]> git.sesse.net Git - stockfish/log
10 years agoExtend full 3 fold detection to PvNodes
Marco Costalba [Fri, 26 Oct 2012 09:08:06 +0000 (11:08 +0200)]
Extend full 3 fold detection to PvNodes

And restore old behaviour of not returning from a RootNode
without updating RootMoves[].

Also renamed is_draw() template parameters to reflect a
'positive' logic (Check instead of Skip) that is easier
to follow.

New bench: 5312693

10 years agoMerge pull request #34 from jromang/repetition
Marco Costalba [Fri, 26 Oct 2012 09:01:10 +0000 (02:01 -0700)]
Merge pull request #34 from jromang/repetition

Improve 3 fold repetition detection

10 years agoFull three fold repetition detection only at root node
Jean-Francois Romang [Thu, 25 Oct 2012 13:57:44 +0000 (15:57 +0200)]
Full three fold repetition detection only at root node

10 years agoEnable true 3 fold detection in search
Jean-Francois Romang [Thu, 25 Oct 2012 06:54:05 +0000 (06:54 +0000)]
Enable true 3 fold detection in search

10 years agoAllow full repetition detection
Jean-Francois Romang [Tue, 23 Oct 2012 20:33:45 +0000 (20:33 +0000)]
Allow full repetition detection

Based on sshivaji 6ee19aa5389ce60181907ba53bbb50642f2d5657 commit

10 years agoFix an assert when we stop the search
Marco Costalba [Wed, 24 Oct 2012 22:07:16 +0000 (00:07 +0200)]
Fix an assert when we stop the search

When signal 'stop' is raised we return bestValue
that could be still set at -VALUE_INFINITE and
this triggers an assert. Fix it by returning
a value we know for sure is not +-VALUE_INFINITE.

Reported by 平岡拓也 Hiraoka.

No functional change.

10 years agoRename RootPosition and shuffle think()
Marco Costalba [Wed, 24 Oct 2012 12:37:52 +0000 (14:37 +0200)]
Rename RootPosition and shuffle think()

Just slightly code reshuffles, noting interesting here...

No functional change.

10 years agoDrop Chess960 and UCIMultiPV globals and rename MultiPV
Marco Costalba [Wed, 24 Oct 2012 11:14:16 +0000 (13:14 +0200)]
Drop Chess960 and UCIMultiPV globals and rename MultiPV

No functional change.

10 years agoWrap in a class Skill Level code
Marco Costalba [Wed, 24 Oct 2012 10:05:20 +0000 (12:05 +0200)]
Wrap in a class Skill Level code

Note that the actual pickup is done in the class
d'tor so to be sure it is always triggered, even
in case of a sudden exit due to a 'stop' signal.

No functional change.

10 years agoShuffle aspiration window loop
Marco Costalba [Wed, 24 Oct 2012 08:34:36 +0000 (10:34 +0200)]
Shuffle aspiration window loop

No functional change.

10 years agoRetire refine_eval()
Marco Costalba [Mon, 22 Oct 2012 07:50:00 +0000 (09:50 +0200)]
Retire refine_eval()

Inline its content and better comment what it does through
some renaming.

No functional change.

10 years agoDon't copy a full Position object in print()
Marco Costalba [Sun, 21 Oct 2012 22:53:17 +0000 (00:53 +0200)]
Don't copy a full Position object in print()

Function move_to_san() requires the Position to be
passed by referenced because a do/undo move is done
inside the function to detect a possible mate and to
add to the san string the corresponding '#' suffix.

Instead of passing a copy of current position pass
directly the original position object after const
casting it. This has the advantage to avoid a costly
Position copy, on the down side a bench test could
report different searched nodes if print(move) is
used, due to the additionals do_move() calls.

No functional change.

10 years agoDon't need to init board[] with NO_PIECE
Marco Costalba [Sun, 21 Oct 2012 12:03:44 +0000 (14:03 +0200)]
Don't need to init board[] with NO_PIECE

Now that NO_PIECE == 0 the common memset() will
do the work.

No functional change.

10 years agoChange NO_PIECE value and shrink PieceValue[]
Marco Costalba [Sun, 21 Oct 2012 09:49:05 +0000 (11:49 +0200)]
Change NO_PIECE value and shrink PieceValue[]

This requires changing color_of() definition.

No functional change.

10 years agoUse self-describing constants instead of numbers
Marco Costalba [Sun, 21 Oct 2012 08:41:23 +0000 (10:41 +0200)]
Use self-describing constants instead of numbers

And remove now useless comments.

No functional change.

10 years agoMove RootColor from Eval to Search
Marco Costalba [Sun, 21 Oct 2012 07:12:02 +0000 (09:12 +0200)]
Move RootColor from Eval to Search

No functional change.

10 years agoContempt factor: use DrawValue also in case of stealmates
Marco Costalba [Sat, 20 Oct 2012 09:02:37 +0000 (11:02 +0200)]
Contempt factor: use DrawValue also in case of stealmates

Spotted by Jörg Oster.

No functional change (when contempt factor is not used).

10 years agoFix compatibility with old Windows 95 and 98
Marco Costalba [Sat, 20 Oct 2012 08:32:35 +0000 (10:32 +0200)]
Fix compatibility with old Windows 95 and 98

Report and patches by bnemias.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoMerge pull request #33 from daylen/master
Marco Costalba [Sat, 20 Oct 2012 08:07:27 +0000 (01:07 -0700)]
Merge pull request #33 from daylen/master

Further improve OS X compatibility

Change the minimum supported version from 10.6 to 10.0

No functional change.

10 years agoFurther improve compatibility when compiling on OS X
Daylen Yang [Tue, 16 Oct 2012 07:13:16 +0000 (00:13 -0700)]
Further improve compatibility when compiling on OS X

Change the minimum supported version from 10.6 to 10.0

10 years agoFix Makefile for PowerPC with prefetch enabled
Marco Costalba [Sun, 14 Oct 2012 23:13:41 +0000 (01:13 +0200)]
Fix Makefile for PowerPC with prefetch enabled

Existing Makefile is buggy for PowerPC, it has no
SSE, yet it is given it if Prefetch is enabled,
because it isn't ARMv7.

Patch from Matthew Brades.

No functional change.

10 years agoDocument why is safe ttValue == VALUE_NONE
Marco Costalba [Sun, 14 Oct 2012 10:36:05 +0000 (12:36 +0200)]
Document why is safe ttValue == VALUE_NONE

We can have ttValue == VALUE_NONE when we use a TT
slot to just save a position static evaluation, but
in this case we also save DEPTH_NONE so to avoid
using the ttValue in search. This happens to work,
but due to a number of lucky and tricky cases that
we now documnet through a bunch of asserts and a
little change to value_from_tt() that has no real
effect but clarifing the code.

No functional change.

10 years agoSet TT size to 32 MB during 'bench'
Marco Costalba [Sun, 14 Oct 2012 07:02:04 +0000 (09:02 +0200)]
Set TT size to 32 MB during 'bench'

On some platforms 128 MB of RAM for TT is too much,
so run 'bench' with the default 32 MB size.

No functional change although of course now 'bench'
reports a different number: 5545018

10 years agoMove all Contempt Factor code to search.cpp
Marco Costalba [Sat, 13 Oct 2012 12:38:08 +0000 (14:38 +0200)]
Move all Contempt Factor code to search.cpp

Where it is used.

No functional change.

10 years agoRetire BitCount8Bit[] table
Marco Costalba [Sat, 13 Oct 2012 12:21:27 +0000 (14:21 +0200)]
Retire BitCount8Bit[] table

Use popcount() instead in the only calling place.
It is used only at initialization so there is no
speed regression and anyhow even initialization
itself is not slowed down: magic bitboard setup
stays around 175 msec on my slow 32bit Core Duo.

No functional change.

10 years agoUse new 64 bit De Bruijn BitScan
Marco Costalba [Sat, 13 Oct 2012 11:34:50 +0000 (13:34 +0200)]
Use new 64 bit De Bruijn BitScan

Allows to sync 32 and 64 bits implementations.

Idea by Kim Walisch, reported by Gerd Isenberg:

No functional change.

10 years agoRetire can_return_tt() and rewirte TT-hit code
Marco Costalba [Sat, 13 Oct 2012 09:32:31 +0000 (11:32 +0200)]
Retire can_return_tt() and rewirte TT-hit code

Simplify the code and doing this introduce a couple
of (very small) functional changes:

- Always compare to depth even in "mate value" condition
- TT cut-off in qsearch also in case of PvNode, as in search

Verified against regression with 2500 games at 30"+0.05
on 2 threads: 451 - 444 - 1602

Functional changed: new bench is 5544977

10 years agoScale contempt factor to zero at endgame
Marco Costalba [Sat, 13 Oct 2012 08:40:38 +0000 (10:40 +0200)]
Scale contempt factor to zero at endgame

Contempt Factor is more effective at opening/middle game
and seems harmful at endgame phase. See:



Therefore we scale down the contempt factor while going
on with the game so to reach zero at endgame phase.

No functional change.

10 years agoFix a minor bug in search
Marco Costalba [Thu, 11 Oct 2012 19:11:11 +0000 (21:11 +0200)]
Fix a minor bug in search

As Joona says: "The problem is that when doing full
window search (-VALUE_INFINITE, VALUE_INFINITE), and
pruning all the moves will return fail low which is
mate score, because only clause touching alpha is
"mate distance pruning". So we are returning mate score
although we are just pruning all the moves. In reality
there probably is no mate in sight.

Bug spotted and fixed by Joona.

10 years agoARM lsb/msb assembly
Jean-Francois Romang [Thu, 11 Oct 2012 14:55:25 +0000 (22:55 +0800)]
ARM lsb/msb assembly

Implement lsb/msb using armv7 assembly instructions.
msb is the easiest one, using a gcc intrinsic that generates
code using the ARM's clz instruction. lsb is also using this
clz instruction, but with the help of ARM's 'rbit' (bit
reversing) instruction. This leads to a >2% speed gain.

I also renamed 'arm-32' to the more meaningfull 'armv7' in the Makefile

No functional change.

10 years agoIntroduce arm-32 ARCH in Makefile
Jean-Francois Romang [Wed, 10 Oct 2012 11:34:06 +0000 (19:34 +0800)]
Introduce arm-32 ARCH in Makefile

No functional change.

10 years agoRetire is_dangerous() and inline its content
Marco Costalba [Wed, 10 Oct 2012 06:09:52 +0000 (08:09 +0200)]
Retire is_dangerous() and inline its content

No functional change.

10 years agoRearrange prefetch code
Marco Costalba [Mon, 8 Oct 2012 09:19:50 +0000 (11:19 +0200)]
Rearrange prefetch code

No functional change.

10 years agoMerge pull request #29 from RyanTaker/patch-3
Marco Costalba [Sun, 7 Oct 2012 22:36:08 +0000 (15:36 -0700)]
Merge pull request #29 from RyanTaker/patch-3

Add Contempt Factor to Polyglot.ini

10 years agoAdd Contempt Factor in Polyglot
RyanTaker [Sun, 7 Oct 2012 16:49:55 +0000 (09:49 -0700)]
Add Contempt Factor in Polyglot

The contempt factor was previously not included in polyglot.ini

10 years agoSync qsearch with search
Marco Costalba [Sun, 7 Oct 2012 07:34:04 +0000 (09:34 +0200)]
Sync qsearch with search

Port to qsearch() the same changes we recently
added to search().

Overall this search refactoring series shows
almost 2% speed up on gcc compile.

No functional change.

10 years agoImprove compatibility with older versions of Mac OS X
Daylen Yang [Sun, 7 Oct 2012 00:56:12 +0000 (17:56 -0700)]
Improve compatibility with older versions of Mac OS X

Use the -mmacosx-version-min flag to support older versions of Mac OS X
(like version 10.6 and 10.7) when compiled on a machine running version

10 years agoFix POPCNT support on mingw 64
Marco Costalba [Sat, 6 Oct 2012 10:53:41 +0000 (12:53 +0200)]
Fix POPCNT support on mingw 64

When using asm 'popcnt' instruction the given
operand registers must be of the same type.

No functional change.

10 years agoFix mingw compile with ARCH=x86-64
Marco Costalba [Sat, 6 Oct 2012 10:25:13 +0000 (12:25 +0200)]
Fix mingw compile with ARCH=x86-64

When using the Makefile (as for the mingw case),
IS_64BIT and USE_BSFQ are already set with
ARCH=x86-64 and do not need to be redefined.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoFix Contempt Factor implementation
Marco Costalba [Sat, 6 Oct 2012 08:12:34 +0000 (10:12 +0200)]
Fix Contempt Factor implementation

First disable Contempt Factor during analysis, then
calculate the modified draw score from the point of
view of the player, so from the point of view of
RootPosition color.

Thanks to Ryan Taker for suggesting the fixes.

No functional change.

10 years agoFix fallouts from previous patch
Marco Costalba [Sat, 6 Oct 2012 07:09:27 +0000 (09:09 +0200)]
Fix fallouts from previous patch

These kind of arch specific code is really nasty
to make it right becuase you need to verify on
all the platforms.

Now should compile properly also on ARM

Reported by Jean-Francois.

No functional change.

10 years agoAllow prefetching on non-x86 platforms with gcc
Jean-Francois Romang [Fri, 5 Oct 2012 20:30:22 +0000 (04:30 +0800)]
Allow prefetching on non-x86 platforms with gcc

In particular on ARM processors. Original patch by
Jean-Francois, sligtly modified by me to preserve
the meaning of NO_PREFETCH flag.

Verified with gcc, clang and icc that prefetch instruction
is correctly created.

No functional change.

10 years agoRetire futility_move_count()
Marco Costalba [Fri, 5 Oct 2012 13:00:35 +0000 (15:00 +0200)]
Retire futility_move_count()

And remove (bestValue < beta) condition from
moves loop.

No functional change.

10 years agoRewrite search best value update
Marco Costalba [Wed, 3 Oct 2012 12:11:20 +0000 (14:11 +0200)]
Rewrite search best value update

A simplification and also a small speed-up of
about 1% mainly due to reducing calls to

Worst case split point recovering time after a
cut-off occurred is limited to 3 msec on my slow
PC, and usually is below 1 msec, so it seems safe
to remove the cutoff_occurred() check.

No functional change.

10 years agoAdd experimental contempt factor
Marco Costalba [Fri, 5 Oct 2012 06:23:56 +0000 (08:23 +0200)]
Add experimental contempt factor

This is very crude and very basic: simply in case
of a draw for repetition or 50 moves rule return
a negative score instead of zero according to the
contempt factor (in centipawns). If contempt is
positive engine will try to avoid draws (to use
with weaker opponents), if negative engine will
try to draw. If zero (default) there are no changes.

No functional change.

10 years agoRetire EasyMoveMargin
Marco Costalba [Wed, 3 Oct 2012 08:34:42 +0000 (10:34 +0200)]
Retire EasyMoveMargin

Use a value related to PawnValue instead.

This is a different patch from previous one because
could affect game play and skill levels, although
in a mostly unmeasurable way. Indeed thresold has
been raised so easy move is a bit harder to trigger
and skill level is a bit more prone to blunders.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoDon't hide thresolds values
Marco Costalba [Wed, 3 Oct 2012 08:08:10 +0000 (10:08 +0200)]
Don't hide thresolds values

Show the real value in the code, not hide it
behind a variable name, especially when there
is only once occurence.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoFurther push singular extension
Marco Costalba [Tue, 2 Oct 2012 04:18:21 +0000 (06:18 +0200)]
Further push singular extension

Extend for an extra half-ply in case the node is (probably)
going to fail high. In this case the added overhead is limited.

A novelity is the way this patch has been tested: Always in
self-play but with a much longer TC to allow the singular
extension to fully kick in and also (my impression) to have
less noisy results.

Ater 1015 games on my QUAD at 60"+0.05
Mod vs Orig 173 - 150 - 692 ELO +8

10 years agoFurther rearrange search()
Marco Costalba [Mon, 1 Oct 2012 07:33:13 +0000 (09:33 +0200)]
Further rearrange search()

No functional change.

10 years agoDon't push on the stack 200KB in one go
Marco Costalba [Sun, 30 Sep 2012 09:32:42 +0000 (11:32 +0200)]
Don't push on the stack 200KB in one go

This could be a limit on some platforms (as it seems
to be in Native Client). Patch from a SF fork on github:


No functional change.

10 years agoAdd support for node limited search
Marco Costalba [Sun, 30 Sep 2012 04:49:56 +0000 (06:49 +0200)]
Add support for node limited search

Handle also the SMP case. This has been quite tricky, not
trivial to enforce the node limit in SMP case becuase
with "helpful master" concept we can have recursive split
points and we cannot lock them all at once so there is the
risk of counting the same nodes more than once.

Anyhow this patch should be race free and counted nodes are

No functional change.

10 years agoSkip some useless initializations in search()
Marco Costalba [Sat, 29 Sep 2012 16:26:17 +0000 (18:26 +0200)]
Skip some useless initializations in search()

And rearrange a bit the initialization code. Still
some polishing to do, but it is a first step.

No functional change.

10 years agoDon't need to early check PV moves for legality
Marco Costalba [Sat, 29 Sep 2012 15:41:53 +0000 (17:41 +0200)]
Don't need to early check PV moves for legality

As long as isPvMove (renamed to pvMove) is set after
legality check, we can postpone legality even in PV case.

Patch aligns the PV case with the common non-pv one.

No functional change.

10 years agoDrop a magic in book.cpp
Marco Costalba [Sat, 22 Sep 2012 09:08:10 +0000 (11:08 +0200)]
Drop a magic in book.cpp

Mask out move's spacial flags without relying
on internal Move representation.

No functional change.

10 years agoTweaks to bitcount functions
Marco Costalba [Sat, 22 Sep 2012 07:26:25 +0000 (09:26 +0200)]
Tweaks to bitcount functions

Seems even a bit faster now (almost 1% in 32bit case).

No functional change.

10 years agoRestore development version
Marco Costalba [Fri, 21 Sep 2012 22:37:18 +0000 (00:37 +0200)]
Restore development version

10 years agoStockfish 2.3.1
Marco Costalba [Fri, 21 Sep 2012 22:18:00 +0000 (00:18 +0200)]
Stockfish 2.3.1

Stockfish bench signature is: 5423738

10 years agoBonus for rook/queen attacking pawns on same rank
Gary Linscott [Sun, 16 Sep 2012 13:59:30 +0000 (09:59 -0400)]
Bonus for rook/queen attacking pawns on same rank

Patch and tuning by Gary Linscott from an idea of Ryan Taker.

Double tested by Gary:

Wins: 3390 Losses: 2972 Draws: 11323
LOS: 99.999992%
ELO: 8.213465 +- 99%: 6.746506 95%: 5.124415
Win%: 51.181792 +- 99%: 0.969791 95%: 0.736740

And by me:

After 5612 games 1255  1085  3271 +11 ELO

10 years agoFix compile on 64 bits
Marco Costalba [Thu, 20 Sep 2012 17:25:27 +0000 (19:25 +0200)]
Fix compile on 64 bits

Reported by Quocvuong82.

No functional change.

10 years agoSimplify BSFTable initialization
Marco Costalba [Tue, 18 Sep 2012 08:02:20 +0000 (10:02 +0200)]
Simplify BSFTable initialization

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoFix crash under Chess 960
Marco Costalba [Sun, 16 Sep 2012 12:06:13 +0000 (14:06 +0200)]
Fix crash under Chess 960

We have a crash with this position:

rkqbnnbr/pppppppp/8/8/8/8/PPPPPPPP/RKQBNNBR w HAha -

What happens is that even if we are castling QUEEN_SIDE,
in this case we have kfrom (B8) < kto (C8) so the loop
that checks for attackers runs forever leading to a crash.

The fix is to check for (kto > kfrom) instead of
Side == KING_SIDE, but this is slower in the normal case of
ortodhox chess, so rewrite generate_castle() to handle the
chess960 case as a template parameter and allow the compiler
to optimize out the comparison in case of normal chess.

Reported by Ray Banks.

10 years agoRename class Book to PolyglotBook
Marco Costalba [Sun, 16 Sep 2012 08:30:32 +0000 (10:30 +0200)]
Rename class Book to PolyglotBook

And move struct BookEntry out of the header where it is
not needed.

No functional change.

10 years agoFix KpsK endgame
Marco Costalba [Sun, 16 Sep 2012 06:52:38 +0000 (08:52 +0200)]
Fix KpsK endgame

Broken by commit a44c5cf4f77b05a03 of 3 /12 / 2011 that
was labeled "No functional change" because our 'bench'
test didn't triggered that particular endgame. Indeed
we need to run a specific bench on a set of endgames
position when touching endgame.cpp because normal bench
does not cover endgames properly.

Found by MSVC 2012 code analyzer.

10 years agoRestore development version
Marco Costalba [Sat, 15 Sep 2012 07:34:31 +0000 (09:34 +0200)]
Restore development version

No functional change.

10 years agoStockfish 2.3
Marco Costalba [Sat, 30 Jun 2012 07:00:48 +0000 (08:00 +0100)]
Stockfish 2.3

Stockfish bench signature is: 5416292

10 years agoFix compile with Intel 13.0
Marco Costalba [Sat, 15 Sep 2012 08:50:23 +0000 (10:50 +0200)]
Fix compile with Intel 13.0

It seems Intel is unable to properly workout templates with 'static'
storage specifier.

Workaround using an anonymous namespace instead.

No functional change.

10 years agoDon't exit earlier from aspiration window loop
Marco Costalba [Mon, 10 Sep 2012 16:24:53 +0000 (18:24 +0200)]
Don't exit earlier from aspiration window loop

Currently we exit the loop when

abs(bestValue) >= VALUE_KNOWN_WIN

but there is no logical reason for this. It seems more
natural to re-search again with full open window.

This has practically no impact in most cases, we have a
'no functional change' running 'bench' command.

10 years agoFix MSVC 2012 64bits warnings
Marco Costalba [Fri, 14 Sep 2012 07:42:59 +0000 (09:42 +0200)]
Fix MSVC 2012 64bits warnings

Reported by Rein.

No functional change.

10 years agoSpeed-up generate<LEGAL>
Marco Costalba [Tue, 11 Sep 2012 18:00:58 +0000 (20:00 +0200)]
Speed-up generate<LEGAL>

The trick here is to check for legality only in the
(rare) cases we have pinned pieces or a king move
or an en-passant.

This trick is able to increase the speed of perft
of more then 20%!

No functional change.

10 years agoRemove redundancy in move generation
Marco Costalba [Sun, 9 Sep 2012 12:30:22 +0000 (14:30 +0200)]
Remove redundancy in move generation

Introduce generate_all_moves() and remove a good
bunch of redundant code.

No functional change.

10 years agoSimplify generate_castle()
Marco Costalba [Sun, 9 Sep 2012 08:49:25 +0000 (10:49 +0200)]
Simplify generate_castle()

Skipping the calls to std::min(), std::man() we get
even a nice speed-up on perft.

No functional change.

10 years agoRename *last to *end
Marco Costalba [Sun, 9 Sep 2012 08:17:31 +0000 (10:17 +0200)]
Rename *last to *end

It is a more correct name because it points past the
last move of the list.

No functional change.

10 years agoEnable link time optimization only when optimizing
Marco Costalba [Sun, 9 Sep 2012 08:02:11 +0000 (10:02 +0200)]
Enable link time optimization only when optimizing

Because it is quite slow, skip it when 'optimize' flag is 'no'

No functional change.

10 years agoRevert "Simplify Option c'tor"
Marco Costalba [Fri, 7 Sep 2012 13:21:50 +0000 (15:21 +0200)]
Revert "Simplify Option c'tor"

std::to_string() is C++11 material, not c++03.

So revert the patch.

10 years agoSimplify Option c'tor
Marco Costalba [Thu, 6 Sep 2012 16:16:37 +0000 (18:16 +0200)]
Simplify Option c'tor

No functional change.

10 years agoGet rid of struct Time
Marco Costalba [Tue, 4 Sep 2012 07:38:51 +0000 (09:38 +0200)]
Get rid of struct Time

We just need the milliseconds of current system
time for our needs. This allows to simplify the

No functional change.

10 years agoRename current_time() to now()
Marco Costalba [Sun, 2 Sep 2012 14:39:01 +0000 (16:39 +0200)]
Rename current_time() to now()

Follow C++11 naming conventions.

No functional change.

10 years agoGreatly speed up SEE
Marco Costalba [Sat, 1 Sep 2012 22:15:27 +0000 (00:15 +0200)]
Greatly speed up SEE

Simply reshuffling the code inverting the condition in next_attacker()
yields a miraculous speed up of more than 3% under gcc!

On my laptop a bench run goes from 320Knps to 330Knps

No functional change.

10 years agoUnroll least valuable attacker loop in SEE
Marco Costalba [Sat, 1 Sep 2012 09:45:14 +0000 (11:45 +0200)]
Unroll least valuable attacker loop in SEE

This allows to reduce the scanning for new X-ray attacks
according to the capturing piece type.

It seems to be just a very small speed increase in MSVC 64
bit and gcc 32 bit, I guess cache issues value more than some
instruction less to execute (as usual).

No functional change.

10 years agoSlightly simplify SEE
Marco Costalba [Sat, 1 Sep 2012 09:45:14 +0000 (11:45 +0200)]
Slightly simplify SEE

Some renaming and small code reshuffle.

No fuctional change.

10 years agoRetire Time::restart()
Marco Costalba [Fri, 31 Aug 2012 13:19:35 +0000 (15:19 +0200)]
Retire Time::restart()

Simplify API.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoDon't need to memset HashTable
Marco Costalba [Fri, 31 Aug 2012 12:21:16 +0000 (14:21 +0200)]
Don't need to memset HashTable

Default c'tor Entry() already initializes
to zero all its POD members.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoTerminate threads before to exit main()
Marco Costalba [Wed, 29 Aug 2012 14:43:01 +0000 (16:43 +0200)]
Terminate threads before to exit main()

It is very difficult and risky to assure
that a running thread doesn't access a global
variable. This is currently true, but could
change in the future and we don't want to rely
on code that works 'by accident'. The threads
are still running when ThreadPool destructor is
called (after main() returns) and this could
lead to crashes if a thread accesses a global
that has been already freed. The solution is to
use an exit() function and call it while we are
still in main(), ensuring global variables are
still alive at threads termination time.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoConvert to sync_cout and sync_endl
Marco Costalba [Wed, 29 Aug 2012 11:28:59 +0000 (13:28 +0200)]
Convert to sync_cout and sync_endl

Serialize access to std::cout all over the code.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoIntroduce serialization of accesses to std::cout
Marco Costalba [Wed, 29 Aug 2012 09:25:11 +0000 (11:25 +0200)]
Introduce serialization of accesses to std::cout

When many threds concurrently print you need to serialize
the access to std::cout to avoid output lines are intermixed
with the contents of each thread.

This is not strictly needed at the moment because
only main thread prints out, although some ad-hoc
test could trigger UCI::loop() printing while searching.

Anyhow we want to lift this pretty avoidable constrain
also as a prerequisite for future work.

This patch just introduces the support, next one will enable
the serialization.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoCorrectly handle handover of setup states
Marco Costalba [Mon, 27 Aug 2012 12:39:59 +0000 (14:39 +0200)]
Correctly handle handover of setup states

Before the search we setup the starting position doing all the
moves (sent by GUI) from start position to the position just
before to start searching.

To do this we use a set of StateInfo records used by each
do_move() call. These records shall be kept valid during all
the search because repetition draw detection uses them to back
track all the earlier positions keys. The problem is that, while
searching, the GUI could send another 'position' command, this
calls set_position() that clears the states! Of course a crash
follows shortly.

Before searching all the relevant parameters are copied in
start_searching() just for this reason: to fully detach data
accessed during the search from the UCI protocol handling.
So the natural solution would be to copy also the setup states.
Unfortunatly this approach does not work because StateInfo
contains a pointer to the previous record, so naively copying and
then freeing the original memory leads to a crash.

That's why we use two std::auto_ptr (one belonging to UCI and another
to Search) to safely transfer ownership of the StateInfo records to
the search, after we have setup the root position.

As a nice side-effect all the possible memory leaks are magically
sorted out for us by std::auto_ptr semantic.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoUse std::deque to store setup states
Marco Costalba [Sun, 26 Aug 2012 16:07:54 +0000 (18:07 +0200)]
Use std::deque to store setup states

10 years agoDocument PolyGlotRandoms[] offsets
Marco Costalba [Sat, 25 Aug 2012 17:48:08 +0000 (18:48 +0100)]
Document PolyGlotRandoms[] offsets

Should be more clear from where the 'magic' numbers
come from.

Also bit of reformat while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoExplicitly use threads.size()
Marco Costalba [Fri, 24 Aug 2012 11:40:58 +0000 (12:40 +0100)]
Explicitly use threads.size()

Instead of just size(). Although code is longer,
should be more immediate to understand when reading.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoIntroduce struct Mutex and ConditionVariable
Marco Costalba [Fri, 24 Aug 2012 09:59:41 +0000 (10:59 +0100)]
Introduce struct Mutex and ConditionVariable

To mimics C++11 std::mutex and std::condition_variable,
also rename locks and condition variables to be more
uniform across the classes.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoSlightly microptimize SEE
Marco Costalba [Fri, 24 Aug 2012 08:30:50 +0000 (09:30 +0100)]
Slightly microptimize SEE

Reduce of one instruction. It seems a tad faster on
the profiler now. Very slightly but anyhow it is a
code semplification.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoUse size_t as operator[] argument type
Marco Costalba [Wed, 22 Aug 2012 10:44:43 +0000 (11:44 +0100)]
Use size_t as operator[] argument type

This better mimics std::vector::operator[] and
fixes a warning with MSVC 64bit.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoMerge generate_direct_checks() in generate_moves()
Marco Costalba [Wed, 22 Aug 2012 09:13:09 +0000 (10:13 +0100)]
Merge generate_direct_checks() in generate_moves()

Further reduce redundancy in move generation.
Veirifed no speed regression on MSVC, Clang and gcc.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoStreamline generate_moves()
Marco Costalba [Wed, 22 Aug 2012 07:29:49 +0000 (08:29 +0100)]
Streamline generate_moves()

Greatly simplify these very performace critical functions.
Amazingly we don't have any speed regression actually under
MSVC we have the same assembly for generate_moves() !

In generate_direct_checks() 'target' is calculated only
once being a loop invariant.

On Clang there is even a slight speed up.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoReformat piece values arrays
Marco Costalba [Mon, 20 Aug 2012 17:41:28 +0000 (18:41 +0100)]
Reformat piece values arrays

And rename stuff while there.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoMove zobrist keys out of Position
Marco Costalba [Mon, 20 Aug 2012 17:09:57 +0000 (18:09 +0100)]
Move zobrist keys out of Position

Are used by Position but do not belong to that class,
there is only one instance of them (that's why were
defined as static), so move to a proper namespace instead.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoRetire copy c'tor from class Position
Marco Costalba [Mon, 20 Aug 2012 13:44:30 +0000 (14:44 +0100)]
Retire copy c'tor from class Position

Not needed.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoPrefer a reference to a pointer
Marco Costalba [Mon, 20 Aug 2012 06:54:38 +0000 (07:54 +0100)]
Prefer a reference to a pointer

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoUse enums instead of constants for piece values
Marco Costalba [Sun, 19 Aug 2012 23:04:42 +0000 (00:04 +0100)]
Use enums instead of constants for piece values

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
10 years agoDocument De Bruijn sequences
Marco Costalba [Sun, 19 Aug 2012 13:12:39 +0000 (14:12 +0100)]
Document De Bruijn sequences

Insted of raw magic numbers.

No functional change.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>