Double prefetch on Windows

[stockfish] / src / tt.cpp

diff --git a/src/tt.cpp b/src/tt.cpp
index 862842f2639069b20a28593487207b74c755e884..ffb8b0124568e58101a913af300b60bf2dc195ce 100644 (file)
--- a/src/tt.cpp
+++ b/src/tt.cpp
@@ -34,7 +34,7 @@
 #endif
 
 // This is the number of TTEntry slots for each position
-static const int ClusterSize = 5;
+static const int ClusterSize = 4;
 
 // The main transposition table
 TranspositionTable TT;
@@ -174,14 +174,19 @@ TTEntry* TranspositionTable::retrieve(const Key posKey) const {
 /// blocking function and do not stalls the CPU waiting for data
 /// to be loaded from RAM, that can be very slow. When we will
 /// subsequently call retrieve() the TT data will be already
-/// quickly accessible in L1/l2 CPU cache.
+/// quickly accessible in L1/L2 CPU cache.
 
 void TranspositionTable::prefetch(const Key posKey) const {
 
 #if defined(_MSC_VER)
-  _mm_prefetch((char*)first_entry(posKey), _MM_HINT_T0);
+   char* addr = (char*)first_entry(posKey);
+  _mm_prefetch(addr, _MM_HINT_T0);
+  _mm_prefetch(addr+64, _MM_HINT_T0);
 #else
-  __builtin_prefetch((const void*)first_entry(posKey), 0, 3);
+  // We need to force an asm volatile here because gcc builtin
+  // is optimized away by Intel compiler.
+  char* addr = (char*)first_entry(posKey);
+  asm volatile("prefetcht0 %0" :: "m" (addr));
 #endif
 }