]>
git.sesse.net Git - stockfish/log
Marco Costalba [Mon, 25 Apr 2011 09:46:52 +0000 (10:46 +0100)]
Tidy up benchmark.cpp
Node count is different just becuase now we don't log on
"bench.txt" file anymore so that we avoid some calls to
pretty_pv() that calls Position::do_move().
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 24 Apr 2011 23:22:48 +0000 (00:22 +0100)]
Large API rename in ThreadsManager
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 24 Apr 2011 17:46:26 +0000 (18:46 +0100)]
Don't allocate MAX_THREADS hash tables if not necessary
This prevent crashing on mobile devices with limited RAM,
currently with MAX_THREADS = 32 we would need 44MB that
could be too much for a poor cellphone.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 24 Apr 2011 08:20:03 +0000 (09:20 +0100)]
Move pawn and material tables under Thread class
This change allows to remove some quite a bit of code
and seems the natural thing to do.
Introduced file thread.cpp to move away from search.cpp a lot
of threads related stuff.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 24 Apr 2011 07:54:36 +0000 (08:54 +0100)]
Rename MOVES_MAX in MAX_MOVES
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 24 Apr 2011 07:31:47 +0000 (08:31 +0100)]
Retire extensions as UCI option
There is no real need why an user should change these values.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 24 Apr 2011 07:18:39 +0000 (08:18 +0100)]
Document why we use per-thread pawn and material tables
Arisen from a discussion on talkchess.
No fnctional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 23 Apr 2011 14:42:23 +0000 (15:42 +0100)]
Fix some comments in early stop detection
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 23 Apr 2011 13:57:05 +0000 (14:57 +0100)]
Retire UseLogFile in search.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 22 Apr 2011 13:52:03 +0000 (15:52 +0200)]
Introduce and use SearchLimits
Pack a bit of global variables related to search limits in
a single struct.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 22 Apr 2011 10:02:54 +0000 (12:02 +0200)]
Use move_is_special() in pawn endgame condition
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 22 Apr 2011 09:57:22 +0000 (11:57 +0200)]
Retire SearchStartTime global
Use a static variable inside current_search_time() instead.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 22 Apr 2011 07:55:47 +0000 (09:55 +0200)]
Reduce loops in init_threads() and exit_threads()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 20 Apr 2011 09:18:36 +0000 (11:18 +0200)]
Move wake_sleeping_thread() to Thread class
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 19 Apr 2011 08:24:43 +0000 (10:24 +0200)]
Correctly implementg selDepth feature
Send to GUI the deepest search depth apart from
qsearch of the PV line.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 19 Apr 2011 07:56:59 +0000 (09:56 +0200)]
Move sleepLock and sleepCond under Thread
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 19 Apr 2011 07:29:28 +0000 (09:29 +0200)]
Increase risk of blunders at low skill levels
According to Heinz's tests current setup is in fact too
strong for weak players. This seems the best according
to his tests.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 16 Apr 2011 09:25:22 +0000 (10:25 +0100)]
Use only history to score non captures
It seems gain is practically unuseful, so remove.
After 13554 games:
Mod vs Orig 2252 - 2319 - 8983 ELO -1 (+- 3.4)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 17 Apr 2011 09:31:26 +0000 (10:31 +0100)]
Small simplification in scale_by_game_phase()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 16 Apr 2011 13:41:53 +0000 (14:41 +0100)]
Move ply to SearchStack
Shrink search() signature for better readibility.
We get also a nice 1.3% speed increase.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 16 Apr 2011 10:24:30 +0000 (11:24 +0100)]
Remove one indentation level in get_next_move()
Small renaming and fix some comments.
No functional and no speed change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 15 Apr 2011 16:14:12 +0000 (18:14 +0200)]
Code style in tt.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 15 Apr 2011 15:41:45 +0000 (17:41 +0200)]
Better self-document LMR reduction() formula
Suggested by Onno
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 14 Apr 2011 13:59:49 +0000 (15:59 +0200)]
Promote OptionsMap to a class
And add a bit of documentation too.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 16 Apr 2011 09:21:13 +0000 (10:21 +0100)]
Fix a stale comment
Spotted by Onno
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 14 Apr 2011 06:06:56 +0000 (07:06 +0100)]
Remove src/COPYING file
It is enough the one in the base directory
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 14 Apr 2011 06:04:32 +0000 (07:04 +0100)]
Remove "divide by zero" workaround
It is now useless because of the condition at the beginning.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 13 Apr 2011 13:12:33 +0000 (15:12 +0200)]
Cleanup debug counters
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 13 Apr 2011 09:54:41 +0000 (11:54 +0200)]
Move move_is_legal() under Position class
It is a more logical place than in move generation file.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 13 Apr 2011 07:08:19 +0000 (09:08 +0200)]
Some more cleanup in endgame.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 12 Apr 2011 05:27:00 +0000 (07:27 +0200)]
Fix some warnings and a compile error with icc
Unfortunatly icc does not understand that weakerSide and
strongerSide belongs to the base class :-(
So we have define them in the derived class.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 11 Apr 2011 16:12:41 +0000 (18:12 +0200)]
Move EndgameFunctions to endgame.cpp
And cleanup code while there.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 11 Apr 2011 13:33:35 +0000 (15:33 +0200)]
Increase MaterialTableSize 8 times
Now that we prefetch in material hash table we
can increase its size and gain something.
Hit rate is now of 98% from 92%
Speedup of 0.8%
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 11 Apr 2011 12:50:56 +0000 (14:50 +0200)]
Introduce and use NoPawnsSF[] in material.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 11 Apr 2011 09:32:55 +0000 (11:32 +0200)]
Tempeltize material imbalance
Speedup of almost 1%
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 11 Apr 2011 07:54:41 +0000 (09:54 +0200)]
Sync material.h with pawns.h
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 11 Apr 2011 07:03:34 +0000 (09:03 +0200)]
Prefetch also material tables
Prefetch both pawn and material tables in do_move() and
prefetch always, not only after a pawn move or a capture.
Speed up of 0,7%
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 9 Apr 2011 18:50:56 +0000 (19:50 +0100)]
Assorted code style and comments in pawns.cpp and pawns.h
The only interesting thing is that a backward or isolated
pawn cannot be a candidate passer, so code this condition.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 8 Apr 2011 06:48:05 +0000 (07:48 +0100)]
Retire mate threat extension
It seems we have a lot of totally useless code !
After 8577 games 1504 - 1451 - 5622 ELO +2 (+- 4.4)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 7 Apr 2011 07:09:41 +0000 (08:09 +0100)]
Fix a compile error in debug mode
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 3 Apr 2011 08:19:08 +0000 (09:19 +0100)]
Assorted code style and comments in search.cpp
Nothing really serious....
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 4 Apr 2011 07:53:44 +0000 (09:53 +0200)]
Added -Wshadow option and fixed resulting warnings
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 2 Apr 2011 08:05:53 +0000 (09:05 +0100)]
Teach SF to blunder
Add blunder cabability to skill level feature.
The idea is that instead of choosing the best move at the end
of the ID loop, we now do this at a randomly chosen sampling depth
dependent on SkillLevel, so that at low skill levels we sample when
ID loop has reached only a small depth and so we have an higher
probability to pick up a blunder.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 2 Apr 2011 07:15:05 +0000 (08:15 +0100)]
Use prob cut search to prune bad captures
The idea is to try a shallow search with reduced beta
on bad captures so to quickly prune them out in case
are really bad.
After 5529 games 966 - 868 - 3695 ELO +6 (+- 5.4) LOS 91%
Tested also version without upper limitation to 8 plies:
After 8780 games 1537 - 1398 - 5850 ELO +5 (+- 4.3) LOS 93%
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Mar 2011 11:19:51 +0000 (12:19 +0100)]
Retire recapture extension also for PvNode
Seems that extension is useless.
After 10105 games
Mod vs Orig 1738 - 1702 - 6665 ELO +1 (+- 4)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 1 Apr 2011 12:35:39 +0000 (14:35 +0200)]
Use a constant instead of value_mate_in(PLY_MAX)
And also apply the same to value_mated_in(PLY_MAX)
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 1 Apr 2011 12:19:51 +0000 (14:19 +0200)]
Retire value_is_mate()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 1 Apr 2011 11:48:13 +0000 (13:48 +0200)]
Simplify wait_for_stop_or_ponderhit()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 1 Apr 2011 11:36:14 +0000 (13:36 +0200)]
Retire update_killers()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 27 Mar 2011 17:54:02 +0000 (18:54 +0100)]
Fine tune skill level
Rescaled Skill level from 0 to 20. At level 19 is still
comparable with Crafty 20.14, while at low levels strength
increase is now less steep.
Thanks to Joona and Heinz for testing and valuable
comments.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 27 Mar 2011 10:23:29 +0000 (11:23 +0100)]
Add "Skill level functionality
It is now possible to adjust skill level of Stockfish
from 10 (full strength) to 0.
Skill adjustment is done in such a way that is CPU speed and
time control largely independent, at least at low skills. It
means that given a skill we have same play level on a mobile
phone and on a super OCTAL CPU, at 1' per game or at 180'.
At skill 9 strength is that of an average engine, I have used
Crafty 20.14 to tune and we are more or less there. At skill 0
engine is pretty weak but still shows a realistic play.
When skill is not used we don't have any impact on the regular
code.
Idea to use MultiPV is from Heinz van Saanen, implementation and
formulas by me.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Mar 2011 07:42:38 +0000 (08:42 +0100)]
Fix a compile error with icc
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 24 Mar 2011 10:06:00 +0000 (11:06 +0100)]
Send PV line to GUI only after resolving a fail high
This is how Shredder, Rybka and others do and
avoids user is confused by a fail high (sent to GUI)
followed by a fail low (not sent).
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 24 Mar 2011 09:05:14 +0000 (10:05 +0100)]
Triviality in position.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Mar 2011 09:50:50 +0000 (10:50 +0100)]
Retire move_ambiguity() altogether
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 22 Mar 2011 08:24:14 +0000 (09:24 +0100)]
Simplify move_ambiguity()
And additional small touches in move.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 19 Mar 2011 15:00:09 +0000 (16:00 +0100)]
Do not send ponder move if we don't have it
Has been reported by Justin Blanchard that
this creates problems on some buggy GUI.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 17 Mar 2011 12:47:15 +0000 (13:47 +0100)]
Use intrinsic in pop_1st_bit() under MSVC 64 bits
Around 1% speedup when compiled with MSVC 64
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Mar 2011 15:57:02 +0000 (16:57 +0100)]
Additional cleanup in bitbase.cpp
Also better document what code does.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Mar 2011 11:04:02 +0000 (12:04 +0100)]
Move KPKBitbase[] where it belongs
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Mar 2011 10:22:02 +0000 (11:22 +0100)]
Fix a couple of issues in bitbase.cpp
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 10 Mar 2011 13:39:53 +0000 (14:39 +0100)]
Reintroduce initialization of some bitboards
With off-by-one bug in InFrontBB[] loop fixed.
Also use int instead of File to workaround a bug
in mingw 4.4.0 in first loop that cycles forever.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 9 Mar 2011 08:58:26 +0000 (09:58 +0100)]
Rename NonSlidingAttacksBB[] in StepAttacksBB[]
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 9 Mar 2011 08:20:43 +0000 (09:20 +0100)]
Update copyright notes in rkiss.h
New info after a thread on talkchess:
http://www.talkchess.com/forum/viewtopic.php?t=38313
and some emails exchange with Heinz.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 10 Mar 2011 07:07:07 +0000 (08:07 +0100)]
Revert previous patch due to miscompile under gcc
I need to understand what's going on, in the
meantime revert.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 8 Mar 2011 13:38:17 +0000 (14:38 +0100)]
Change initialization of some bitboards
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 20 Feb 2011 10:07:13 +0000 (11:07 +0100)]
Increase MAX_THREADS to 32
No speed regression after 8731 games:
Mod vs Orig 1394 - 1342 - 5995 ELO +2 (+- 4.1)
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 7 Mar 2011 17:13:02 +0000 (18:13 +0100)]
Introduce and use rot() in rkiss.h
Also fix indentation.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Mar 2011 12:15:13 +0000 (13:15 +0100)]
Simplify aspirationDelta update rule
After 7522 games:
Mod vs Orig 1229 - 1148 - 5145 ELO +3 (+- 4.5) LOS 83%
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Mar 2011 12:32:19 +0000 (13:32 +0100)]
Be sure to read options before to call trace_evaluate()
Otherwise in case we change an option with setoption and
then ask for "eval" command the evaluation is not updated.
Spotted by Justin Blanchard.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 28 Feb 2011 19:17:57 +0000 (20:17 +0100)]
Fix aspiration corner case
Fix a corner case where we start aspiration window and
suddendly we get a VALUE_KNOWN_WIN / MATE score, this makes
aspiration to blow up in a series of researches loops.
Exit aspiration loop in that case.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 28 Feb 2011 08:00:17 +0000 (09:00 +0100)]
Remove a FIXME in id_loop()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Feb 2011 17:29:22 +0000 (18:29 +0100)]
Score root move list during first iteration
Use first iteration to get a proper startup score
and possibly detect an easy move.
After 5180 games:
Mod vs Orig 847 - 823 - 3510 ELO +1 (+- 5.5)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 27 Feb 2011 08:28:30 +0000 (09:28 +0100)]
Remove an useless condition in equal SEE pruning
Because we are never in check there and evaluation cannot
return a mated value the condition is useless.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 22 Feb 2011 22:23:45 +0000 (23:23 +0100)]
Prune moves with equal SEE in qsearch
After 5166 games:
Mod vs Orig 890 - 762 - 3514 ELO +8 (+- 5.5) LOS 96%
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Feb 2011 13:09:58 +0000 (14:09 +0100)]
Add evaluation tracing code
This patch is based on Justin Blanchard's original
work and allows to breakdown evaluation in its sub terms and
show to the user.
Tracing code has zero speed impact when not used.
Note that tracing code is not thread-safe, but this
should not be a problem given the typical usage scenario.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 21 Feb 2011 11:45:51 +0000 (12:45 +0100)]
Correctly round evaluation to grain size
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 17 Feb 2011 07:24:35 +0000 (08:24 +0100)]
Depth dependant singular extension margin
After 7965 games:
Mod vs Orig 1324 - 1249 - 5392 ELO +3 (+- 4.4) LOS 81%
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Feb 2011 20:27:22 +0000 (21:27 +0100)]
Move all enum types definitions to types.h
Cleanup headers while there.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Feb 2011 13:34:00 +0000 (14:34 +0100)]
Retire color.h
Move contents to piece.h and square.h
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Feb 2011 13:25:46 +0000 (14:25 +0100)]
Rename piece_of_color_and_type() to make_piece()
To be aligned with make_square()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Feb 2011 13:17:18 +0000 (14:17 +0100)]
Retire piece_type_from_char()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Feb 2011 11:23:16 +0000 (12:23 +0100)]
Triviality in main.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Feb 2011 09:03:30 +0000 (10:03 +0100)]
Retire enum SquareDelta
Use Square instead. At the end is the same because we were
anyway foreseen operators on mixed terms (Square, SquareDelta).
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 23 Feb 2011 08:43:47 +0000 (09:43 +0100)]
Small simplifications in square.h
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 19 Feb 2011 11:22:23 +0000 (12:22 +0100)]
Introduce and use speed_to_uci()
And retire nps()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 18 Feb 2011 16:02:25 +0000 (17:02 +0100)]
Rename SplitPoint parentSstack
Now that we don't have anymore a search stack array in
SplitPoint we can rename this data member to somthing more
usual.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 18 Feb 2011 14:44:02 +0000 (15:44 +0100)]
Retire SearchStack sstack[] from SplitPoint
Use a local variable instead. To make it work we need to
correctly init next ply search stack at the beginning of the
search because now that ss is allocated on the stack instead
of on the global storage it contains garbage.
As a side effect we can peform a fast search stack
init in id_loop().
With this patch size of SplitPoint goes from 71944 to 136 bytes,
and consequently size of Thread goes from 575568 to 1104 bytes.
Finally the size of ThreadsManager that contains all the thread
info goes from
9209248 to just 17824 bytes !!
No functional change also in faked split.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 17 Feb 2011 07:55:56 +0000 (08:55 +0100)]
Sync Root new depth to what we do in search()
This allow us to restore the old depth 12 benchmark
and fixes one and for all the depth mess.
Test confirms no regression:
After 5658 games 892 - 924 - 3842 ELO -1 (+- 5.2)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 17 Feb 2011 08:24:55 +0000 (09:24 +0100)]
Unify best move update logic
Try to rewrite the Root case using as most as common
code as possible.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 17 Feb 2011 08:08:27 +0000 (09:08 +0100)]
Do not special case reductions for MultiPV case
Note that this introduces an asymmetry in which best move
is searched deeper then others also in MultiPV, but this is
not an error per se.
No functional change when MultiPV = 1
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 16 Feb 2011 07:56:24 +0000 (08:56 +0100)]
Avoid permanent PV entries in TT
This patch removes a condition that allows a PV entry to remain
in TT across games for an unlimited time.
Although this produces a nice ELO boost in the long term it
is an artifact that affects tests results bewteen version
with and without this feature.
So remove now and readd before to release because it actually
seems a strong feature.
As example a verification tournament against SF 2.0.1 starting around
+10 ELO after 4K games sligltly climbed to +21 ELO after 14K games !!!
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 13 Feb 2011 10:13:54 +0000 (11:13 +0100)]
Write the LogFile only at the end of an iteration
Skip writing fail high/low sequences. Note that we don't need
fail high/low markers anymore in pretty_pv().
No functional change but some do/undo move sequences.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Feb 2011 18:04:27 +0000 (19:04 +0100)]
Spell checking fixes in search.cpp
Reported by Eelco on open-chess.org
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Feb 2011 17:40:16 +0000 (18:40 +0100)]
Rename iteration in depth in id_loop()
And retire the redundant one.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Feb 2011 17:09:08 +0000 (18:09 +0100)]
Start to count iterations from 1
First search should be done at iteration = 1, not 2. So offset
the variable by one.
As a nice side effect now search correctly stops at PLY_MAX
included, not after searching (PLY_MAX - 1) as before.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Feb 2011 16:39:48 +0000 (17:39 +0100)]
Fix wrong reported depth
Interestingly this patch will make people complain search depth
is reduced against 2.0.1 ;-) but actually it is only an artifact.
Spotted by Joona.
No functional change apart from a different do / undo move
sequence due to teh fact that we don't call pv_info_to_uci()
anymore before entering id loop.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 11 Feb 2011 05:20:16 +0000 (06:20 +0100)]
Move sending of PV line to id_loop()
No functional change apart form move reordering because
pv_info_to_uci() performs a do / undo_move sequence.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 10 Feb 2011 20:09:20 +0000 (21:09 +0100)]
Maximum aspiration delta of 24
After 9080 games
1430 - 1342 - 6308 ELO +3 (+- 2.9)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 5 Feb 2011 18:24:42 +0000 (20:24 +0200)]
Maximum aspiration delta of 64
After 9242 games
Mod vs Orig: 1483 - 1373 - 6386 ELO +4 (+- 2.9)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>