Use compiler intrinsic instead of assembly for popcnt

This time, do not break compatibility with some AMD machines that have SSE3 and popcnt, but do
not have SSE4.2.

No functional change.

diff --git a/src/Makefile b/src/Makefile
index 187e7b313404fb5df608a62eba6d38cd81b546fc..1e5a16580b188df89614665027b8681453082d51 100644 (file)
--- a/src/Makefile
+++ b/src/Makefile
@@ -285,7 +285,7 @@ endif
 
 ### 3.9 popcnt
 ifeq ($(popcnt),yes)
 
 ### 3.9 popcnt
 ifeq ($(popcnt),yes)

-       CXXFLAGS += -msse3 -DUSE_POPCNT
+       CXXFLAGS += -msse3 -mpopcnt -DUSE_POPCNT

 endif
 
 ### 3.10 pext
 endif
 
 ### 3.10 pext

diff --git a/src/bitcount.h b/src/bitcount.h
index f84c51cb8596caf76e658d126330c898dafde7f8..9feed19f9755a94be0022f3e9cf5c891972589d0 100644 (file)
--- a/src/bitcount.h
+++ b/src/bitcount.h
@@ -96,8 +96,7 @@ inline int popcount<CNT_HW_POPCNT>(Bitboard b) {
 
 #else
 
 
 #else
 

-  __asm__("popcnt %1, %0" : "=r" (b) : "r" (b));
-  return b;
+  return __builtin_popcountll(b);

 
 #endif
 }
 
 #endif
 }