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
execute an implied ucinewgame upon entering the UCI::loop,
to make sure that searches starting with and without an (optional) ucinewgame
command yield the same search.
This is needed now that seach::clear() initializes tables to non-zero default values.
No functional change
Closes#1101Closes#1104
Testing the release candidate revealed only one minor issue, namely a new warning -Wimplicit-fallthrough (part of -Wextra) triggers in the movepicker. This can be silenced by adding a comment, and once we move to c++17 by adding a standard annotation [[fallthrough]];.
No functional change.
Closes#1090
ss->killers can change while the movepicker is active.
The reason ss->killers changes is related to the singular
extension search in the moves loop that calls search<>
recursively with ss instead of ss+1,
effectively using the same stack entry for caller and callee.
By making a copy of the killers,
the movepicker does the right thing nevertheless.
Tested as a bug fix
STC:
http://tests.stockfishchess.org/tests/view/58ff130f0ebc59035df33f37
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 70845 W: 12752 L: 12716 D: 45377
LTC:
http://tests.stockfishchess.org/tests/view/58ff48000ebc59035df33f3d
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 28368 W: 3730 L: 3619 D: 21019
Bench: 6465887
Closes#1085
history related scores are not related to evaluation based scores.
For example, can easily exceed the range -VALUE_INFINITE,VALUE_INFINITE.
As such the current type is confusing, and a plain int is a better match.
tested for no regression:
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 43693 W: 7909 L: 7827 D: 27957
No functional change.
Closes#1070
the order of elements returned by std::partition is implementation defined (since not stable) and could depend on the version of libstdc++ linked.
As std::stable_partition was tested to be too slow (http://tests.stockfishchess.org/tests/view/585cdfd00ebc5903140c6082).
Instead combine partition with our custom implementation of insert_sort, which fixes this issue.
Implementation based on a patch by mstembera (http://tests.stockfishchess.org/tests/view/58d4d3460ebc59035df3315c), which suggests some benefit by itself.
Higher depth moves are all sorted (INT_MIN version), as in current master.
STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 33116 W: 6161 L: 6061 D: 20894
LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 88703 W: 11572 L: 11540 D: 65591
Bench: 6256522
Closes#1058Closes#1065
a single Xeon Phi can present itself as a single numa node with up to 288 threads (4 threads per hardware core).
Tested to work as expected with a Xeon Phi CPU 7230 up to 256 threads.
No functional change
Closes#1045
Simplifies away all associated checks, leading to a ~0.5% speedup.
The code now explicitly checks if moves are OK, rather than using nullptr checks.
Verified for no regression:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 32218 W: 5762 L: 5660 D: 20796
No functional change
Closes#1021
Replaces the HalfDensity array with an equivalent, compact implementation.
Includes suggestions by mcostalba & snicolet.
No functional change
Closes#1004
Fixes failing build for
make ARCH=x86-32 clean && make ARCH=x86-32 optimize=no build
by passing -m32 also to the link step.
Extend travis testing accordingly.
No functional change.
Closes#999
Perform more complex verification and validation.
- signature.sh : extract and optionally compare Bench/Signature/Node count.
- perft.sh : verify perft counts for a number of positions.
- instrumented.sh : run a few commands or uci sequences through valgrind/sanitizer instrumented binaries.
- reprosearch.sh : verify reproducibility of search.
These script can be used from directly from the command line in the src directory.
Update travis script to use these shell scripts.
No functional change.
Fixes the only exception, in razoring.
The code already does assert(PvNode || (alpha == beta - 1)), and it can be verified by studying the program flow that this is indeed the case, also for the modified line.
No functional change.
make skipEarlyPruning a search argument instead of managing this by hand.
Verified for no regression at STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 96754 W: 17089 L: 17095 D: 62570
No functional change.
* Refactor bonus and penalty calculation
Compute common terms in a helper function.
No functional change.
* Further refactoring
Remove some parenthesis that are now useless.
Define prevSq once, use repeatedly.
No functional change.
bench: 5884767 (bench of previous patch is wrong)
This refines the profile-build target to avoid 'touch'ing the sources,
keeping meaningful modification dates and avoiding editor warnings like vi's:
WARNING: The file has been changed since reading it!!!
Do you really want to write to it (y/n)?
Instead of touching sources, the (instrumented) object files are removed,
which has the same effect of rebuilding them in the next step.
As a side effect, this simplifies the Makefile a bit.
No functional change.
Makes the actual number of nodes searched match closely
the number of nodes requested, by increasing the frequency
of checking the number of nodes searched at low node count.
All other searches retain the default checking frequency of
once per 4096 nodes, and are thus unaffected.
Passed STC as non-regression
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 26643 W: 4766 L: 4655 D: 17222
No functional change.
Casting a pointer to a different type with stricter alignment
requirements yields to implementation dependent behaviour.
Practicaly everything is fine for common platforms because the
CPU/OS/compiler will generate correct code, but anyhow it is
better to be safe than sorry.
Testing with dbg_hit_on() shows that the unalignment accesses are
very rare (below 0.1%) so it makes sense to split the code in a
fast path for the common case and a slower path as a fallback.
No functional change (verified with TB enabled).
Small fixes for compilation with sanitize=yes optimize=no,
by always adding -fsanitize=undefined to the LDFLAGS as required.
Updates config-sanity to check&report the status of the flag.
No functional change.
Avoid shifting negative signed integers and use typed
enum to avoids decrementing a variable beyond its defined
range, like:
for (Rank r = RANK_8; r >= RANK_1; --r)
Changes were tested individually and passed SPRT[-3, 1].
With this patch gcc --sanitize builds cleanly.
No functional change.
For a default bench, this fixes the last valgrind
error (jump on uninitialised value).
Passed STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 187869 W: 33303 L: 33463 D: 121103
No functional change.
