The only change done to the Makefile to get a somewhat faster binary as
discussed in #2291 is to add -msse4 to the compile options of the bmi2 build.
Since all processors supporting bmi2 also support sse4 this can be done easily.
It is a useful step to avoid sending around custom and poorly tested builds.
The speedup isn't enough to pass [0,4] but it is roughly 1.15Elo and a LOS of 90%:
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 93009 W: 20519 L: 20316 D: 52174
Also rewrite the documentation for the user when using `make --help`, so that
the order of architectures for x86-64 has the more performant build one on top.
Closes https://github.com/official-stockfish/Stockfish/pull/2300
No functional change
This exploits the recent fractional Skill Level, and is a result from some discussion in #2221 and the older #758.
Basically, if UCI_LimitStrength is set, it will internally convert UCI_Elo to a matching fractional Skill Level.
The Elo estimate is based on games at TC 60+0.6, Hash 64Mb, 8moves_v3.pgn, rated with Ordo, anchored to goldfish1.13 (CCRL 40/4 ~2000).
Note that this is mostly about internal consistency, the anchoring to CCRL is a bit weak, e.g. within this tournament,
goldfish and sungorus only have a 200Elo difference, their rating difference on CCRL is 300Elo.
I propose that we continue to expose 'Skill Level' as an UCI option, for backwards compatibility.
The result of a tournament under those conditions are given by the following table, where the player name reflects the UCI_Elo.
# PLAYER : RATING ERROR POINTS PLAYED (%) CFS(%)
1 Elo2837 : 2792.2 50.8 536.5 711 75 100
2 Elo2745 : 2739.0 49.0 487.5 711 69 100
3 Elo2654 : 2666.4 49.2 418.0 711 59 100
4 Elo2562 : 2604.5 38.5 894.5 1383 65 100
5 Elo2471 : 2515.2 38.1 651.5 924 71 100
6 Elo2380 : 2365.9 35.4 478.5 924 52 100
7 Elo2289 : 2290.0 28.0 864.0 1596 54 100
8 sungorus1.4 : 2204.9 27.8 680.5 1596 43 60
9 Elo2197 : 2201.1 30.1 523.5 924 57 100
10 Elo2106 : 2103.8 24.5 730.5 1428 51 100
11 Elo2014 : 2030.5 30.3 377.5 756 50 98
12 goldfish1.13 : 2000.0 ---- 511.0 1428 36 100
13 Elo1923 : 1928.5 30.9 641.5 1260 51 100
14 Elo1831 : 1829.0 42.1 370.5 756 49 100
15 Elo1740 : 1738.3 42.9 277.5 756 37 100
16 Elo1649 : 1625.0 42.1 525.5 1260 42 100
17 Elo1558 : 1521.5 49.9 298.0 756 39 100
18 Elo1467 : 1471.3 51.3 246.5 756 33 100
19 Elo1375 : 1407.1 51.9 183.0 756 24 ---
It can be observed that all set Elos correspond within the error bars with the observed Ordo rating.
No functional change
The current skill levels (1-20) allow for adjusting playing strengths, but
do so in big steps (e.g. level 10 vs level 11 is a ~143 Elo jump at STC).
Since the 'Skill Level' input can already be a floating point number, this
patch uses the fractional part of the input to provide the user with
fine control, allowing for varying the playing strength essentially
continuously.
The implementation internally still uses integer skill levels (needed since they pick Depths),
but non-deterministically rounds up or down the used skill level such that the average integer
skill corresponds to the input floating point one. As expected, intermediate
(fractional) skill levels yield intermediate playing strenghts.
Tested at STC, playing level 10 against levels between 10 and 11 for 10000 games
level 10.25 ELO: 24.26 +-6.2
level 10.5 ELO: 67.51 +-6.3
level 10.75 ELO: 98.52 +-6.4
level 11 ELO: 143.65 +-6.7
http://tests.stockfishchess.org/tests/view/5cd9c6b40ebc5925cf056791http://tests.stockfishchess.org/tests/view/5cd9d22b0ebc5925cf056989http://tests.stockfishchess.org/tests/view/5cd9cf610ebc5925cf056906http://tests.stockfishchess.org/tests/view/5cd9d2490ebc5925cf05698e
No functional change.
Fixes issues #2126 and #2189 where no progress in rootDepth is made for particular fens:
8/8/3P3k/8/1p6/8/1P6/1K3n2 b - - 0 1
8/1r1rp1k1/1b1pPp2/2pP1Pp1/1pP3Pp/pP5P/P5K1/8 w - - 79 46
the cause are the shuffle extensions. Upon closer analysis, it appears that in these cases a shuffle extension is made for every node searched, and progess can not be made. This patch implements a fix, namely to limit the number of extensions relative to the number of nodes searched. The ratio employed is 1/4, which fixes the issues seen so far, but it is a heuristic, and I expect that certain positions might require an even smaller fraction.
The patch was tested as a bug fix and passed:
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 56601 W: 12633 L: 12581 D: 31387
http://tests.stockfishchess.org/tests/view/5d02b37a0ebc5925cf09f6da
LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 52042 W: 8907 L: 8837 D: 34298
http://tests.stockfishchess.org/tests/view/5d0319420ebc5925cf09fe57
Furthermore, to confirm that the shuffle extension in this form indeed still brings Elo, one more test at VLTC was performed:
VLTC:
LLR: 2.96 (-2.94,2.94) [0.00,3.50]
Total: 142022 W: 20963 L: 20435 D: 100624
http://tests.stockfishchess.org/tests/view/5d03630d0ebc5925cf0a011a
Bench: 3961247
High rootDepths, selDepths and generally searches are increasingly
common with long time control games, analysis, and improving hardware.
In this case, depths of MAX_DEPTH/MAX_PLY (128) can be reached,
and the search tree is truncated.
In principle MAX_PLY can be easily increased, except for a technicality
of storing depths in a signed 8 bit int in the TT. This patch increases
MAX_PLY by storing the depth in an unsigned 8 bit, after shifting by the
most negative depth stored in TT (DEPTH_NONE).
No regression at STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 42235 W: 9565 L: 9484 D: 23186
http://tests.stockfishchess.org/tests/view/5cdb35360ebc5925cf0595e1
Verified to reach high depths on
k1b5/1p1p4/pP1Pp3/K2pPp2/1P1p1P2/3P1P2/5P2/8 w - -
info depth 142 seldepth 154 multipv 1 score cp 537 nodes 26740713110 ...
No bench change.
only the extension part of the shuffle patch is sufficient to
pass [0,3.5] bounds at VLTC as shown by two more tests.
http://tests.stockfishchess.org/tests/view/5cc168bc0ebc5925cf02bda8
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 120684 W: 17875 L: 17400 D: 85409
http://tests.stockfishchess.org/tests/view/5cc14d510ebc5925cf02bcb5
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 68415 W: 10250 L: 9905 D: 48260
this patch proposes to simplify back to this basic and easier to
understand version. In case there is a need to run a [-3, 1] VLTC on
this one, it can be done, but it is resource intensive, and not needed
IMO.
Bench: 3388643
Only called with Squares as argument, so remove unused variants.
As this is just syntax changes, only verified bench at high depth.
No functional change.
always use the implementation of gives_check in position, no need to
hand-inline part of the implementation in search.
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 57895 W: 12632 L: 12582 D: 32681
http://tests.stockfishchess.org/tests/view/5c9eaa4b0ebc5925cfffc9e3
No functional change.
This is covered by the line just before. If we would like to protect
against the piece value of e.g. a N == B, this could be done by an
assert, no need to do this at runtime.
No functional change.
This removes the skipQuiets variable, as was done in an earlier round by
@protonspring, but fixes an oversight which led to wrong mate
announcements. Quiets can only be pruned when there is no mate score, so
set moveCountPruning at the right spot.
tested as a fix at STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 66321 W: 14690 L: 14657 D: 36974
http://tests.stockfishchess.org/tests/view/5c74f3170ebc5925cffd4b3c
and as the full patch at LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 25903 W: 4341 L: 4228 D: 17334
http://tests.stockfishchess.org/tests/view/5c7540030ebc5925cffd506f
Bench: 3292342
stopOnPonderhit is used to stop search quickly on a ponderhit. It is set by mainThread as part of its time management. However, master employs it as a signal between mainThread and the UCI thread. This is not necessary, it is sufficient for the UCI thread to signal that pondering finished, and mainThread should do its usual time-keeping job, and in this case stop immediately.
This patch implements this, removing stopOnPonderHit as an atomic variable from the ThreadPool,
and moving it as a normal variable to mainThread, reducing its scope. In MainThread::check_time() the search is stopped immediately if ponder switches to false, and the variable stopOnPonderHit is set.
Furthermore, ponder has been moved to mainThread, as the variable is only used to exchange signals between the UCI thread and mainThread.
The version has been tested locally (as fishtest doesn't support ponder):
Score of ponderSimp vs master: 2616 - 2528 - 8630 [0.503] 13774
Elo difference: 2.22 +/- 3.54
which indicates no regression.
No functional change.
The new form is likely to trigger a bit more at LTC. Given that LTC
appears to be an improvement, I think that is fine.
The change is not very invasive: it does the same as before, use
potentially less time for moves that are very stable. Most of the
time, the full bonus was given if the bonus was given, so the gradual
part {3, 4, 5} didn't matter much. Whereas previously 'stable' was
expressed as the last 80% of iterations are the same, now I use a
fixed depth (10 iterations). For TCEC style TC, it will presumably
imply some more moves that are played quicker (and thus more time
on the clock when it potentially matters). Note that 10 iterations
of stability means we've been proposing that move for 99.9% of search
time.
passed STC
http://tests.stockfishchess.org/tests/view/5c30d2290ebc596a450c055b
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 70921 W: 15403 L: 15378 D: 40140
passed LTC
http://tests.stockfishchess.org/tests/view/5c31ae240ebc596a450c1881
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17422 W: 2968 L: 2842 D: 11612
No functional change.
This addresses partially issue #1911 in that it documents in our
Readme the command that users can use to verifying the md5sum of
their downloaded tablebase files.
Additionally, a quick check of the file size (the size of each
tablebase file modulo 64 is 16 as pointed out by @syzygy1) has been
implemented at launch time in Stockfish.
Closes https://github.com/official-stockfish/Stockfish/pull/1927
and https://github.com/official-stockfish/Stockfish/issues/1911
No functional change.
The `&& (ss-1)->killers[0] ` conditions are there seemingly to protect
accessing ss-5.
This is unneeded and not so intuitive (as the killer is checked for equality
with currentMove, and that one is non-zero once we're high enough in the stack,
this protects access to ss-5). We can just extend the stack from ss-4 to ss-5,
so we can call update_continuation_histories(ss-1, ..) always in search.
This goes a bit further than #1881 and addresses a comment in #1878.
passed STC:
http://tests.stockfishchess.org/tests/view/5c1aa8d50ebc5902ba127ad0
LLR: 3.12 (-2.94,2.94) [-3.00,1.00]
Total: 53515 W: 11734 L: 11666 D: 30115
passed LTC:
http://tests.stockfishchess.org/tests/view/5c1b272c0ebc5902ba12858d
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 140176 W: 23123 L: 23192 D: 93861
Bench: 3451321
this patch fixes a rare but reproducible segfault observed playing a
multi-threaded match, it is discussed somewhat in fishcooking.
From the core file, it could be observed that the issue was in qsearch, namely:
````
ss->pv[0] = MOVE_NONE;
````
and the backtrace shows the it arrives there via razoring, called from the rootNode:
````
(gdb) bt
alpha=-19, beta=682, depth=DEPTH_ZERO) at search.cpp:1247
````
Indeed, ss->pv can indeed by a nullptr at the rootNode. However, why is the
segfault so rare ?
The reason is that the condition that guards razoring:
````
(depth < 2 * ONE_PLY && eval <= alpha - RazorMargin)
````
is almost never true, since at the root alpha for depth < 5 is -VALUE_INFINITE.
Nevertheless with the new failHigh scheme, this is not guaranteed, and rootDepth > 5,
can still result in a depth < 2 search at the rootNode. If now another thread,
via the hash, writes a new low eval to the rootPos qsearch can be entered.
Rare but not unseen... I assume that some of the crashes in fishtest recently
might be due to this.
Closes https://github.com/official-stockfish/Stockfish/pull/1860
No functional change
As noticed in the forum, a crash in extract_ponder_from_tt could result
if hash size is set before the ponder move is printed. While it is arguably
a GUI issue (but it got me on the cli), it is easy to avoid this issue.
Closes https://github.com/official-stockfish/Stockfish/pull/1856
No functional change.
Because of aggressive time management and optimistic assumptions
about move overhead, it's still very easy to get Stockfish to forfeit
on time when we hit an endgame and have Syzygy EGTB on a spinning
drive. The latency from serving a few thousand EGTB probes (~10ms each),
of which there can currently be up to 4000 outstanding before a time
check, will easily overwhelm the default Move Overhead of 30ms.
This problem was first raised by Gian-Carlo Pascutto and some solutions
and improvements were discussed in the following pull requests:
https://github.com/official-stockfish/Stockfish/pull/1471https://github.com/official-stockfish/Stockfish/pull/1623https://github.com/official-stockfish/Stockfish/pull/1783
This patch is a minimal change proposed by Marco Costalba to lower
the impact of the bug. We now force a check of the clock right after
each tablebase read.
No functional change.
The recently committed Fail-High patch (081af90805)
had a number of changes beyond adjusting the depth of search on fail high, with
some undesirable side effects.
1) Decreasing depth on PV output, confusing GUIs and players alike as described in
issue #1787. The depth printed is anyway a convention, let's consider adjustedDepth
an implementation detail, and continue to print rootDepth. Depth, nodes, time and
move quality all increase as we compute more. (fixing this output has no effect on
play).
2) Fixes go depth output (now based on rootDepth again, no effect on play), also
reported in issue #1787
3) The depth lastBestDepth is used to compute how long a move is stable, a new move
found during fail-high is incorrectly considered stable if based on adjustedDepth
instead of rootDepth (this changes time management). Reverting this passed STC
and LTC:
STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 82982 W: 17810 L: 17808 D: 47364
http://tests.stockfishchess.org/tests/view/5bd391a80ebc595e0ae1e993
LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 109083 W: 17602 L: 17619 D: 73862
http://tests.stockfishchess.org/tests/view/5bd40c820ebc595e0ae1f1fb
4) In the thread voting scheme, the rank of the fail-high thread is now artificially
low, incorrectly since the quality of the move is much better than what adjustedDepth
suggests (e.g. if it takes 10 iterations to find VALUE_KNOWN_WIN, it has very low
depth). Further evidence comes from a test that showed that the move of highest
depth is not better than that of the last PV (which is potentially of much lower
adjustedDepth).
I.e. this test http://tests.stockfishchess.org/tests/view/5bd37a120ebc595e0ae1e7c3
failed SPRT[0, 5]:
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10609 W: 2266 L: 2345 D: 5998
In a running 5+0.05 th 8 test (more than 10000 games) a positive Elo estimate is
shown (strong enough for a [-3,1], possibly not [0,4]):
http://tests.stockfishchess.org/tests/view/5bd421be0ebc595e0ae1f315
LLR: -0.13 (-2.94,2.94) [0.00,4.00]
Total: 13644 W: 2573 L: 2532 D: 8539
Elo 1.04 [-2.52,4.61] / LOS 71%
Thus, restore old behavior as a bugfix, keeping the core of the fail-high patch
idea as resolving scheme. This is non-functional for bench, but changes searches
via time management and in the threaded case.
Bench: 3556672
The patch adds a small random component (+-1) to VALUE_DRAW for the evaluation
of draw positions (mostly 3folds). This random component is not static, but
potentially different for each visit of the node (hence derived from the node
counter). The effect is that in positions with many 3fold draw lines, different
lines are followed at each iteration. This keeps the search much more dynamic,
as opposed to being locked to one particular 3fold.
An example of a position where master suffers from 3fold-blindness and this patch
solves quickly is the famous TCEC game 53:
FEN: 3r2k1/pr6/1p3q1p/5R2/3P3p/8/5RP1/3Q2K1 b - - 0 51
master doesn't see that this is a lost position (draw eval up to depth 50) as
Qf6-e6 d4-d5 (found by patch at depth 23) leads to a loss.
The 3fold-blindness is more important at longer TC, the patch was yellow STC and
LTC, but passed VLTC:
STC
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 46328 W: 10048 L: 9953 D: 26327
http://tests.stockfishchess.org/tests/view/5b9c0ca20ebc592cf275f7c7
LTC
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 54663 W: 8938 L: 8846 D: 36879
http://tests.stockfishchess.org/tests/view/5b9ca1610ebc592cf27601d3
VLTC
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 31789 W: 4512 L: 4284 D: 22993
http://tests.stockfishchess.org/tests/view/5b9d1a670ebc592cf276076d
Credit to @crossbr for pointing to this problem repeatedly, and giving the hint
that many draw lines are typical in those situations.
Bench: 4756639
Storing unconditionally the current generation and bound is equivalent to master.
Part of the condition was added as a speed optimization in #429.
Here the branch is fully eliminated.
passed STC single-threaded:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 73515 W: 16378 L: 16359 D: 40778
http://tests.stockfishchess.org/tests/view/5b2fc38c0ebc5902b2e57fd5
passed STC multi-threaded:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 63725 W: 12916 L: 12874 D: 37935
http://tests.stockfishchess.org/tests/view/5b307b8f0ebc5902b2e5895f
The multithreaded test was run after a plausible suggestion by @mstembera that the effect of this could be larger with many cores. The result seems to indicate this doesn't really matter on the 8core architecture abundantly available on fishtest.
No functional change
After a helpful suggestion from AppVeyor support staff, moving the Stockfish
execution from ps to cmd seems to work. Alternative to PR #1624 tested in PR #1637.
No functional change.
Makes sure the potential benefit of first touch does not depend on
the order of the UCI commands Threads and Hash, by reallocating the
hash if a Threads is issued. The cost is zeroing the TT once more
than needed. In case the prefered order (first Threads than Hash)
is employed, this amounts to zeroing the default sized TT (16Mb),
which is essentially instantaneous.
Follow up for https://github.com/official-stockfish/Stockfish/pull/1601
where additional data and discussion is available.
Closes https://github.com/official-stockfish/Stockfish/pull/1620
No functional change.
define Color us and use this instead of pos.side_to_move() and nmp_odd. The latter allows to clarify the nmp verification criterion.
Tested for no regression:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 76713 W: 15303 L: 15284 D: 46126
http://tests.stockfishchess.org/tests/view/5b046a0d0ebc5914abc12971
No functional change.
This patch simplifies the control flow in search(), removing an if
and a goto. A side effect of the patch is that Stockfish is now a
little bit more selective at low depths, because we allow razoring,
futility pruning and probcut pruning after a null move.
passed STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32035 W: 6523 L: 6422 D: 19090
http://tests.stockfishchess.org/tests/view/5af142ca0ebc597fb3d39bb6
passed LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 41431 W: 6187 L: 6097 D: 29147
http://tests.stockfishchess.org/tests/view/5af148770ebc597fb3d39bc1
Ideas for further work:
• Use the nodes credit opened by the patch (the increased selectivity)
to try somewhat higher razoring, futility or probcut margins at [0..4].
Bench: 4855031
Currently the make strip target is broken on mingw as the exe name is wrong (stockfish instead of stockfish.exe).
Needs some testing by mingw users (both profile-build and strip, native and cross).
No functional change.
Extends valgrind/sanitizer testing to cover syzygy code.
The script downloads 4 man syzygy as needed. The time needed for the
additional testing is small (in fact hard to see a difference compared
to the large fluctuations in testing time in travis).
Possible follow-ups:
* include more TB sensitive positions in bench.
* include the test script of recent commit "Refactor tbprobe.cpp".
* verify unchanged bench with TB (with a long run).
* make the TB part of the continuation integration tests optional.
Closes https://github.com/official-stockfish/Stockfish/pull/1518
and https://github.com/official-stockfish/Stockfish/pull/1490
No functional change.
Rewrite the MovePicker class using lambda expressions for move filtering.
Includes code style changes by @mcostalba.
Verified for speed, passed STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 43191 W: 9391 L: 9312 D: 24488
http://tests.stockfishchess.org/tests/view/5a99b9df0ebc590297cc8f04
This rewrite of MovePicker.cpp seems to trigger less random crashes on Ryzen
machines than the version in previous master (reported by Bojun Guo).
Closes https://github.com/official-stockfish/Stockfish/pull/1454
No functional change.
If search depth is less than ONE_PLY call qsearch(), no need to check the
depth condition at various call sites of search().
Passed STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 14568 W: 3011 L: 2877 D: 8680
http://tests.stockfishchess.org/tests/view/5aa846190ebc59029781015b
Also helps gcc to find some optimizations (smaller binary, some speedup).
Thanks to Aram and Stefan for identifying an oversight in an early version.
Closes https://github.com/official-stockfish/Stockfish/pull/1487
No functional change.
Avoid duplicated code after recent commit "Use evaluation trend
to adjust futility margin". We initialize the improving variable
to true in the check case, which allows to avoid redundant code
in the general case.
Tested for speed by snicolet, patch seems about 0.4% faster.
No functional change.
Note: initializing the improving variable to false in the check
case was tested as a functional change, ending yellow in both STC
and LTC. This change is not included in the commit, but it is an
interesting result that could become part of a future patch about
improving or LMR. Reference of the LTC yellow test:
http://tests.stockfishchess.org/tests/view/5aa131560ebc590297cb636e
Move the first killer move out of the capture stage, combining treatment
of first and second killer move.
passed STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 55777 W: 12367 L: 12313 D: 31097
http://tests.stockfishchess.org/tests/view/5a88617e0ebc590297cc8351
Similar to an earlier proposition of Günther Demetz, see pull request #1075.
I think it is more robust and readable than master, why hand-unroll the loop
over the killer array, and duplicate code ?
This version includes review comments from Marco Costalba.
Bench: 5227124
as discussed in issue #1349, the way pages are allocated with calloc might imply some overhead on first write.
This overhead can be large and slow down the first search after a TT resize significantly, especially for large TT.
Using an explicit clear of the TT on resize fixes this problem.
Not implemented, but possibly useful for large TT, is to do this zero-ing using all search threads. Not only would this be faster, it could also lead to a more favorable memory allocation on numa systems with a first touch policy.
No functional change.
The heuristic to avoid thread binding if less than 8 threads are requested resulted in the first 7 threads not being bound.
The branch was verified to yield a roughly 13% speedup by @CoffeeOne on the appropriate hardware and OS, and an earlier version of this patch tested well on his machine:
http://tests.stockfishchess.org/tests/view/5a3693480ebc590ccbb8be5a
ELO: 9.24 +-4.6 (95%) LOS: 100.0%
Total: 5000 W: 634 L: 501 D: 3865
To make sure all threads (including mainThread) are bound as soon as the total number exceeds 7, recreate all threads on a change of thread number.
To do this, unify Threads::init, Threads::exit and Threads::set are unified in a single Threads::set function that goes through the needed steps.
The code includes several suggestions from @joergoster.
Fixes issue #1312
No functional change
For efficiency reasons current master only allows for transposition table sizes that are N = 2^k in size, the index computation can be done efficiently as (hash % N) can be written instead as (hash & 2^k - 1). On a typical computer (with 4, 8... etc Gb of RAM), this implies roughly half the RAM is left unused in analysis.
This issue was mentioned on fishcooking by Mindbreaker:
http://tests.stockfishchess.org/tests/view/5a3587de0ebc590ccbb8be04
Recently a neat trick was proposed to map a hash into the range [0,N[ more efficiently than (hash % N) for general N, nearly as efficiently as (hash % 2^k):
https://lemire.me/blog/2016/06/27/a-fast-alternative-to-the-modulo-reduction/
namely computing (hash * N / 2^32) for 32 bit hashes. This patch implements this trick and now allows for general hash sizes. Note that for N = 2^k this just amounts to using a different subset of bits from the hash. Master will use the lower k bits, this trick will use the upper k bits (of the 32 bit hash).
There is no slowdown as measured with [-3, 1] test:
http://tests.stockfishchess.org/tests/view/5a3587de0ebc590ccbb8be04
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 128498 W: 23332 L: 23395 D: 81771
There are two (smaller) caveats:
1) the patch is implemented for a 32 bit hash (so that a 64 bit multiply can be used), this effectively limits the number of clusters that can be used to 2^32 or to 128Gb of transpostion table. That's a change in the maximum allowed TT size, which could bother those using 256Gb or more regularly.
2) Already in master, an excluded move is hashed into the position key in rather simple way, essentially only affecting the lower 16 bits of the key. This is OK in master, since bits 0-15 end up in the index, but not in the new scheme, which picks the higher bits. This is 'fixed' by shifting the excluded move a few bits up. Eventually a better hashing scheme seems wise.
Despite these two caveats, I think this is a nice improvement in usability.
Bench: 5346341
Current master can yield different staticEvals depending on the path
used to reach the position. The reason for this is that the evaluation after a
null move is always computed subtracting 2 * Eval::Tempo, while this is not
the case for lazy or specialized evals. This patch always adds tempo to evals,
which doesn't affect playing strength:
LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 59911 W: 7616 L: 7545 D: 44750
STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 104947 W: 18897 L: 18919 D: 67131
Fixes issue #1335
Bench: 5208264
If the PV leads to a draw (3-fold / 50-moves) position
and we're ahead of time, think a little longer, possibly
finding a better way.
As this is most likely effective at higher draw rates,
tried speculative LTC after a yellow STC:
STC:
http://tests.stockfishchess.org/tests/view/59eb173a0ebc590ccbb8975d
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 56095 W: 10013 L: 9902 D: 36180
elo = 0.688 +- 1.711 LOS: 78.425%
LTC:
http://tests.stockfishchess.org/tests/view/59eba1670ebc590ccbb897b4
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 59579 W: 7577 L: 7273 D: 44729
elo = 1.773 +- 1.391 LOS: 99.381%
bench: 5234652
If the search is quit before skill.pick_best is called,
skill.best_move might be MOVE_NONE.
Ensure skill.best is always assigned anyhow.
Also retire the tricky best_move() and let the underlying
semantic to be clear and explicit.
No functional change.
In light of issue #1232, a test was performed about the value of '-fno-exceptions' and a second one of the combination '-fno-exceptions -fno-rtti'. It turns out these options are can be removed without introducing slowdown.
STC for removing '-fno-exceptions'
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 13678 W: 2572 L: 2439 D: 8667
STC for removing '-fno-exceptions -fno-rtti' (current patch)
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 32557 W: 6074 L: 5973 D: 20510
No functional change.
After increasing the number of threads, the histories were not cleared,
resulting in uninitialized memory usage.
This patch fixes this by clearing threads histories in Thread c'tor as
is the idomatic way.
This fixes issue 1227
No functional change.
credit goes to @mstembera for suggesting this approach.
SEE now deals with castling, promotion and en passant in a similar way.
passed STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32902 W: 6079 L: 5979 D: 20844
passed LTC
LLR: 3.92 (-2.94,2.94) [-3.00,1.00]
Total: 110698 W: 14198 L: 14145 D: 82355
Bench: 5713905
If any thread found a 'mate in x' stop the search. Previously only
mainThread would do so. Requires the bestThread selection to be
adjusted to always prefer mate scores, even if the search depth is less.
I've tried to collect some data for this patch. On 30 cores, mate finding
seems 5-30% faster on average. It is not so easy to get numbers for this,
as the time to find a mate fluctuates significantly with multi-threaded runs,
so it is an average over 100 searches for the same position. Furthermore,
hash size and position make a difference as well.
Bench: 5965302
The only call site of Time.maximum() corrected by 10.
Do this directly in remaining().
Ponder increased Time.optimum by 25% in init(). Idem.
Delete unused includes.
No functional change.
In this rare case (e.g. go infinite on a stalemate),
just spin till ponderhit/stop comes.
The Thread::wait() is a renmant of the old YBWC
code, today with lazy SMP, threads don't need to
wait when outside of their idle loop.
No functional change.
Limits::ponder was used as a signal between uci and search threads,
but is not an atomic variable, leading to the following race as
flagged by a sanitized binary.
Expect input:
```
spawn ./stockfish
send "uci\n"
expect "uciok"
send "setoption name Ponder value true\n"
send "go wtime 4000 btime 4000\n"
expect "bestmove"
send "position startpos e2e4 d7d5\n"
send "go wtime 4000 btime 4000 ponder\n"
sleep 0.01
send "ponderhit\n"
expect "bestmove"
send "quit\n"
expect eof
```
Race:
```
WARNING: ThreadSanitizer: data race (pid=7191)
Read of size 4 at 0x0000005c2260 by thread T1:
Previous write of size 4 at 0x0000005c2260 by main thread:
Location is global 'Search::Limits' of size 88 at 0x0000005c2220 (stockfish+0x0000005c2260)
```
The reason of teh race is that ponder is not just set in UCI go()
assignment but also is signaled by an async ponderhit in uci.cpp:
else if (token == "ponderhit")
Search::Limits.ponder = 0; // Switch to normal search
The fix is to add an atomic bool to the threads structure to
signal the ponder status, letting Search::Limits to reflect just
what was passed to 'go'.
No functional change.
as a lower level routine, movepicker should not depend on the
search stack or the thread class, removing a circular dependency.
Instead of copying the search stack into the movepicker object,
as well as accessing the thread class for one of the histories,
pass the required fields explicitly to the constructor (removing
the need for thread.h and implicitly search.h in movepick.cpp).
The signature is thus longer, but more explicit:
Also some renaming of histories structures while there.
passed STC [-3,1], suggesting a small elo impact:
LLR: 3.13 (-2.94,2.94) [-3.00,1.00]
Total: 381053 W: 68071 L: 68551 D: 244431
elo = -0.438 +- 0.660 LOS: 9.7%
No functional change.
in the last month a couple of timeouts have been seen in travis valgrind testing, leading to undesired false positives. The precise cause of this is unclear: a normal valgrind instrumented run is about 6min, the timeout is 10min. Either there are rare hangs (not reproduced locally), or maybe the actual runtime fluctuates on the travis infrastructure (which uses VMs on AWS as far as I know). This patch leads to roughly a 2x speedup of the instrumented testing by reducing the depth from 10 to 9. If timeouts persist, it needs further analysis.
No functional change.
Closes#1171
Instead of having Signals in the search namespace,
make the stop variables part of the Threads structure.
This moves more of the shared (atomic) variables towards
the thread-related structures, making their role more clear.
No functional change
Closes#1149
the nodes, tbHits, rootDepth and lastInfoTime variables are read by multiple threads, but not declared atomic, leading to data races as found by -fsanitize=thread. This patch fixes this issue. It is based on top of the CI-threading branch (PR #1129), and should fix the corresponding CI error messages.
The patch passed an STC check for no regression:
http://tests.stockfishchess.org/tests/view/5925d5590ebc59035df34b9f
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 169597 W: 29938 L: 30066 D: 109593
Whereas rootDepth and lastInfoTime are not performance critical, nodes and tbHits are. Indeed, an earlier version using relaxed atomic updates on the latter two variables failed STC testing (http://tests.stockfishchess.org/tests/view/592001700ebc59035df34924), which can be shown to be due to x86-32 (http://tests.stockfishchess.org/tests/view/592330ac0ebc59035df34a89). Indeed, the latter have no instruction to atomically update a 64bit variable. The proposed solution thus uses a variable in Position that is accessed only by one thread, which is copied every few thousand nodes to the shared variable in Thread.
No functional change.
Closes#1130Closes#1129
TT.new_search() was being called by mainThread in Thread::search(). However, mainThread is the last to start searching, and helper threads could reach a measured rootDepth 10 (on 64 cores) before mainThread increments the TT generation. Fixed by moving the call to MaintThread::search() before helper threads start searching.
No functional change.
Closes#1134
Fixes a bug in Search::clear, where the filling of CounterMoveStats&, overwrote (currently presumably unused) memory because sizeof(cm) returns the size in bytes, whereas elements was needed.
No functional change
Closes#1119
