aa5ff68247
This PR fixes bugs in HLSL floating conversions. HLSL always has `half`, `float` and `double` types, which promote in the order: `half`->`float`->`double` and convert in the order: `double`->`float`->`half` As with other conversions in C++, promotions are preferred over conversions. We do have floating conversions documented in the draft language specification (https://microsoft.github.io/hlsl-specs/specs/hlsl.pdf [Conv.rank.float]) although the exact language is still in flux (https://github.com/microsoft/hlsl-specs/pull/206). Resolves #81047
64 lines
1.6 KiB
HLSL
64 lines
1.6 KiB
HLSL
// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -o - -fsyntax-only %s -verify
|
|
// XFAIL: *
|
|
|
|
// https://github.com/llvm/llvm-project/issues/81047
|
|
|
|
// expected-no-diagnostics
|
|
void Fn4(int64_t2 L);
|
|
void Fn4(int2 I);
|
|
|
|
void Call4(int16_t H) { Fn4(H); }
|
|
|
|
int test_builtin_dot_bool_type_promotion(bool p0, bool p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
float test_dot_scalar_mismatch(float p0, int p1) { return dot(p0, p1); }
|
|
|
|
float test_dot_element_type_mismatch(int2 p0, float2 p1) { return dot(p0, p1); }
|
|
|
|
float test_builtin_dot_vec_int_to_float_promotion(int2 p0, float2 p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
int64_t test_builtin_dot_vec_int_to_int64_promotion(int64_t2 p0, int2 p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
float test_builtin_dot_vec_half_to_float_promotion(float2 p0, half2 p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
float test_builtin_dot_vec_int16_to_float_promotion(float2 p0, int16_t2 p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
half test_builtin_dot_vec_int16_to_half_promotion(half2 p0, int16_t2 p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
int test_builtin_dot_vec_int16_to_int_promotion(int2 p0, int16_t2 p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
int64_t test_builtin_dot_vec_int16_to_int64_promotion(int64_t2 p0,
|
|
int16_t2 p1) {
|
|
return dot(p0, p1);
|
|
}
|
|
|
|
float4 test_frac_int4(int4 p0) { return frac(p0); }
|
|
|
|
float test_frac_int(int p0) { return frac(p0); }
|
|
|
|
float test_frac_bool(bool p0) { return frac(p0); }
|
|
|
|
// This resolves the wrong overload. In clang this converts down to an int, in
|
|
// DXC it extends the scalar to a vector.
|
|
void Fn(int) {}
|
|
void Fn(vector<int64_t,2>) {}
|
|
|
|
void Call() {
|
|
int64_t V;
|
|
Fn(V);
|
|
}
|