================
@@ -94,3 +122,60 @@ double3 test_abs_double3(double3 p0) { return abs(p0); }
// CHECK-LABEL: define noundef nofpclass(nan inf) <4 x double>
@_Z16test_abs_double4
// CHECK: call reassoc nnan ninf nsz arcp afn <4 x double> @llvm.fabs.v4f64(
double4 test_abs_double4(double4 p0) { return abs(p0); }
+
+
+// CHECK-LABEL: define {{.*}}hlsl3abs{{.*}}(i32
+// CHECK: [[Alloca:%.*]] = alloca i32
----------------
llvm-beanz wrote:
I don't think we should be testing the implementation of `constexpr` in this
PR. That really just adds extra testing time without meaningfully adding
coverage since `constexpr` is extensively tested in the compiler elsewhere.
If this were not a constexpr-able function we should have gotten an error
elsewhere and boy would that be a big a surprise given what the function does!
https://github.com/llvm/llvm-project/pull/128257
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
