Stockfish Development Versions are build automatically if there are changes on the master branch in the git repository (https://github.com/official-stockfish/Stockfish). Use it at your own risk.
They are compiled with gcc 5.3.1/mingw 5.3.1 on Ubuntu 16.04.
2.Jan.2016: BMI build now slightly faster due to profiling.
Development builds for OS X are available at stockfishchess.de
Read the FAQ or leave a comment.

goto Gull
!! latest version !!


Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: VoyagerOne
Date: Tue Aug 30 09:09:55 2016 +0200
Timestamp: 1472540995

Remove condition on killers in history pruning

Now allows main killer to be history prune.

STC:
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 15852 W: 2910 L: 2781 D: 10161

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 56428 W: 7610 L: 7537 D: 41281

Bench: 8032058
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Stefan Geschwentner
Date: Mon Aug 29 22:15:03 2016 +0200
Timestamp: 1472501703

Tweak probcut threshold

Use better threshold for capture move generation.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 23265 W: 4415 L: 4188 D: 14662

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 36618 W: 5083 L: 4836 D: 26699

bench: 7030088
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Andrew Grant
Date: Sun Aug 28 09:47:30 2016 +0200
Timestamp: 1472370450

Removed an extra space

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Alain SAVARD
Date: Sun Aug 28 08:49:40 2016 +0200
Timestamp: 1472366980

Move king tropism to evaluate_king

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Sun Aug 28 08:08:13 2016 +0200
Timestamp: 1472364493

Retire CheckInfo

Move its content directly under StateInfo.

Verified for no speed regression.

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Sat Aug 27 12:16:13 2016 +0200
Timestamp: 1472292973

Silence some warnings with MSVC 2013

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Stéphane Nicolet
Date: Sat Aug 27 09:53:26 2016 +0200
Timestamp: 1472284406

Move CheckInfo under StateInfo

This greately simplifies usage because hides to the
search the implementation specific CheckInfo.

This is based on the work done by Marco in pull request #716,
implementing on top of it the ideas in the discussion: caching
the calls to slider_blockers() in the CheckInfo structure,
and simplifying the slider_blockers() function by removing its
first parameter.

Compared to master, bench is identical but the number of calls
to slider_blockers() during bench goes down from 22461515 to 18853422,
hopefully being a little bit faster overall.

archlinux, gcc-6
make profile-build ARCH=x86-64-bmi2
50 runs each

bench:
base = 2356320 +/- 981
test = 2403811 +/- 981
diff = 47490 +/- 1828

speedup = 0.0202
P(speedup > 0) = 1.0000

perft 6:
base = 175498484 +/- 429925
test = 183997959 +/- 429925
diff = 8499474 +/- 469401

speedup = 0.0484
P(speedup > 0) = 1.0000

perft 7 (but only 10 runs):
base = 185403228 +/- 468705
test = 188777591 +/- 468705
diff = 3374363 +/- 476687

speedup = 0.0182
P(speedup > 0) = 1.0000

$ ./pyshbench ../Stockfish/master ../Stockfish/test 20
run base test diff
...

base = 2501728 +/- 182034
test = 2532997 +/- 182034
diff = 31268 +/- 5116

speedup = 0.0125
P(speedup > 0) = 1.0000

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Sat Aug 27 09:12:25 2016 +0200
Timestamp: 1472281945

Make engine ONE_PLY value independent

This non-functional change patch is a deep work to allow SF to be independent
from the actual value of ONE_PLY (currently set to 1). I have verified SF is
now independent for ONE_PLY values 1, 2, 4, 8, 16, 32 and 256.

This patch gives consistency to search code and enables future work, opening
the door to safely tweaking the ONE_PLY value for any reason.

Verified for no speed regression at STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 95643 W: 17728 L: 17737 D: 60178

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: gamander
Date: Sat Aug 27 08:48:07 2016 +0200
Timestamp: 1472280487

Fixed wrong definition of WhiteCamp and BlackCamp

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: hxim
Date: Sat Aug 27 08:39:32 2016 +0200
Timestamp: 1472279972

Simplify stats update

Simplify code by moving countermove and follow-up move
history update into procedure.

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Thu Aug 25 09:51:50 2016 +0200
Timestamp: 1472111510

Reformat stats update

Rewritten in a way to have explicit in the search
the bonus/penalty we apply: hopefully this will lead
to further simplification/fix of current rather messy
stats update code.

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: VoyagerOne
Date: Wed Aug 24 08:26:32 2016 +0200
Timestamp: 1472019992

Refutation penalty on captures

Apply refutation penalty for prior PV quiet move on captures

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 8208 W: 1153 L: 1008 D: 6047
http://tests.stockfishchess.org/tests/view/57bc5a9f0ebc59030fbe47b5

Only LTC because a very similar patch already passed STC + LTC

bench: 7038730
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: VoyagerOne
Date: Fri Aug 19 08:59:18 2016 +0200
Timestamp: 1471589958

Simplify IID

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 30468 W: 5687 L: 5582 D: 19199
http://tests.stockfishchess.org/tests/view/57b1ddd80ebc591c761f63e2

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 87406 W: 11756 L: 11725 D: 63925
http://tests.stockfishchess.org/tests/view/57b212590ebc591c761f63f9

bench: 6554900
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: VoyagerOne
Date: Fri Aug 19 08:50:14 2016 +0200
Timestamp: 1471589414

Do LMR on captures

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 5361 W: 1086 L: 936 D: 3339
http://tests.stockfishchess.org/tests/view/57b31b0f0ebc591c761f643d

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 54694 W: 7591 L: 7287 D: 39816
http://tests.stockfishchess.org/tests/view/57b3442b0ebc591c761f6450

bench: 6881120
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Thu Aug 18 10:40:56 2016 +0200
Timestamp: 1471509656

Remove a stale assignment

No more used after previous patch.

Spotted by Jekaa .

No functional change.
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Stefano80
Date: Thu Aug 18 10:16:56 2016 +0200
Timestamp: 1471508216

Retire pawn span

Retire pawn span and replace with pawn count in evaluate_scale_factor.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 26482 W: 4929 L: 4818 D: 16735

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 61938 W: 8400 L: 8335 D: 45203

Bench: 7662861
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: VoyagerOne
Date: Mon Aug 15 14:45:40 2016 +0200
Timestamp: 1471265140

Use predicted depth for history pruning

STC: (Yellow)
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 69115 W: 12880 L: 12797 D: 43438

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 124163 W: 16923 L: 16442 D: 90798

Note: Note based off past experiments / patches... history pruning
is quite TC sensitive. I believe the reason for this TC dependency
is that the CMH/FMH is a very large table that takes time to fill
up with. In addition having more time for will increase the accuracy
of the stats' value.

Bench: 7351698
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Luca Brivio
Date: Sat Aug 13 10:12:09 2016 +0200
Timestamp: 1471075929

Cap space evaluation bonus

When computing space evaluation, limit the bonus square count to 16.

STC @ 10+0.1 th 1:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 30793 W: 5910 L: 5648 D: 19235

LTC @ 60+0.6 th 1:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 31361 W: 4410 L: 4184 D: 22767

Bench: 7165385
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Alain SAVARD
Date: Wed Aug 10 12:49:24 2016 +0200
Timestamp: 1470826164

Simplify space formula

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 16868 W: 3260 L: 3132 D: 10476

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 16910 W: 2381 L: 2255 D: 12274

bench: 6663531
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: VoyagerOne
Date: Tue Aug 2 09:17:14 2016 +0200
Timestamp: 1470122234

Use Color-From-To history stats to help sort moves

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 33502 W: 6498 L: 6223 D: 20781
http://tests.stockfishchess.org/tests/view/578abb940ebc5972faa169e2

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 50782 W: 7124 L: 6832 D: 36826
http://tests.stockfishchess.org/tests/view/578b8e5d0ebc5972faa169fd

LTC: (Sanity test against latest master)
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 32759 W: 4600 L: 4370 D: 23789
http://tests.stockfishchess.org/tests/view/5798b7d30ebc591c761f5b72

bench: 6985912

P.S. Thanks @mstembera for rewriting my code to make it smp compatible. A BIG thank you!
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: VoyagerOne
Date: Mon Aug 1 21:47:42 2016 +0200
Timestamp: 1470080862

Futility tweak

Use a different margin for pruning child nodes.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16692 W: 3251 L: 3051 D: 10390
http://tests.stockfishchess.org/tests/view/579b95d10ebc591c761f5c03

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 24140 W: 3501 L: 3297 D: 17342
http://tests.stockfishchess.org/tests/view/579bb15d0ebc591c761f5c0b

Bench: 7927017
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: ajithcj
Date: Sun Jul 24 09:54:07 2016 +0200
Timestamp: 1469346847

Allow null pruning at depth 1

This removes a check that prevents null pruning at depth 1 PLY.

STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 23445 W: 4638 L: 4521 D: 14286

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 61416 W: 8627 L: 8563 D: 44226

bench: 8145304
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Stefano80
Date: Sun Jul 24 08:55:58 2016 +0200
Timestamp: 1469343358

See prune at higher depth

Allow SEE pruning at higher depths in shallow depth
pruning using a threshold increasing with depth.

STC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 35366 W: 7011 L: 6724 D: 21631

LTC
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 15578 W: 2243 L: 2070 D: 11265

Bench: 8417887
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: ajithcj
Date: Sat Jul 16 07:35:31 2016 +0200
Timestamp: 1468647331

Gradually relax the NMP staticEval check

Gradually relax the NMP staticEval check as we go to
higher depths.

Use tuned values.

STC
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 16745 W: 3371 L: 3168 D: 10206

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 5906 W: 875 L: 736 D: 4295

bench: 8548212
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Lyudmil Antonov
Date: Sat Jul 16 07:06:13 2016 +0200
Timestamp: 1468645573

Workaround gcc stack alignment bug

GCC uses SSE instructions to move data but in 32-bit gcc version used
by abrok the stack is not 16-byte aligned due to a bug.

This patch workaround teh bug by not using the stack
to store KingFlank[]

Fixes issue #721.

No functional change.

next page >