0df).

STC:
LLR: 2.98 (-2.94,2.94) {-0.25,1.25}
Total: 108328 W: 14048 L: 13719 D: 80561
Ptnml(0-2): 842, 10039, 32062, 10390, 831
https://tests.stockfishchess.org/tests/view/5f50e053ba100690c5cc5f00

LTC:
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 13872 W: 1059 L: 890 D: 11923
Ptnml(0-2): 30, 724, 5270, 871, 41
https://tests.stockfishchess.org/tests/view/5f51821fba100690c5cc5f36

closes https://github.com/official-stockfish/Stockfish/pull/3104

Bench: 3832716

10 months agoLMR Simplification
VoyagerOne [Wed, 2 Sep 2020 20:58:44 +0000 (16:58 -0400)]
LMR Simplification

remove reduction at non-check cut nodes for second move at low depths

STC:
LLR: 2.95 (-2.94,2.94) {-1.25,0.25}
Total: 61712 W: 6594 L: 6543 D: 48575
Ptnml(0-2): 293, 5085, 20082, 5070, 326
https://tests.stockfishchess.org/tests/view/5f5007d6ba100690c5cc5ea9

LTC:
LLR: 2.94 (-2.94,2.94) {-0.75,0.25}
Total: 57544 W: 2983 L: 2925 D: 51636
Ptnml(0-2): 47, 2568, 23495, 2604, 58
https://tests.stockfishchess.org/tests/view/5f50c597ba100690c5cc5ef7

closes https://github.com/official-stockfish/Stockfish/pull/3103

Bench: 3952302

10 months agoSimplify singularQuietLMR
Unai Corzo [Thu, 3 Sep 2020 10:18:42 +0000 (12:18 +0200)]
Simplify singularQuietLMR

remove formerPV dependence

STC https://tests.stockfishchess.org/tests/view/5f4cb922ba100690c5cc5d35
LLR: 2.96 (-2.94,2.94) {-1.25,0.25}
Total: 113672 W: 12347 L: 12368 D: 88957
Ptnml(0-2): 566, 9537, 36699, 9420, 614

LTC https://tests.stockfishchess.org/tests/view/5f4e8474ba100690c5cc5e12
LLR: 2.93 (-2.94,2.94) {-0.75,0.25}
Total: 43032 W: 2298 L: 2227 D: 38507
Ptnml(0-2): 45, 1940, 17475, 2011, 45

closes https://github.com/official-stockfish/Stockfish/pull/3102

bench: 3290084

10 months agoLess pruning in qsearch
Unai Corzo [Wed, 2 Sep 2020 14:45:49 +0000 (16:45 +0200)]
Less pruning in qsearch

do not prune moves that give discovery checks, even if with negative SSE.

STC https://tests.stockfishchess.org/tests/view/5f4cb5e8ba100690c5cc5d25
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 91328 W: 9940 L: 9667 D: 71721
Ptnml(0-2): 491, 7345, 29693, 7670, 465

LTC https://tests.stockfishchess.org/tests/view/5f4dbc2eba100690c5cc5dac
LLR: 2.97 (-2.94,2.94) {0.25,1.25}
Total: 52448 W: 2799 L: 2586 D: 47063
Ptnml(0-2): 53, 2220, 21459, 2445, 47

closes https://github.com/official-stockfish/Stockfish/pull/3098

bench: 4031192

10 months agoRestore development version
Joost VandeVondele [Fri, 4 Sep 2020 05:46:06 +0000 (07:46 +0200)]
Restore development version

have fun!

No functional change

10 months agoStockfish 12
Joost VandeVondele [Wed, 2 Sep 2020 07:12:04 +0000 (09:12 +0200)]
Stockfish 12

Official release version of Stockfish 12

Bench: 3624569

-----------------------

It is our pleasure to release Stockfish 12 to users world-wide

Downloads will be freely available at

https://stockfishchess.org/download/

This version 12 of Stockfish plays significantly stronger than
any of its predecessors. In a match against Stockfish 11,
Stockfish 12 will typically win at least ten times more game pairs
than it loses.

This jump in strength, visible in regular progression tests during
development[1], results from the introduction of an efficiently
updatable neural network (NNUE) for the evaluation in Stockfish[2],
and associated tuning of the engine as a whole. The concept of the
NNUE evaluation was first introduced in shogi, and ported to
Stockfish afterward. Stockfish remains a CPU-only engine, since the
NNUE networks can be very efficiently evaluated on CPUs. The
recommended parameters of the NNUE network are embedded in
distributed binaries, and Stockfish will use NNUE by default.

Both the NNUE and the classical evaluations are available, and
can be used to assign values to positions that are later used in
alpha-beta (PVS) search to find the best move. The classical
evaluation computes this value as a function of various chess
concepts, handcrafted by experts, tested and tuned using fishtest.
The NNUE evaluation computes this value with a neural network based
on basic inputs. The network is optimized and trained on the
evaluations of millions of positions.

The Stockfish project builds on a thriving community of enthusiasts
that contribute their expertise, time, and resources to build a free
and open source chess engine that is robust, widely available, and
very strong. We invite chess fans to join the fishtest testing
framework and programmers to contribute on github[3].

Stay safe and enjoy chess!

The Stockfish team

[1] https://github.com/glinscott/fishtest/wiki/Regression-Tests
[2] https://github.com/official-stockfish/Stockfish/commit/84f3e867903f62480c33243dd0ecbffd342796fc
[3] https://stockfishchess.org/get-involved/

10 months agoUpdate CPU contributors list
Joost VandeVondele [Wed, 2 Sep 2020 06:05:08 +0000 (08:05 +0200)]
Update CPU contributors list

with fishtest data of Sept. 2 2020

closes https://github.com/official-stockfish/Stockfish/pull/3095

No functional change

10 months agoOnly use MADV_RANDOM if defined
Joost VandeVondele [Tue, 1 Sep 2020 10:22:47 +0000 (12:22 +0200)]
Only use MADV_RANDOM if defined

needed to compile on Haiku.

fixes https://github.com/official-stockfish/Stockfish/issues/3093

closes https://github.com/official-stockfish/Stockfish/pull/3094

No functional change

10 months agoLMR Root Node Simplification
VoyagerOne [Sun, 30 Aug 2020 17:57:57 +0000 (13:57 -0400)]
LMR Root Node Simplification

Simplify LMR at Root node

STC:
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 71520 W: 7649 L: 7614 D: 56257
Ptnml(0-2): 346, 5845, 23349, 5868, 352
https://tests.stockfishchess.org/tests/view/5f4be8c0ba100690c5cc5cbb

LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 74832 W: 3997 L: 3948 D: 66887
Ptnml(0-2): 77, 3422, 30362, 3485, 70
https://tests.stockfishchess.org/tests/view/5f4c603eba100690c5cc5d0e

closes https://github.com/official-stockfish/Stockfish/pull/3091

Bench: 3624569

10 months agoAlways print an info line before a bestmove
Joost VandeVondele [Mon, 31 Aug 2020 20:53:20 +0000 (22:53 +0200)]
Always print an info line before a bestmove

if very few nodes are being searched before a bestmove is reported,
an info line might be missing.

fixes https://github.com/official-stockfish/Stockfish/issues/2757

closes https://github.com/official-stockfish/Stockfish/pull/3088

no functional change

10 months agoUse llvm linker with clang on windows for LTO.
mstembera [Mon, 31 Aug 2020 03:48:10 +0000 (20:48 -0700)]
Use llvm linker with clang on windows for LTO.

other linkers might fail to link during the LTO phase.

The linker might have to be installed using
`pacman -Syu mingw-w64-x86_64-lld`

closes https://github.com/official-stockfish/Stockfish/pull/3084

No functional change.

10 months agoUse stable sort to make sure bench with TB yields same results everywhere.
mstembera [Mon, 31 Aug 2020 01:40:49 +0000 (18:40 -0700)]
Use stable sort to make sure bench with TB yields same results everywhere.

std::sort() is not stable so different implementations can produce different results:
use the stable version instead.

Observed for '8/6k1/5r2/8/8/8/1K6/Q7 w - - 0 1' yielding different bench results for gcc and MSVC
and 3-4-5 syzygy TB prior to this patch.

closes https://github.com/official-stockfish/Stockfish/pull/3083

No functional change.

11 months agoUpdate parametes in classical evaluation.
FauziAkram [Sun, 30 Aug 2020 10:58:05 +0000 (13:58 +0300)]
Update parametes in classical evaluation.

Passed STC (NNUE=False):
https://tests.stockfishchess.org/tests/view/5f42edfe5089a564a10d84a0
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 13840 W: 2591 L: 2336 D: 8913
Ptnml(0-2): 194, 1453, 3387, 1676, 210

Passed LTC (NNUE=False):
https://tests.stockfishchess.org/tests/view/5f4369795089a564a10d84d8
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 159744 W: 19430 L: 18850 D: 121464
Ptnml(0-2): 960, 14185, 49030, 14709, 988

closes https://github.com/official-stockfish/Stockfish/pull/3080

bench: 3736029

11 months agoQS Pruning Simplification
VoyagerOne [Sun, 30 Aug 2020 01:13:05 +0000 (21:13 -0400)]
QS Pruning Simplification

Remove depth dependence in QS pruning

STC:
LLR: 2.95 (-2.94,2.94) {-1.25,0.25}
Total: 40536 W: 4442 L: 4358 D: 31736
Ptnml(0-2): 209, 3330, 13118, 3390, 221
https://tests.stockfishchess.org/tests/view/5f49035b3def6407861152f9

LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 97104 W: 5164 L: 5130 D: 86810
Ptnml(0-2): 103, 4478, 39377, 4470, 124
https://tests.stockfishchess.org/tests/view/5f4939d53def640786115322

closes https://github.com/official-stockfish/Stockfish/pull/3077

Bench: 3865238

11 months agoAdd / remove leaves from search tree ttPv
MJZ1977 [Fri, 28 Aug 2020 10:06:36 +0000 (12:06 +0200)]
Add / remove leaves from search tree ttPv

add if previous leaf is in search tree and we didn't find a counter move
else remove the position if the leaf is the last one in search tree.

STC : https://tests.stockfishchess.org/tests/view/5f49203c3def640786115314
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 29968 W: 3381 L: 3195 D: 23392
Ptnml(0-2): 146, 2432, 9671, 2560, 175

LTC : https://tests.stockfishchess.org/tests/view/5f494bea3def640786115336
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 84952 W: 4619 L: 4333 D: 76000
Ptnml(0-2): 86, 3765, 34481, 4065, 79

closes https://github.com/official-stockfish/Stockfish/pull/3075

Bench 3527337

11 months agoReintroduce depth reduction
Unai Corzo [Fri, 28 Aug 2020 07:27:15 +0000 (09:27 +0200)]
Reintroduce depth reduction

Reintroduce depth reduction if the position is not in TT.

STC https://tests.stockfishchess.org/tests/view/5f4652e85089a564a10d868c
LLR: 2.97 (-2.94,2.94) {-0.25,1.25}
Total: 40240 W: 4535 L: 4331 D: 31374
Ptnml(0-2): 215, 3276, 12969, 3410, 250

LTC https://tests.stockfishchess.org/tests/view/5f46ca5e5089a564a10d86f3
LLR: 2.93 (-2.94,2.94) {0.25,1.25}
Total: 63096 W: 3426 L: 3188 D: 56482
Ptnml(0-2): 51, 2798, 25645, 2970, 84

closes https://github.com/official-stockfish/Stockfish/pull/3072

bench: 3611906