================
@@ -162,6 +162,62 @@ __gpu_shuffle_idx_u64(uint64_t __lane_mask, uint32_t
__idx, uint64_t __x,
((uint64_t)__gpu_shuffle_idx_u32(__lane_mask, __idx, __lo, __width));
}
+// Returns a bitmask marking all lanes that have the same value of __x.
+_DEFAULT_FN_ATTRS static __inline__ uint64_t
----------------
jhuber6 wrote:
Honestly I got confused myself whether or not the return value was a number or
a mask. I think that having a `typedef uint64_t __lanemask_t` would be helpful
for keeping that unambiguous, but I think that it's best just to keep it
`uint64_t` for all targets to simplify code and keep it portable. I might do
that in a follow-up patch.
https://github.com/llvm/llvm-project/pull/127504
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
