Change updating rule after a TT hit to match
the same one at the end of the search.
Small change in functionality, but we want to
have uniform rules in the code.
bench: 7767864
We avoid to use an ad-hoc table at the cost of a
relative_rank() call in advanced_pawn_push().
On my 32 bit system it is even slightly faster (on 64bit
may be different). This is the speed in nps alternating
old and new bench runs:
new
368890
368825
369972
old
367798
367635
368026
No functional change.
To align to same named Position function and
avoid using std::cout directly.
Also remove some stale <iostream> include while
there.
No functional change.
It seems to intorduce a regression when tested
with 3 threads at 15+0.05:
ELO: -2.26 +-2.2 (95%) LOS: 2.4%
Total: 30000 W: 4813 L: 5008 D: 20179
bench: 8331357
SMP case is very tricky and raises an assert in stage_moves():
assert(stage == KILLERS_S1 || stage == QUIETS_1_S1 || stage == QUIETS_2_S1)
So rewrite the code to just return moves[] when we are sure
we are in quiet moves stages.
Also rename stage_moves to quiet_moves to reflect that.
No functional change (but needs testing in SMP case)
Use MovePicker moves[] to access already tried
quiet moves. A bit of care shall be taken
to avoid calling stage_moves() when we are still
at ttMove stage, because moves are yet to be
generated. Actually our staging move generation
makes this code a bit more tricky than what I'd
like, but removing an ausiliary redundant
array like quietsSearched[] is a good thing.
Idea by DiscoCheck
bench: 9355734
Use the newly introduced LineBB[] to simplify this
super hot-path function.
Verified with perft we don't have any speed regression, although
the number of squares removed is less than before in case of
contact check.
Insipred by DiscoCheck implementation.
Perft numbers are the same, but we have an harmless functional
change due to reorder of moves, because now some illegal moves
are no more detected at generation time, but in the search.
bench: 8331357
As, Gary (that analyzed the bug) says:
SF does not print a PV when the original best move fails low,
we hit our time allowance, and stop the search. The output from
the SF search is below. It was failing low on Ne1 at depth 34.
Then, we get bestmove Qd3, but no PV change.
info depth 34 seldepth 45 score cp 38 upperbound nodes 483484489 nps 15464575 time 31264 multipv 1 pv f3e1 h5h4 e1d3 h4g3 f2g3 a6f6 f1f6 e7f6 d1a4 f6e7 a1f1 d8f8 a4b3 b7b6 b3c2 f7f6 c2a4 h3g5 b2b3 g5f7 a4c6 f7d6 h1g2 f6f5 e4f5 d6f5
info depth 34 seldepth 45 score cp 38 upperbound nodes 483484489 nps 15464575 time 31264 multipv 1 pv f3e1 h5h4 e1d3 h4g3 f2g3 a6f6 f1f6 e7f6 d1a4 f6e7 a1f1 d8f8 a4b3 b7b6 b3c2 f7f6 c2a4 h3g5 b2b3 g5f7 a4c6 f7d6 h1g2 f6f5 e4f5 d6f5
info depth 34 seldepth 47 score cp 30 upperbound nodes 2112334132 nps 17255517 time 122415 multipv 1 pv f3e1 h5h4 d1a4 a6f6 e1d3 d8f8 a4c2 h4g3 f2g3 f6f1 a1f1 h7g8 b2b3 f7f6 a3a4 b7b6
info depth 34 seldepth 47 score cp 30 upperbound nodes 2112334132 nps 17255517 time 122415 multipv 1 pv f3e1 h5h4 d1a4 a6f6 e1d3 d8f8 a4c2 h4g3 f2g3 f6f1 a1f1 h7g8 b2b3 f7f6 a3a4 b7b6
info nodes 18235667001 time 969824
bestmove e2d3 ponder c8d7
Looking at the code, if we hit Signals.stop, we return from id_loop
before printing any PV. It is possible for us to have resorted the
RootMove list though, which will change the move that is actually
played.
No functional change.
And remove a complex (and broken) formula.
Indeed previous code was broken in case of TC with big
time increments where available_time() was too similar
to total time yielding to many time losses, so for instance:
go wtime 2600 winc 2600
info nodes 4432770 time 2601 <-- time forfeit!
maximum search time = 2530 ms
available_time = 2300 ms
For a reference and further details see:
https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/dCPAvQDcm2E
Speed tested with bench disabling timer alltogheter vs timer set at
max resolution, showed we have no speed regressions both in single
core and when using all physical cores.
No functional change.
A combo of two patches that failed SPRT with score
higher than 50% but togheter they succeed:
SPRT at 60+0.05
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 7312 W: 1276 L: 1139 D: 4897
bench: 8029334
Dumb down a bit the code and trade some possible
speed (but this is far from hot path anyhow) for
some added readability for the layman.
No functional change.
This is the first chain bonus version
from Ralph that also passed both
Short TC:
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 23460 W: 4727 L: 4556 D: 14177
And long TC:
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 31858 W: 5497 L: 5240 D: 21121
And performed better against current
committed version, always at 60secs:
LLR: -2.94 (-2.94,2.94) [-3.00,3.00]
Total: 26301 W: 4477 L: 4580 D: 17244
This test was done by Leonid.
bench: 8455956
After 40K games at 60 secs, result is still
not clear, but not a regression against SF 4
After
ELO: 50.11 +-2.1 (95%) LOS: 100.0%
Total: 40000 W: 10547 L: 4817 D: 24636
Before
ELO: 49.51 +-2.1 (95%) LOS: 100.0%
Total: 40000 W: 10483 L: 4821 D: 24696
So re-apply the patch to avoid to
special-case this one.
bench: 7403882
It was never updated !
Currently it only affects evaluate_passed_pawns()
and in particularly the rule to increase the bonus
if we have more non-pawn pieces. We could simply use
popcount() instead and avoid the little slowdown
in put_piece() and remove_piece(), but this would
leave a very subtle and tricky hole where people
are forced to remember that pos.count<ALL_PIECES>()
does not work. This is not obvious and so dangerous.
Thanks to Ronald de Man for spotting this.
bench: 7931424
Due to a strange issue (bug?) the ternary
operator does not return a BitCountType for
icc, so revert to the expression.
The same patch was already applied in
9749f1f14c
Thanks to NssY Wanyonyi for pointing out
this.
No functional change.
This reverts commit 4bc2374450 for
two reasons.
First regression testing shows almost equal
score:
Before the patch:
ELO: 49.75 +-2.5 (95%) LOS: 100.0%
Total: 27205 W: 7113 L: 3244 D: 16848
After the patch:
ELO: 48.87 +-2.9 (95%) LOS: 100.0%
Total: 20860 W: 5478 L: 2563 D: 12819
Second, and more sensible to me, this patch
increases safe check bonuses to 4 times their
original value (!) and considering:
- Values were already well tuned
- Values are highly critical
- King safety is highly critical, very TC
dependent and very difficult to test
- Our testing coverage is partial (self-testing,
blitz times)
I think is better to be safe than sorry and so
I revert the patch.
bench: 8440524
Don't need a struct here. Speed test shows
result is teh same. Moreover RKISS is used
mainly at startup to compute magics, so
prefer to keep it simple...RKISS ;-)
Also some assorted triviality while there.
No functional change.
Increase bench default depth from 12 to 13 and
add 15 new endgame positions to have broader
coverage and also more reliable nps calulcation
used for fishtest framework.
Due to the new endgame positions, where nps is higher,
the total nps is increased of about 15%.
Thanks to Lucas and Jörg for the suggestions.
No functional change, but bench number is now:
bench: 8336338
For some users -stack_size,0x4000 does not work,
so revert for now.
osX 10.6.8
gcc version 4.7.3 (MacPorts gcc47 4.7.3_2)
g++: error: unrecognized command line option '-stack_size,0x4000'
make[2]: *** [stockfish] Error 1
make[1]: *** [gcc-profile-make] Error 2
make: *** [profile-build] Error 2
No functional change.
This reverts commit 800410eef1 and instead increases
stack size.
I went through the old emails with Daylen that reported the
crash issue on Mac OS X and was fixed by 0049d3f337.
It was reported default stack size for a thread in Mac OS X is 8
megabytes while the patch that we are reverting allows to reduce
stack size at max of about 217KB, so the reason for the crash was
only marginal in MAX_MOVES value. On those emails Daylen also
hinted how to increase stack size for Mac OS X to 16MB.
So prefer to increase stack size to 16MB instad of re-inventing
the wheel and do our home grown stack as we did with the patch
that we are now reverting (it will remain anyhow in git history
for documentation purposes).
No functional change.
This should be enough for any legal position, even
the handcrafted ones, like the one presented by Reuven:
1Q5R/4Q1K1/B1Q5/B4Q2/N2Q4/pQ4Q1/pn2Q3/krQ4R w - -
Where currently we crash. This reverts the patch
0049d3f337 of 8/4/2012 where stack
was shrinked due to crashes while in deep analysys.
No functional change.
This greately reduces stack usage and is a
prerequisite for next patch.
Verified with 40K games both in single and SMP
case that there are no regressions.
No functional change.
This is an even safer setup proposed and tested
by Alexandre Meirelles.
Regression testing of 40K games at 10+0.05 show
result is stable both against current master:
ELO: -0.29 +-2.2 (95%) LOS: 39.7%
Total: 40000 W: 8010 L: 8043 D: 23947
and again original master (the one with smallest
time parameters):
ELO: 1.71 +-2.2 (95%) LOS: 93.8%
Total: 40000 W: 8325 L: 8128 D: 23547
Alexandre verified with LittleBlitzer time losses are
greately reduced with this setup:
Games Completed = 2100 of 3000 (Avg game length = 35.745 sec)
Settings = RR/128MB/15000ms+50ms/M 1000cp for 12 moves, D 150 moves/
Time = 39200 sec elapsed, 16800 sec remaining
1. Stockfish 190913 1091.5/2100 803-720-577 (L: m=313 t=1 i=0 a=406) (D: r=278 i=91 f=136 s=8 a=64) (tpm=212.5 d=14.75 nps=925427)
2. Houdini 2.0 w32 1008.5/2100 720-803-577 (L: m=250 t=299 i=0 a=254) (D: r=278 i=91 f=136 s=8 a=64) (tpm=204.1 d=12.04 nps=1326351)
No functional change.
Goes in the direction of avoiding time losses and seems
equivalent after almost 40K games at super fast TC of 10+0.05
ELO: 2.61 +-2.2 (95%) LOS: 99.1%
Total: 39869 W: 8258 L: 7959 D: 23652
No functional change.
Goes in the direction of avoiding time losses and seems
equivalent after almost 40K games at super fast TC of 10+0.05
ELO: 2.41 +-2.3 (95%) LOS: 98.1%
Total: 37222 W: 7843 L: 7585 D: 21794
No functional change.
Instead of current code, give a bonus according to the frontmost
square among candidate + passed pawns.
This is a big simplification that removes a lot of accurate code
substituting it with a statistically based one using the common
'bonus' scheme, leaving to the search to sort out the details.
Results are equivalent but code is much less and, as an added bonus,
we now store candidates bitboard in pawns hash and allow this
info to be used in evaluation. This paves the way to possible
candidate pawns evaluations together with all the other pieces,
as we do for passed.
Patch passed short TC
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 16927 W: 3462 L: 3308 D: 10157
Then failed (quite quickly) at long TC
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 8451 W: 1386 L: 1448 D: 5617
But when ran with a conclusive 40K fixed games at 60 secs it proved
almost equivalent to original one.
ELO: 1.08 +-2.0 (95%) LOS: 85.8%
Total: 40000 W: 6739 L: 6615 D: 26646
bench: 3884003
ENABLE_OPERATORS_ON has incorrect definitions of
post-increment and post-decrement operators.
In particularly the returned value is the variable
already incremented/decremented, while instead they
should return the variable _before_ inc/dec.
This has no real effect because are only used in loops
and where the returned value is never used, neverthless
it is wrong. The fix would be to copy the variable to a
dummy, then inc/dec the variable, then return the dummy.
So instead, rename to pre-increment that can be implemented
without the dummy, actually the current implementation
it is already the correct pre-increment, with the only change
to return a reference (an l-value) and not a copy, so
to properly mimic the pre-increment on native integers.
Spotted by Kojirion.
No functional change.
We always attempt to keep at least this emergencyBaseTime
at clock. But if available time is very low it means that
we will force ourself to play immediately to satisfy the
emergencyBaseTime constrain and so leading to blunders.
Patch is good at short and very short TC (15secs and 5secs respectively)
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 26590 W: 5426 L: 5245 D: 15919
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 5767 W: 1397 L: 1268 D: 3102
Instead seems has no influence at longer TC (60 secs)
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 79862 W: 13623 L: 13339 D: 52900
So it is committed to have a broader testing but is
to be consider still EXPERIMENTAL and can be reverted
easily.
No functional change.
In case we have less then 10ms to think as soon as
we wake up the timer, it immediately fires and calls
check_time() where due to condition:
elapsed > TimeMgr.maximum_time() - 2 * TimerResolution
the stop flag is set and search returns immediately, without
actually search anything.
Here the somewhat hacky fix is to start the timer after
at least one iteration as been completed.
No functional change.
The possible maximum mobility cardinality (plus one in case of
zero squares available) is:
- Knights: max. 8 squares -> max. 9 entries
- Bishops: max. 13 squares -> max. 14 entries
- Rooks: max. 14 squares -> max. 15 entries
- Queen: max. 27 squares -> max. 28 entries
So remove the extra entries in the table.
Spotted by Dariusz Orzechowski.
No functional change.
Idea is sound but implementation is partial. Ryan and Joona noticed that
we leave an hole in material table. Also we got another report by an user
of an odd behaviour. Namely, if you start stockfish and from the prompt
give 'bench' you get 3453941, then if you run again bench you get 3453940.
The reason is that two different positions with the same number of pieces,
but one with a bishop pair and another without have the same material key.
But after Eelco patch also different material imbalance and this yields
to this issue.
Restesting at long TC shows the patch does not really contribute at
ELO improvement. Actually patch failed at long TC.
LLR: -2.97 (-2.94,2.94)
Total: 23109 W: 4104 L: 4092 D: 14913
So revert.
bench: 3453945
This was thought to be a draw but the bishops generally win. However,
it takes up to 66 moves. The position in the diagram was thought to be
a draw for over one hundred years, but tablebases show that White wins
in 45 moves. All of the long wins go through this type of semi-fortress
position. It takes several moves to force Black out of the temporary
fortress in the corner; then precise play with the bishops prevents Black
from forming the temporary fortress in another corner (Nunn 1995:265ff).
Before computer analysis, Speelman listed this position as unresolved,
but "probably a draw" (Speelman 1981:109).
bench: 3453945
STANDALONE-TOOLCHAIN.html in Android NDK says:
It is recommended to use the -mthumb compiler flag to force the generation
of 16-bit Thumb-1 instructions (the default being 32-bit ARM ones).
If you want to target the 'armeabi-v7a' ABI, you will need ensure that the
following two flags are being used:
CFLAGS='-march=armv7-a -mfloat-abi=softfp'
Note: The first flag enables Thumb-2 instructions, and the second one
enables H/W FPU instructions while ensuring that floating-point
parameters are passed in core registers, which is critical for
ABI compatibility. Do *not* use these flags separately!
Thanks to Peter Osterlund for pointout this doc and for showing me
an example Makefile to follow.
No functional change.
Becuase castle is coded as "king captures the rook"
the to_sq(move), A1/8 or H1/8 is empty after the move,
leading to assert assert(p != NO_PIECE) in color_of().
Teach allows() asserts about castle and fix the crash.
Bug reported by Ryan Takker and tracked down by Tom Vijlbrief.
No functional change.
The case of two lone kings on the board is already considered
by the "No pawns" scaling factor rules in material.cpp as is
KBK and KNK.
Moreover we had a small leak in endgames map because for
KK endgame it happens white and black material keys are the
same (both equal to zero), so when adding the black endgame in
Endgames::add() we were overwriting the already exsisting
white one, leading to a memory leak found by Valgrind.
So remove the endgames althogheter and rely on scaling
to correctly set the endgames value to a draw.
No functional change.
Instead of classical flags, throw an
exception when we want to immediately halt
the search. Currently only one type
is used for both UCI stop and threads
cut off.
No functional change.
Search::RootColor is a global parameter set
before to start a search, it is not something
trace() should change.
This patch allows to add trace() calls, for
debugging, inside search itself without altering
the bench, and also ensures that the values
returned by trace() and evaluate() are fully
equivalent.
No functional change.
The rounding formula is different between
positive and negative scores due to the
GrainSize/2 term that is asymmetric.
So use truncation instead of rounding. This
guarantees that evaluation is rounded to zero
in the same way for both positive and negative
scores.
Found with position's flip
bench: 4634244
Because VALUE_NONE is 30002, it happens that
after a check the next move is never an improving
one.
After this patch bench signature is independent from
VALUE_NONE actual value.
bench: 4303194
Apply to LMR the same Eelco's idea
applied to move count pruning.
This is the result of a series of
attempts started by Thomas Kolarik.
Passed both short TC
LLR: 2.95 (-2.94, 2.94)
Total: 5675 W: 1241 L: 1117 D: 3317
And long TC:
LLR: 2.95 (-2.94, 2.94)
Total: 8748 W: 1689 L: 1539 D: 5520
bench: 4356801
Introduce ThreadBase struct that is search
agnostic and just handles low level stuff,
and derive all the other specialized classes
form here.
In particular TimerThread does not hinerits
anymore all the search related stuff from Thread.
Also some renaming while there.
Suggested by Steven Edwards
No functional change.
At thread creation start_routine() is called
and from there the virtual function idle_loop()
because we do this inside Thread c'tor, where the
virtual mechanism is disabled, it could happen that
the base class idle_loop() is called instead.
The issue happens with TimerThread and MainThread
where, at launch, start_routine calls
Thread::idle_loop instead of the derived ones.
Normally this bug is hidden because c'tor finishes
before start_routine() is actually called in the
just created execution thread, but on some platforms
and in some cases this is not guaranteed and the
engine hangs.
Reported by Ted Wong on talkchess
No functional change.
The endgame king + minor vs king is erroneusly
detected as king + minor vs king + minor
Here the fix is to detect king + minor earlier,
in particular to add these trivial cases to
endgame evaluation functions.
Spotted by Reuven Peleg
bench: 4727133
A big simplification and removing of useless code.
Finished at 50% both at short TC (with SPRT) than
at long TC at fixed number of games:
ELO: -0.14 +-3.4 (95%) LOS: 46.8%
Total: 15206 W: 2836 L: 2842 D: 9528
bench: 5059948
This reverts commit 4b3a0fdab0.
As Gary says: " It failed when I tried it at long TC previously, and only
barely passed this time. Some anecdotal evidence is that it hurts vs other
engines as well (the Lightspeed rating list showed a 16 elo drop from previous
best version - still +- 5 error bars on both, but that's still significant)"
I also agree that if we have some doubts (like in this case) it is better to
be safe than sorry.
bench: 4615572
Marco Costalba