Hi rnk, hansw, thakis,
Implement _umul128; it provides the high and low halves of a 128-bit
multiply. We can simply use our __int128 arithmetic to implement this,
we generate great code for it:
movq %rdx, %rax
mulq %rcx
movq %rdx, (%r8)
retq
http://reviews.llvm.org/D6486
Files:
lib/Headers/Intrin.h
Index: lib/Headers/Intrin.h
===================================================================
--- lib/Headers/Intrin.h
+++ lib/Headers/Intrin.h
@@ -417,9 +417,14 @@
#endif
unsigned __int64 _shlx_u64(unsigned __int64, unsigned int);
unsigned __int64 shrx_u64(unsigned __int64, unsigned int);
-unsigned __int64 _umul128(unsigned __int64 _Multiplier,
- unsigned __int64 _Multiplicand,
- unsigned __int64 *_HighProduct);
+static __inline__ unsigned __int64 __attribute__((__always_inline__, __nodebug__))
+_umul128(unsigned __int64 _Multiplier, unsigned __int64 _Multiplicand,
+ unsigned __int64 *_HighProduct) {
+ __int128 FullProduct =
+ (unsigned __int128)_Multiplier * (unsigned __int128)_Multiplicand;
+ *_HighProduct = FullProduct >> 64;
+ return FullProduct;
+}
void __cdecl _xrstor64(void const *, unsigned __int64);
void __cdecl _xsave64(void *, unsigned __int64);
void __cdecl _xsaveopt64(void *, unsigned __int64);
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
