summary |
shortlog |
log |
commit | commitdiff |
tree
raw |
patch |
inline | side by side (from parent 1:
77b4f4c)
When changing 'search' and 'splitPointsSize' we have to
use thread locks, not split point ones, because can_join()
is called under the formers.
Verified succesfully with 24 hours toruture tests with 20
cores machine by Louis Zulli: it does not hangs.
Verifyed for no regressions with STC, 7 threads:
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 52804 W: 8159 L: 8087 D: 36558
No functional change.
searching = false;
activePosition = nullptr;
searching = false;
activePosition = nullptr;
+
+ spinlock.release();
+
sp->slavesMask.reset(idx);
sp->allSlavesSearching = false;
sp->nodes += pos.nodes_searched();
sp->slavesMask.reset(idx);
sp->allSlavesSearching = false;
sp->nodes += pos.nodes_searched();
// After releasing the lock we can't access any SplitPoint related data
// in a safe way because it could have been released under our feet by
// the sp master.
// After releasing the lock we can't access any SplitPoint related data
// in a safe way because it could have been released under our feet by
// the sp master.
sp->spinlock.release();
// Try to late join to another split point if none of its slaves has
sp->spinlock.release();
// Try to late join to another split point if none of its slaves has
SplitPoint& sp = splitPoints[splitPointsSize];
sp.spinlock.acquire(); // No contention here until we don't increment splitPointsSize
SplitPoint& sp = splitPoints[splitPointsSize];
sp.spinlock.acquire(); // No contention here until we don't increment splitPointsSize
sp.master = this;
sp.parentSplitPoint = activeSplitPoint;
sp.master = this;
sp.parentSplitPoint = activeSplitPoint;
++splitPointsSize;
activeSplitPoint = &sp;
activePosition = nullptr;
++splitPointsSize;
activeSplitPoint = &sp;
activePosition = nullptr;
// Try to allocate available threads
Thread* slave;
// Try to allocate available threads
Thread* slave;
Thread::idle_loop(); // Force a call to base class idle_loop()
Thread::idle_loop(); // Force a call to base class idle_loop()
- sp.spinlock.acquire();
- spinlock.acquire();
-
// In the helpful master concept, a master can help only a sub-tree of its
// split point and because everything is finished here, it's not possible
// for the master to be booked.
assert(!searching);
assert(!activePosition);
// In the helpful master concept, a master can help only a sub-tree of its
// split point and because everything is finished here, it's not possible
// for the master to be booked.
assert(!searching);
assert(!activePosition);
// We have returned from the idle loop, which means that all threads are
// finished. Note that decreasing splitPointsSize must be done under lock
// protection to avoid a race with Thread::can_join().
// We have returned from the idle loop, which means that all threads are
// finished. Note that decreasing splitPointsSize must be done under lock
// protection to avoid a race with Thread::can_join().
+ spinlock.acquire();
+
+ searching = true;
--splitPointsSize;
activeSplitPoint = sp.parentSplitPoint;
activePosition = &pos;
--splitPointsSize;
activeSplitPoint = sp.parentSplitPoint;
activePosition = &pos;
+
+ spinlock.release();
+
+ // Split point data cannot be changed now, so no need to lock protect
pos.set_nodes_searched(pos.nodes_searched() + sp.nodes);
*bestMove = sp.bestMove;
*bestValue = sp.bestValue;
pos.set_nodes_searched(pos.nodes_searched() + sp.nodes);
*bestMove = sp.bestMove;
*bestValue = sp.bestValue;
-
- spinlock.release();
- sp.spinlock.release();