HCha/Stockfish
1
0
Fork 0
mirror of https://github.com/HChaZZY/Stockfish.git synced 2025-12-20 09:06:45 +08:00
Code Issues Packages Projects Releases Wiki Activity

Use atomics instead of volatile

Browse Source 
Rely on well defined behaviour for message passing, instead of volatile. Three
versions have been tested, to make sure this wouldn't cause a slowdown on any
platform.

v1: Sequentially consistent atomics

No mesurable regression, despite the extra memory barriers on x86. Even with 15
threads and extreme time pressure, both acting as a magnifying glass:

threads=15, tc=2+0.02
ELO: 2.59 +-3.4 (95%) LOS: 93.3%
Total: 18132 W: 4113 L: 3978 D: 10041

threads=7, tc=2+0.02
ELO: -1.64 +-3.6 (95%) LOS: 18.8%
Total: 16914 W: 4053 L: 4133 D: 8728

v2: Acquire/Release semantics

This version generates no extra barriers for x86 (on the hot path). As expected,
no regression either, under the same conditions:

threads=15, tc=2+0.02
ELO: 2.85 +-3.3 (95%) LOS: 95.4%
Total: 19661 W: 4640 L: 4479 D: 10542

threads=7, tc=2+0.02
ELO: 0.23 +-3.5 (95%) LOS: 55.1%
Total: 18108 W: 4326 L: 4314 D: 9468

As suggested by Joona, another test at LTC:

threads=15, tc=20+0.05
ELO: 0.64 +-2.6 (95%) LOS: 68.3%
Total: 20000 W: 3053 L: 3016 D: 13931

v3: Final version: SeqCst/Relaxed

threads=15, tc=10+0.1
ELO: 0.87 +-3.9 (95%) LOS: 67.1%
Total: 9541 W: 1478 L: 1454 D: 6609

Resolves #474
This commit is contained in:
lucasart
2015-10-24 22:50:51 +01:00
committed by Joona Kiiski
parent 3428a28751
commit 00d9e9fd28
4 changed files with 22 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
src/search.cpp
View File

@@ -37,7 +37,7 @@
namespace Search {
volatile SignalsType Signals;
SignalsType Signals;
LimitsType Limits;
StateStackPtr SetupStates;
}
@@ -581,8 +581,8 @@ namespace {
if (!RootNode)
{
// Step 2. Check for aborted search and immediate draw
if (Signals.stop || pos.is_draw() || ss->ply >= MAX_PLY)
return ss->ply >= MAX_PLY && !inCheck ? evaluate(pos)
if (Signals.stop.load(std::memory_order_relaxed) || pos.is_draw() || ss->ply >= MAX_PLY)
return ss->ply >= MAX_PLY && !inCheck ? evaluate(pos)
: DrawValue[pos.side_to_move()];
// Step 3. Mate distance pruning. Even if we mate at the next move our score
@@ -841,7 +841,7 @@ moves_loop: // When in check search starts from here
if (RootNode && thisThread == Threads.main())
{
Signals.firstRootMove = (moveCount == 1);
Signals.firstRootMove = moveCount == 1;
if (Time.elapsed() > 3000)
sync_cout << "info depth " << depth / ONE_PLY
@@ -1008,7 +1008,7 @@ moves_loop: // When in check search starts from here
// Finished searching the move. If a stop occurred, the return value of
// the search cannot be trusted, and we return immediately without
// updating best move, PV and TT.
if (Signals.stop)
if (Signals.stop.load(std::memory_order_relaxed))
return VALUE_ZERO;
if (RootNode)
@@ -1577,7 +1577,7 @@ void check_time() {
{
bool stillAtFirstMove = Signals.firstRootMove
&& !Signals.failedLowAtRoot
&& elapsed > Time.available() * 75 / 100;
&& elapsed > Time.available() * 3 / 4;
if ( stillAtFirstMove
|| elapsed > Time.maximum() - 2 * TimerThread::Resolution)