[HLSL][DXIL] Implementation of round intrinsic

[farzonl]

hlsl_intrinsics.h - add the round api
DXIL.td add the llvm intrinsic to DXIL lowering mapping
This change reuses llvm's existing intrinsic __builtin_elementwise_round\ int_round
This change implements: #70077

[llvmbot]

@llvm/pr-subscribers-backend-x86
@llvm/pr-subscribers-hlsl

@llvm/pr-subscribers-backend-directx

Author: Farzon Lotfi (farzonl)

Changes

hlsl_intrinsics.h - add the round api
DXIL.td add the llvm intrinsic to DXIL lowering mapping
This chane reuses llvms existing __builtin_elementwise_round`int_round`
This change implements: #70077

---

Full diff: https://github.com/llvm/llvm-project/pull/83570.diff

5 Files Affected:

• (modified) clang/lib/Headers/hlsl/hlsl_intrinsics.h (+34)
• (added) clang/test/CodeGenHLSL/builtins/round.hlsl (+53)
• (added) clang/test/SemaHLSL/BuiltIns/round-errors.hlsl (+27)
• (modified) llvm/lib/Target/DirectX/DXIL.td (+3)
• (added) llvm/test/CodeGen/DirectX/round.ll (+43)

diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 0aa8651ba80dc4..312304e13979ff 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -831,6 +831,40 @@ uint64_t3 reversebits(uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_elementwise_bitreverse)
 uint64_t4 reversebits(uint64_t4);
 
+//===----------------------------------------------------------------------===//
+// frac builtins
+//===----------------------------------------------------------------------===//
+
+/// \fn T round(T x)
+/// \brief Rounds the specified value to the nearest integer. Halfway cases are
+/// rounded to the nearest even. \a x parameter. \param x The specified input
+/// value.
+///
+/// The return value is the \a x parameter, rounded to the nearest integer
+/// within a floating-point type.
+
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+half round(half);
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+half2 round(half2);
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+half3 round(half3);
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+half4 round(half4);
+
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+float round(float);
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+float2 round(float2);
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+float3 round(float3);
+_HLSL_BUILTIN_ALIAS(__builtin_elementwise_round)
+float4 round(float4);
+
 //===----------------------------------------------------------------------===//
 // sin builtins
 //===----------------------------------------------------------------------===//
diff --git a/clang/test/CodeGenHLSL/builtins/round.hlsl b/clang/test/CodeGenHLSL/builtins/round.hlsl
new file mode 100644
index 00000000000000..b9f35bd3712d18
--- /dev/null
+++ b/clang/test/CodeGenHLSL/builtins/round.hlsl
@@ -0,0 +1,53 @@
+// RUN: %clang_cc1 -finclude-default-header -x hlsl -triple \
+// RUN:   dxil-pc-shadermodel6.3-library %s -fnative-half-type \
+// RUN:   -emit-llvm -disable-llvm-passes -o - | FileCheck %s \ 
+// RUN:   --check-prefixes=CHECK,NATIVE_HALF
+// RUN: %clang_cc1 -finclude-default-header -x hlsl -triple \
+// RUN:   dxil-pc-shadermodel6.3-library %s -emit-llvm -disable-llvm-passes \
+// RUN:   -o - | FileCheck %s --check-prefixes=CHECK,NO_HALF
+
+// NATIVE_HALF: define noundef half @
+// NATIVE_HALF: %elt.round = call half @llvm.round.f16(
+// NATIVE_HALF: ret half %elt.round
+// NO_HALF: define noundef float @"?test_round_half@@YA$halff@$halff@@Z"(
+// NO_HALF: %elt.round = call float @llvm.round.f32(
+// NO_HALF: ret float %elt.round
+half test_round_half(half p0) { return round(p0); }
+// NATIVE_HALF: define noundef <2 x half> @
+// NATIVE_HALF: %elt.round = call <2 x half> @llvm.round.v2f16
+// NATIVE_HALF: ret <2 x half> %elt.round
+// NO_HALF: define noundef <2 x float> @
+// NO_HALF: %elt.round = call <2 x float> @llvm.round.v2f32(
+// NO_HALF: ret <2 x float> %elt.round
+half2 test_round_half2(half2 p0) { return round(p0); }
+// NATIVE_HALF: define noundef <3 x half> @
+// NATIVE_HALF: %elt.round = call <3 x half> @llvm.round.v3f16
+// NATIVE_HALF: ret <3 x half> %elt.round
+// NO_HALF: define noundef <3 x float> @
+// NO_HALF: %elt.round = call <3 x float> @llvm.round.v3f32(
+// NO_HALF: ret <3 x float> %elt.round
+half3 test_round_half3(half3 p0) { return round(p0); }
+// NATIVE_HALF: define noundef <4 x half> @
+// NATIVE_HALF: %elt.round = call <4 x half> @llvm.round.v4f16
+// NATIVE_HALF: ret <4 x half> %elt.round
+// NO_HALF: define noundef <4 x float> @
+// NO_HALF: %elt.round = call <4 x float> @llvm.round.v4f32(
+// NO_HALF: ret <4 x float> %elt.round
+half4 test_round_half4(half4 p0) { return round(p0); }
+
+// CHECK: define noundef float @
+// CHECK: %elt.round = call float @llvm.round.f32(
+// CHECK: ret float %elt.round
+float test_round_float(float p0) { return round(p0); }
+// CHECK: define noundef <2 x float> @
+// CHECK: %elt.round = call <2 x float> @llvm.round.v2f32
+// CHECK: ret <2 x float> %elt.round
+float2 test_round_float2(float2 p0) { return round(p0); }
+// CHECK: define noundef <3 x float> @
+// CHECK: %elt.round = call <3 x float> @llvm.round.v3f32
+// CHECK: ret <3 x float> %elt.round
+float3 test_round_float3(float3 p0) { return round(p0); }
+// CHECK: define noundef <4 x float> @
+// CHECK: %elt.round = call <4 x float> @llvm.round.v4f32
+// CHECK: ret <4 x float> %elt.round
+float4 test_round_float4(float4 p0) { return round(p0); }
diff --git a/clang/test/SemaHLSL/BuiltIns/round-errors.hlsl b/clang/test/SemaHLSL/BuiltIns/round-errors.hlsl
new file mode 100644
index 00000000000000..fed4573063acb5
--- /dev/null
+++ b/clang/test/SemaHLSL/BuiltIns/round-errors.hlsl
@@ -0,0 +1,27 @@
+
+// RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm -disable-llvm-passes -verify -verify-ignore-unexpected
+
+float test_too_few_arg() {
+  return __builtin_elementwise_round();
+  // expected-error@-1 {{too few arguments to function call, expected 1, have 0}}
+}
+
+float2 test_too_many_arg(float2 p0) {
+  return __builtin_elementwise_round(p0, p0);
+  // expected-error@-1 {{too many arguments to function call, expected 1, have 2}}
+}
+
+float builtin_bool_to_float_type_promotion(bool p1) {
+  return __builtin_elementwise_round(p1);
+  // expected-error@-1 {{1st argument must be a vector, integer or floating point type (was 'bool')}}
+}
+
+float builtin_round_int_to_float_promotion(int p1) {
+  return __builtin_elementwise_round(p1);
+  // expected-error@-1 {{1st argument must be a floating point type (was 'int')}}
+}
+
+float2 builtin_round_int2_to_float2_promotion(int2 p1) {
+  return __builtin_elementwise_round(p1);
+  // expected-error@-1 {{1st argument must be a floating point type (was 'int2' (aka 'vector<int, 2>'))}}
+}
diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 67ef7986622092..1d4633490d8597 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -220,6 +220,9 @@ def Sin  : DXILOpMapping<13, unary, int_sin,
                          "Returns sine(theta) for theta in radians.">;
 def UMax : DXILOpMapping<39, binary, int_umax,
                          "Unsigned integer maximum. UMax(a,b) = a > b ? a : b">;
+def Round : DXILOpMapping<26, unary, int_round,
+                         "Returns the input rounded to the nearest integer"
+                         "within a floating-point type.">;
 def ThreadId : DXILOpMapping<93, threadId, int_dx_thread_id,
                              "Reads the thread ID">;
 def GroupId  : DXILOpMapping<94, groupId, int_dx_group_id,
diff --git a/llvm/test/CodeGen/DirectX/round.ll b/llvm/test/CodeGen/DirectX/round.ll
new file mode 100644
index 00000000000000..9f1fb92c4afe4a
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/round.ll
@@ -0,0 +1,43 @@
+; RUN: opt -S -dxil-op-lower < %s | FileCheck %s
+
+; Make sure dxil operation function calls for round are generated for float and half.
+; CHECK:call float @dx.op.unary.f32(i32 26, float %{{.*}})
+; CHECK:call half @dx.op.unary.f16(i32 26, half %{{.*}})
+
+target datalayout = "e-m:e-p:32:32-i1:32-i8:8-i16:16-i32:32-i64:64-f16:16-f32:32-f64:64-n8:16:32:64"
+target triple = "dxil-pc-shadermodel6.7-library"
+
+; Function Attrs: noinline nounwind optnone
+define noundef float @_Z3foof(float noundef %a) #0 {
+entry:
+  %a.addr = alloca float, align 4
+  store float %a, ptr %a.addr, align 4
+  %0 = load float, ptr %a.addr, align 4
+  %elt.round = call float @llvm.round.f32(float %0)
+  ret float %elt.round
+}
+
+; Function Attrs: nocallback nofree nosync nounwind readnone speculatable willreturn
+declare float @llvm.round.f32(float) #1
+
+; Function Attrs: noinline nounwind optnone
+define noundef half @_Z3barDh(half noundef %a) #0 {
+entry:
+  %a.addr = alloca half, align 2
+  store half %a, ptr %a.addr, align 2
+  %0 = load half, ptr %a.addr, align 2
+  %elt.round = call half @llvm.round.f16(half %0)
+  ret half %elt.round
+}
+
+; Function Attrs: nocallback nofree nosync nounwind readnone speculatable willreturn
+declare half @llvm.round.f16(half) #1
+
+attributes #0 = { noinline nounwind optnone "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
+attributes #1 = { nocallback nofree nosync nounwind readnone speculatable willreturn }
+
+!llvm.module.flags = !{!0}
+!llvm.ident = !{!1}
+
+!0 = !{i32 1, !"wchar_size", i32 4}
+!1 = !{!"clang version 15.0.0 (https://github.com/llvm/llvm-project.git 73417c517644db5c419c85c0b3cb6750172fcab5)"}
\ No newline at end of file

[bharadwajy]

A few simple suggestions. Otherwise looks good.

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.

[damyanp]

This chane reuses llvms existing intrinsic __builtin_elementwise_round\ int_round

I suspect a typo - chane -> change?