7659157f1b
(x + y) >> 1 gets changed to: (x>>1) + (y>>1) + (x&y&1) Saves us having to do any llvm assembly and overflow checking in the addition. Reviewed-by: Tom Stellard <thomas.stellard@amd.com> llvm-svn: 188476
7 lines
332 B
C++
7 lines
332 B
C++
//hadd = (x+y)>>1
|
|
//This can be simplified to x>>1 + y>>1 + (1 if both x and y have the 1s bit set)
|
|
//This saves us having to do any checks for overflow in the addition sum
|
|
_CLC_OVERLOAD _CLC_DEF __CLC_GENTYPE hadd(__CLC_GENTYPE x, __CLC_GENTYPE y) {
|
|
return (x>>(__CLC_GENTYPE)1)+(y>>(__CLC_GENTYPE)1)+(x&y&(__CLC_GENTYPE)1);
|
|
}
|