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

Per-thread TB hit counters

Browse Source 
Use a per-thread counter to reduce contention
with many cores and endgame positions.

Measured around 1% speed-up on a 12 core and 8%
on 28 cores with 6-men, searching on:
 7R/1p3k2/2p2P2/3nR1P1/8/3b1P2/7K/r7 b - - 3 38

Also retire the unused set_nodes_searched() and fix
a couple of return types and naming conventions.

No functional change.
This commit is contained in:
syzygy
2016-10-20 21:16:09 +02:00
committed by Marco Costalba
parent 3686e719a1
commit ca67752645
4 changed files with 28 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
src/thread.cpp
View File

@@ -35,7 +35,7 @@ ThreadPool Threads; // Global object
Thread::Thread() {
resetCalls = exit = false;
maxPly = callsCnt = 0;
maxPly = callsCnt = tbHits = 0;
history.clear();
counterMoves.clear();
idx = Threads.size(); // Start from 0
@@ -158,15 +158,26 @@ void ThreadPool::read_uci_options() {
/// ThreadPool::nodes_searched() returns the number of nodes searched
int64_t ThreadPool::nodes_searched() {
uint64_t ThreadPool::nodes_searched() {
int64_t nodes = 0;
uint64_t nodes = 0;
for (Thread* th : *this)
nodes += th->rootPos.nodes_searched();
return nodes;
}
/// ThreadPool::tb_hits() returns the number of TB hits
uint64_t ThreadPool::tb_hits() {
uint64_t hits = 0;
for (Thread* th : *this)
hits += th->tbHits;
return hits;
}
/// ThreadPool::start_thinking() wakes up the main thread sleeping in idle_loop()
/// and starts a new search, then returns immediately.