-
Notifications
You must be signed in to change notification settings - Fork 14.3k
Hlsl asuint16 function #132315
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Hlsl asuint16 function #132315
Conversation
@llvm/pr-subscribers-clang @llvm/pr-subscribers-backend-spir-v Author: None (metkarpoonam) ChangesImplemented the asuint16 function and added test cases for codegen, Sema, and SPIR-V backend. Full diff: https://github.com/llvm/llvm-project/pull/132315.diff 4 Files Affected:
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index a48a8e998a015..70cbed851f0f0 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -80,6 +80,34 @@ void asuint(double3, out uint3, out uint3);
_HLSL_BUILTIN_ALIAS(__builtin_hlsl_elementwise_splitdouble)
void asuint(double4, out uint4, out uint4);
+//===----------------------------------------------------------------------===//
+// asuint16 builtins
+//===----------------------------------------------------------------------===//
+
+/// \fn uint16_t asuint16(T X)
+/// \brief Interprets the bit pattern of \a X as an 16-bit unsigned integer.
+/// \param X The input value.
+#ifdef __HLSL_ENABLE_16_BIT
+
+template <typename T, int N>
+constexpr __detail::enable_if_t<__detail::is_same<int16_t, T>::value ||
+ __detail::is_same<uint16_t, T>::value ||
+ __detail::is_same<half, T>::value,
+ vector<uint16_t, N>>
+asuint16(vector<T, N> V) {
+ return __detail::bit_cast<uint16_t, T, N>(V);
+}
+
+template <typename T>
+constexpr __detail::enable_if_t<__detail::is_same<int16_t, T>::value ||
+ __detail::is_same<uint16_t, T>::value ||
+ __detail::is_same<half, T>::value,
+ uint16_t>
+asuint16(T F) {
+ return __detail::bit_cast<uint16_t, T>(F);
+}
+#endif
+
//===----------------------------------------------------------------------===//
// distance builtins
//===----------------------------------------------------------------------===//
diff --git a/clang/test/CodeGenHLSL/builtins/asuint16.hlsl b/clang/test/CodeGenHLSL/builtins/asuint16.hlsl
new file mode 100644
index 0000000000000..0efda5f9331a9
--- /dev/null
+++ b/clang/test/CodeGenHLSL/builtins/asuint16.hlsl
@@ -0,0 +1,60 @@
+// RUN: %clang_cc1 -finclude-default-header -x hlsl -triple dxil-pc-shadermodel6.2-library %s -fnative-half-type -emit-llvm -O1 -o - | FileCheck %s
+
+//CHECK-LABEL: define {{.*}}test_ints
+//CHECK-SAME: {{.*}}(i16 {{.*}} [[VAL:%.*]]){{.*}}
+// CHECK-NOT: bitcast
+//CHECK: entry:
+// CHECK: ret i16 [[VAL]]
+uint16_t test_int(int16_t p0)
+{
+ return asuint16(p0);
+}
+
+//CHECK-LABEL: define {{.*}}test_uint
+//CHECK-SAME: {{.*}}(i16 {{.*}} [[VAL:%.*]]){{.*}}
+//CHECK-NOT: bitcast
+//CHECK: entry:
+//CHECK-NEXT: ret i16 [[VAL]]
+uint16_t test_uint(uint16_t p0)
+{
+ return asuint16(p0);
+}
+
+//CHECK-LABEL: define {{.*}}test_half
+//CHECK-SAME: {{.*}}(half {{.*}} [[VAL:%.*]]){{.*}}
+//CHECK: [[RES:%.*]] = bitcast half [[VAL]] to i16
+//CHECK-NEXT: ret i16 [[RES]]
+uint16_t test_half(half p0)
+{
+ return asuint16(p0);
+}
+
+//CHECK-LABEL: define {{.*}}test_vector_int
+//CHECK-SAME: {{.*}}(<4 x i16> {{.*}} [[VAL:%.*]]){{.*}}
+//CHECK-NOT: bitcast
+//CHECK: entry:
+//CHECK-NEXT: ret <4 x i16> [[VAL]]
+uint16_t4 test_vector_int(int16_t4 p0)
+{
+ return asuint16(p0);
+}
+
+//CHECK-LABEL: define {{.*}}test_vector_uint
+//CHECK-SAME: {{.*}}(<4 x i16> {{.*}} [[VAL:%.*]]){{.*}}
+//CHECK-NOT: bitcast
+//CHECK: entry:
+//CHECK-NEXT: ret <4 x i16> [[VAL]]
+uint16_t4 test_vector_uint(uint16_t4 p0)
+{
+ return asuint16(p0);
+}
+
+//CHECK-LABEL: define {{.*}}fn
+//CHECK-SAME: {{.*}}(<4 x half> {{.*}} [[VAL:%.*]]){{.*}}
+//CHECK: [[RES:%.*]] = bitcast <4 x half> [[VAL]] to <4 x i16>
+//CHECK-NEXT: ret <4 x i16> [[RES]]
+uint16_t4 fn(half4 p1)
+{
+ return asuint16(p1);
+}
+
diff --git a/clang/test/SemaHLSL/BuiltIns/asuint16-errors.hlsl b/clang/test/SemaHLSL/BuiltIns/asuint16-errors.hlsl
new file mode 100644
index 0000000000000..364eb828d81f4
--- /dev/null
+++ b/clang/test/SemaHLSL/BuiltIns/asuint16-errors.hlsl
@@ -0,0 +1,51 @@
+// RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.2-library %s -fnative-half-type -verify
+
+uint16_t test_asuint16_less_argument()
+{
+ return asuint16();
+ // expected-error@-1 {{no matching function for call to 'asuint16'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate function template not viable: requires single argument 'V', but no arguments were provided}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate function template not viable: requires single argument 'F', but no arguments were provided}}
+
+}
+
+int16_t4 test_asuint16_too_many_arg(uint16_t p0, uint16_t p1)
+{
+ return asuint16(p0, p1);
+ // expected-error@-1 {{no matching function for call to 'asuint16'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate function template not viable: requires single argument 'V', but 2 arguments were provided}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate function template not viable: requires single argument 'F', but 2 arguments were provided}}
+
+}
+
+int16_t test_asuint16_int(int p1)
+{
+ return asuint16(p1);
+ // expected-error@-1 {{no matching function for call to 'asuint16'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: could not match 'vector<T, N>' against 'int'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: substitution failure [with T = int]: no type named 'Type'}}
+}
+
+int16_t test_asuint16_float(float p1)
+{
+ return asuint16(p1);
+ // expected-error@-1 {{no matching function for call to 'asuint16'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: could not match 'vector<T, N>' against 'float'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: substitution failure [with T = float]: no type named 'Type'}}
+}
+
+int16_t4 test_asuint16_vector_int(int4 p1)
+{
+ return asuint16(p1);
+ // expected-error@-1 {{no matching function for call to 'asuint16'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: substitution failure [with T = int, N = 4]: no type named 'Type'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: substitution failure [with T = int4]: no type named 'Type'}}
+}
+
+int16_t4 test_asuint16_vector_float(float4 p1)
+{
+ return asuint16(p1);
+ // expected-error@-1 {{no matching function for call to 'asuint16'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: substitution failure [with T = float, N = 4]: no type named 'Type'}}
+ // expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: substitution failure [with T = float4]: no type named 'Type'}}
+}
diff --git a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/asuint16.ll b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/asuint16.ll
new file mode 100644
index 0000000000000..3c9e6733c1805
--- /dev/null
+++ b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/asuint16.ll
@@ -0,0 +1,37 @@
+; RUN: llc -O0 -verify-machineinstrs -mtriple=spirv-unknown-unknown %s -o - | FileCheck %s
+; RUN: %if spirv-tools %{ llc -O0 -mtriple=spirv-unknown-unknown %s -o - -filetype=obj | spirv-val %}
+
+; CHECK-DAG: %[[#half:]] = OpTypeFloat 16
+; CHECK-DAG: %[[#uint_16:]] = OpTypeInt 16 0
+; CHECK-DAG: %[[#v4uint_16:]] = OpTypeVector %[[#uint_16]] 4
+; CHECK-DAG: %[[#v4half:]] = OpTypeVector %[[#half]] 4
+
+define i16 @test_half(half nofpclass(nan inf) %p0) {
+entry:
+ %0 = bitcast half %p0 to i16
+ ret i16 %0
+
+ ; CHECK: %[[#arg0:]] = OpFunctionParameter %[[#half]]
+ ; CHECK: %[[#bit_cast:]] = OpBitcast %[[#uint_16]] %[[#arg0]]
+ ; CHECK: OpReturnValue %[[#bit_cast]]
+}
+
+define noundef <4 x i16> @test_vector_half(<4 x half> nofpclass(nan inf) %p1) {
+entry:
+ %0 = bitcast <4 x half> %p1 to <4 x i16>
+ ret <4 x i16> %0
+
+ ; CHECK: %[[#arg0:]] = OpFunctionParameter %[[#v4half]]
+ ; CHECK: %[[#bit_cast:]] = OpBitcast %[[#v4uint_16]] %[[#arg0]]
+ ; CHECK: OpReturnValue %[[#bit_cast]]
+}
+
+attributes #0 = { alwaysinline mustprogress nofree norecurse nosync nounwind willreturn memory(none) "approx-func-fp-math"="true" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
+
+!llvm.module.flags = !{!0}
+!dx.valver = !{!1}
+!llvm.ident = !{!2}
+
+!0 = !{i32 1, !"wchar_size", i32 4}
+!1 = !{i32 1, i32 8}
+!2 = !{!"clang version 21.0.0git (https://github.com/llvm/llvm-project.git 8037234c865d98219b54a70d9d63aee1d1f2be1c)"}
|
; RUN: llc -O0 -verify-machineinstrs -mtriple=spirv-unknown-unknown %s -o - | FileCheck %s | ||
; RUN: %if spirv-tools %{ llc -O0 -mtriple=spirv-unknown-unknown %s -o - -filetype=obj | spirv-val %} | ||
|
||
; CHECK-DAG: %[[#half:]] = OpTypeFloat 16 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
delete this file. the tests you have in the asint16 pr are sufficient to for testing OpBitcast
works for both scalars and vectors.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've removed the asuint16.ll as suggested.
|
||
//CHECK-LABEL: define {{.*}}test_ints | ||
//CHECK-SAME: {{.*}}(i16 {{.*}} [[VAL:%.*]]){{.*}} | ||
// CHECK-NOT: bitcast |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
all the check lines in the file should line up remove space here and line 7
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've removed the spaces as suggested.
#ifdef __HLSL_ENABLE_16_BIT | ||
|
||
template <typename T, int N> | ||
constexpr __detail::enable_if_t<__detail::is_same<int16_t, T>::value || |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
you need to add the Shader Model 6.2 availability macro to these two functions.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Added suggested shader model macro.
✅ With the latest revision this PR passed the C/C++ code formatter. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM but I'll let Farzon approve
// expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: could not match 'vector<T, N>' against 'int'}} | ||
// expected-note@hlsl/hlsl_intrinsics.h:* {{candidate template ignored: substitution failure [with T = int]: no type named 'Type'}} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Probably not an actionable item unless @farzonl has any ideas
The notes here feel quite unhelpful from a user perspective. I'm guessing the error is that int
doesn't fit in int16_t
but if I was a user and I saw this error then I'd be extremely confusing by what's actually wrong. That said the actual error at least points at the function and tells you it's the wrong overload so maybe not
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Since this is a language overload resolution rule and not one we defined in sema there is not much we can do in this PR. Atleast nothing that woudn't be considered scope creep. Worth talk with Chris about though and see if there is something we can change.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Implemented the asuint16 function and added test cases for codegen, Sema, and SPIR-V backend.
fixes #99185