================
@@ -738,6 +739,16 @@ def UMin : DXILOp<40, binary> {
let attributes = [Attributes<DXIL1_0, [ReadNone]>];
}
+def UAddc : DXILOp<44, binaryWithCarryOrBorrow > {
+ let Doc = "Unsigned 32-bit integer arithmetic add with carry. uaddc(a,b) =
(a+b, a+b overflowed ? 1 : 0)";
+ let intrinsics = [IntrinSelect<int_uadd_with_overflow>];
----------------
Icohedron wrote:
Whatever does the lowering for `int_uadd_with_overflow` generically would still
need to know that the 32-bit `int_uadd_with_overflow` maps directly to the
`UAddc` DXIL op in order to codegen optimally, as opposed to creating some
expansion that does the same thing using more instructions.
https://github.com/llvm/llvm-project/pull/127137
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
