llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT--> @llvm/pr-subscribers-llvm-ir @llvm/pr-subscribers-backend-x86 Author: Farzon Lotfi (farzonl) <details> <summary>Changes</summary> This change is an implementation of #<!-- -->87367's investigation on supporting IEEE math operations as intrinsics. Which was discussed in this RFC: https://discourse.llvm.org/t/rfc-all-the-math-intrinsics/78294 If you want an overarching view of how this will all connect see: https://github.com/llvm/llvm-project/pull/90088 Changes: - `clang/docs/LanguageExtensions.rst` - Document the new elementwise tan builtin. - `clang/include/clang/Basic/Builtins.td` - Implement the tan builtin. - `clang/lib/CodeGen/CGBuiltin.cpp` - invoke the tan intrinsic on uses of the builtin - `clang/lib/Headers/hlsl/hlsl_intrinsics.h` - Associate the tan builtin with the equivalent hlsl apis - `clang/lib/Sema/SemaChecking.cpp` - Add generic sema checks as well as HLSL specifc sema checks to the tan builtin - `llvm/include/llvm/IR/Intrinsics.td` - Create the tan intrinsic - `llvm/docs/LangRef.rst` - Document the tan intrinsic --- Full diff: https://github.com/llvm/llvm-project/pull/90276.diff 15 Files Affected: - (modified) clang/docs/LanguageExtensions.rst (+1) - (modified) clang/include/clang/Basic/Builtins.td (+6) - (modified) clang/lib/CodeGen/CGBuiltin.cpp (+3-1) - (modified) clang/lib/Headers/hlsl/hlsl_intrinsics.h (+23) - (modified) clang/lib/Sema/SemaChecking.cpp (+2) - (modified) clang/test/CodeGen/builtins-elementwise-math.c (+16) - (modified) clang/test/CodeGen/strictfp-elementwise-bulitins.cpp (+10) - (added) clang/test/CodeGenHLSL/builtins/tan.hlsl (+59) - (modified) clang/test/Sema/aarch64-sve-vector-trig-ops.c (+6) - (modified) clang/test/Sema/builtins-elementwise-math.c (+21) - (modified) clang/test/Sema/riscv-rvv-vector-trig-ops.c (+6) - (modified) clang/test/SemaCXX/builtins-elementwise-math.cpp (+7) - (modified) clang/test/SemaHLSL/BuiltIns/half-float-only-errors.hlsl (+1) - (modified) llvm/docs/LangRef.rst (+37) - (modified) llvm/include/llvm/IR/Intrinsics.td (+1) ``````````diff diff --git a/clang/docs/LanguageExtensions.rst b/clang/docs/LanguageExtensions.rst index f18b946efd4bfa..aa61e239d6984c 100644 --- a/clang/docs/LanguageExtensions.rst +++ b/clang/docs/LanguageExtensions.rst @@ -656,6 +656,7 @@ Unless specified otherwise operation(±0) = ±0 and operation(±infinity) = ±in T __builtin_elementwise_ceil(T x) return the smallest integral value greater than or equal to x floating point types T __builtin_elementwise_sin(T x) return the sine of x interpreted as an angle in radians floating point types T __builtin_elementwise_cos(T x) return the cosine of x interpreted as an angle in radians floating point types + T __builtin_elementwise_tan(T x) return the tangent of x interpreted as an angle in radians floating point types T __builtin_elementwise_floor(T x) return the largest integral value less than or equal to x floating point types T __builtin_elementwise_log(T x) return the natural logarithm of x floating point types T __builtin_elementwise_log2(T x) return the base 2 logarithm of x floating point types diff --git a/clang/include/clang/Basic/Builtins.td b/clang/include/clang/Basic/Builtins.td index de721a87b3341d..11982af3fa609b 100644 --- a/clang/include/clang/Basic/Builtins.td +++ b/clang/include/clang/Basic/Builtins.td @@ -1326,6 +1326,12 @@ def ElementwiseSqrt : Builtin { let Prototype = "void(...)"; } +def ElementwiseTan : Builtin { + let Spellings = ["__builtin_elementwise_tan"]; + let Attributes = [NoThrow, Const, CustomTypeChecking]; + let Prototype = "void(...)"; +} + def ElementwiseTrunc : Builtin { let Spellings = ["__builtin_elementwise_trunc"]; let Attributes = [NoThrow, Const, CustomTypeChecking]; diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp index d08ab539148914..f941e93b0ced2b 100644 --- a/clang/lib/CodeGen/CGBuiltin.cpp +++ b/clang/lib/CodeGen/CGBuiltin.cpp @@ -3821,7 +3821,9 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID, case Builtin::BI__builtin_elementwise_sin: return RValue::get( emitUnaryBuiltin(*this, E, llvm::Intrinsic::sin, "elt.sin")); - + case Builtin::BI__builtin_elementwise_tan: + return RValue::get( + emitUnaryBuiltin(*this, E, llvm::Intrinsic::tan, "elt.tan")); case Builtin::BI__builtin_elementwise_trunc: return RValue::get( emitUnaryBuiltin(*this, E, llvm::Intrinsic::trunc, "elt.trunc")); diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h b/clang/lib/Headers/hlsl/hlsl_intrinsics.h index 06409c6fc77417..3390f0962f67d0 100644 --- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h +++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h @@ -1441,6 +1441,29 @@ float3 sqrt(float3); _HLSL_BUILTIN_ALIAS(__builtin_elementwise_sqrt) float4 sqrt(float4); +//===----------------------------------------------------------------------===// +// tan builtins +//===----------------------------------------------------------------------===// +#ifdef __HLSL_ENABLE_16_BIT +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +half tan(half); +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +half2 tan(half2); +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +half3 tan(half3); +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +half4 tan(half4); +#endif + +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +float tan(float); +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +float2 tan(float2); +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +float3 tan(float3); +_HLSL_BUILTIN_ALIAS(__builtin_elementwise_tan) +float4 tan(float4); + //===----------------------------------------------------------------------===// // trunc builtins //===----------------------------------------------------------------------===// diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp index e33113ab9c4c1d..fcc901392e1161 100644 --- a/clang/lib/Sema/SemaChecking.cpp +++ b/clang/lib/Sema/SemaChecking.cpp @@ -3049,6 +3049,7 @@ Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl, unsigned BuiltinID, case Builtin::BI__builtin_elementwise_nearbyint: case Builtin::BI__builtin_elementwise_sin: case Builtin::BI__builtin_elementwise_sqrt: + case Builtin::BI__builtin_elementwise_tan: case Builtin::BI__builtin_elementwise_trunc: case Builtin::BI__builtin_elementwise_canonicalize: { if (PrepareBuiltinElementwiseMathOneArgCall(TheCall)) @@ -5664,6 +5665,7 @@ bool Sema::CheckHLSLBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) { case Builtin::BI__builtin_elementwise_roundeven: case Builtin::BI__builtin_elementwise_sin: case Builtin::BI__builtin_elementwise_sqrt: + case Builtin::BI__builtin_elementwise_tan: case Builtin::BI__builtin_elementwise_trunc: { if (CheckFloatOrHalfRepresentations(this, TheCall)) return true; diff --git a/clang/test/CodeGen/builtins-elementwise-math.c b/clang/test/CodeGen/builtins-elementwise-math.c index 1c667e5bff1eab..1b5466abd347d7 100644 --- a/clang/test/CodeGen/builtins-elementwise-math.c +++ b/clang/test/CodeGen/builtins-elementwise-math.c @@ -604,6 +604,22 @@ void test_builtin_elementwise_sqrt(float f1, float f2, double d1, double d2, vf2 = __builtin_elementwise_sqrt(vf1); } +void test_builtin_elementwise_tan(float f1, float f2, double d1, double d2, + float4 vf1, float4 vf2) { + // CHECK-LABEL: define void @test_builtin_elementwise_tan( + // CHECK: [[F1:%.+]] = load float, ptr %f1.addr, align 4 + // CHECK-NEXT: call float @llvm.tan.f32(float [[F1]]) + f2 = __builtin_elementwise_tan(f1); + + // CHECK: [[D1:%.+]] = load double, ptr %d1.addr, align 8 + // CHECK-NEXT: call double @llvm.tan.f64(double [[D1]]) + d2 = __builtin_elementwise_tan(d1); + + // CHECK: [[VF1:%.+]] = load <4 x float>, ptr %vf1.addr, align 16 + // CHECK-NEXT: call <4 x float> @llvm.tan.v4f32(<4 x float> [[VF1]]) + vf2 = __builtin_elementwise_tan(vf1); +} + void test_builtin_elementwise_trunc(float f1, float f2, double d1, double d2, float4 vf1, float4 vf2) { // CHECK-LABEL: define void @test_builtin_elementwise_trunc( diff --git a/clang/test/CodeGen/strictfp-elementwise-bulitins.cpp b/clang/test/CodeGen/strictfp-elementwise-bulitins.cpp index fdf865ebbe8911..c72d5949916911 100644 --- a/clang/test/CodeGen/strictfp-elementwise-bulitins.cpp +++ b/clang/test/CodeGen/strictfp-elementwise-bulitins.cpp @@ -187,6 +187,16 @@ float4 strict_elementwise_sqrt(float4 a) { return __builtin_elementwise_sqrt(a); } +// CHECK-LABEL: define dso_local noundef <4 x float> @_Z22strict_elementwise_tanDv4_f +// CHECK-SAME: (<4 x float> noundef [[A:%.*]]) local_unnamed_addr #[[ATTR2]] { +// CHECK-NEXT: entry: +// CHECK-NEXT: [[ELT_TAN:%.*]] = tail call <4 x float> @llvm.tan.v4f32(<4 x float> [[A]]) #[[ATTR4]] +// CHECK-NEXT: ret <4 x float> [[ELT_TAN]] +// +float4 strict_elementwise_tan(float4 a) { + return __builtin_elementwise_tan(a); +} + // CHECK-LABEL: define dso_local noundef <4 x float> @_Z24strict_elementwise_truncDv4_f // CHECK-SAME: (<4 x float> noundef [[A:%.*]]) local_unnamed_addr #[[ATTR2]] { // CHECK-NEXT: entry: diff --git a/clang/test/CodeGenHLSL/builtins/tan.hlsl b/clang/test/CodeGenHLSL/builtins/tan.hlsl new file mode 100644 index 00000000000000..aa542fac226d0b --- /dev/null +++ b/clang/test/CodeGenHLSL/builtins/tan.hlsl @@ -0,0 +1,59 @@ +// 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: spirv-unknown-vulkan-compute %s -emit-llvm -disable-llvm-passes \ +// RUN: -o - | FileCheck %s --check-prefixes=CHECK,NO_HALF + +// CHECK-LABEL: test_tan_half +// NATIVE_HALF: call half @llvm.tan.f16 +// NO_HALF: call float @llvm.tan.f32 +half test_tan_half ( half p0 ) { + return tan ( p0 ); +} + +// CHECK-LABEL: test_tan_half2 +// NATIVE_HALF: call <2 x half> @llvm.tan.v2f16 +// NO_HALF: call <2 x float> @llvm.tan.v2f32 +half2 test_tan_half2 ( half2 p0 ) { + return tan ( p0 ); +} + +// CHECK-LABEL: test_tan_half3 +// NATIVE_HALF: call <3 x half> @llvm.tan.v3f16 +// NO_HALF: call <3 x float> @llvm.tan.v3f32 +half3 test_tan_half3 ( half3 p0 ) { + return tan ( p0 ); +} + +// CHECK-LABEL: test_tan_half4 +// NATIVE_HALF: call <4 x half> @llvm.tan.v4f16 +// NO_HALF: call <4 x float> @llvm.tan.v4f32 +half4 test_tan_half4 ( half4 p0 ) { + return tan ( p0 ); +} + +// CHECK-LABEL: test_tan_float +// CHECK: call float @llvm.tan.f32 +float test_tan_float ( float p0 ) { + return tan ( p0 ); +} + +// CHECK-LABEL: test_tan_float2 +// CHECK: call <2 x float> @llvm.tan.v2f32 +float2 test_tan_float2 ( float2 p0 ) { + return tan ( p0 ); +} + +// CHECK-LABEL: test_tan_float3 +// CHECK: call <3 x float> @llvm.tan.v3f32 +float3 test_tan_float3 ( float3 p0 ) { + return tan ( p0 ); +} + +// CHECK-LABEL: test_tan_float4 +// CHECK: call <4 x float> @llvm.tan.v4f32 +float4 test_tan_float4 ( float4 p0 ) { + return tan ( p0 ); +} diff --git a/clang/test/Sema/aarch64-sve-vector-trig-ops.c b/clang/test/Sema/aarch64-sve-vector-trig-ops.c index 7ca941f578c70d..5039599d43ad81 100644 --- a/clang/test/Sema/aarch64-sve-vector-trig-ops.c +++ b/clang/test/Sema/aarch64-sve-vector-trig-ops.c @@ -16,3 +16,9 @@ svfloat32_t test_cos_vv_i8mf8(svfloat32_t v) { return __builtin_elementwise_cos(v); // expected-error@-1 {{1st argument must be a vector, integer or floating point type}} } + +svfloat32_t test_tan_vv_i8mf8(svfloat32_t v) { + + return __builtin_elementwise_tan(v); + // expected-error@-1 {{1st argument must be a vector, integer or floating point type}} +} diff --git a/clang/test/Sema/builtins-elementwise-math.c b/clang/test/Sema/builtins-elementwise-math.c index 2e05337273ee41..2e4319d158e7a0 100644 --- a/clang/test/Sema/builtins-elementwise-math.c +++ b/clang/test/Sema/builtins-elementwise-math.c @@ -626,6 +626,27 @@ void test_builtin_elementwise_sqrt(int i, float f, double d, float4 v, int3 iv, // expected-error@-1 {{1st argument must be a floating point type (was 'unsigned4' (vector of 4 'unsigned int' values))}} } +void test_builtin_elementwise_tan(int i, float f, double d, float4 v, int3 iv, unsigned u, unsigned4 uv) { + + struct Foo s = __builtin_elementwise_tan(f); + // expected-error@-1 {{initializing 'struct Foo' with an expression of incompatible type 'float'}} + + i = __builtin_elementwise_tan(); + // expected-error@-1 {{too few arguments to function call, expected 1, have 0}} + + i = __builtin_elementwise_tan(i); + // expected-error@-1 {{1st argument must be a floating point type (was 'int')}} + + i = __builtin_elementwise_tan(f, f); + // expected-error@-1 {{too many arguments to function call, expected 1, have 2}} + + u = __builtin_elementwise_tan(u); + // expected-error@-1 {{1st argument must be a floating point type (was 'unsigned int')}} + + uv = __builtin_elementwise_tan(uv); + // expected-error@-1 {{1st argument must be a floating point type (was 'unsigned4' (vector of 4 'unsigned int' values))}} +} + void test_builtin_elementwise_trunc(int i, float f, double d, float4 v, int3 iv, unsigned u, unsigned4 uv) { struct Foo s = __builtin_elementwise_trunc(f); diff --git a/clang/test/Sema/riscv-rvv-vector-trig-ops.c b/clang/test/Sema/riscv-rvv-vector-trig-ops.c index a457e484860602..ee4c596c431841 100644 --- a/clang/test/Sema/riscv-rvv-vector-trig-ops.c +++ b/clang/test/Sema/riscv-rvv-vector-trig-ops.c @@ -17,3 +17,9 @@ vfloat32mf2_t test_cos_vv_i8mf8(vfloat32mf2_t v) { return __builtin_elementwise_cos(v); // expected-error@-1 {{1st argument must be a vector, integer or floating point type}} } + +vfloat32mf2_t test_tan_vv_i8mf8(vfloat32mf2_t v) { + + return __builtin_elementwise_tan(v); + // expected-error@-1 {{1st argument must be a vector, integer or floating point type}} +} diff --git a/clang/test/SemaCXX/builtins-elementwise-math.cpp b/clang/test/SemaCXX/builtins-elementwise-math.cpp index 44a44ab055e997..499f2795ddb272 100644 --- a/clang/test/SemaCXX/builtins-elementwise-math.cpp +++ b/clang/test/SemaCXX/builtins-elementwise-math.cpp @@ -111,6 +111,13 @@ void test_builtin_elementwise_sin() { static_assert(!is_const<decltype(__builtin_elementwise_sin(b))>::value); } +void test_builtin_elementwise_tan() { + const float a = 42.0; + float b = 42.3; + static_assert(!is_const<decltype(__builtin_elementwise_tan(a))>::value); + static_assert(!is_const<decltype(__builtin_elementwise_tan(b))>::value); +} + void test_builtin_elementwise_sqrt() { const float a = 42.0; float b = 42.3; diff --git a/clang/test/SemaHLSL/BuiltIns/half-float-only-errors.hlsl b/clang/test/SemaHLSL/BuiltIns/half-float-only-errors.hlsl index ef0928f8fef0d6..4089188134d321 100644 --- a/clang/test/SemaHLSL/BuiltIns/half-float-only-errors.hlsl +++ b/clang/test/SemaHLSL/BuiltIns/half-float-only-errors.hlsl @@ -9,6 +9,7 @@ // RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only -disable-llvm-passes -verify -DTEST_FUNC=__builtin_elementwise_sin // RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only -disable-llvm-passes -verify -DTEST_FUNC=__builtin_elementwise_sqrt // RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only -disable-llvm-passes -verify -DTEST_FUNC=__builtin_elementwise_roundeven +// RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only -disable-llvm-passes -verify -DTEST_FUNC=__builtin_elementwise_tan // RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only -disable-llvm-passes -verify -DTEST_FUNC=__builtin_elementwise_trunc double2 test_double_builtin(double2 p0) { diff --git a/llvm/docs/LangRef.rst b/llvm/docs/LangRef.rst index f169ab941c457b..144733202f9e0e 100644 --- a/llvm/docs/LangRef.rst +++ b/llvm/docs/LangRef.rst @@ -15270,6 +15270,43 @@ trapping or setting ``errno``. When specified with the fast-math-flag 'afn', the result may be approximated using a less accurate calculation. +'``llvm.tan.*``' Intrinsic +^^^^^^^^^^^^^^^^^^^^^^^^^^ + +Syntax: +""""""" + +This is an overloaded intrinsic. You can use ``llvm.tan`` on any +floating-point or vector of floating-point type. Not all targets support +all types however. + +:: + + declare float @llvm.tan.f32(float %Val) + declare double @llvm.tan.f64(double %Val) + declare x86_fp80 @llvm.tan.f80(x86_fp80 %Val) + declare fp128 @llvm.tan.f128(fp128 %Val) + declare ppc_fp128 @llvm.tan.ppcf128(ppc_fp128 %Val) + +Overview: +""""""""" + +The '``llvm.tan.*``' intrinsics return the tangent of the operand. + +Arguments: +"""""""""" + +The argument and return value are floating-point numbers of the same type. + +Semantics: +"""""""""" + +Return the same value as a corresponding libm '``tan``' function but without +trapping or setting ``errno``. + +When specified with the fast-math-flag 'afn', the result may be approximated +using a less accurate calculation. + '``llvm.pow.*``' Intrinsic ^^^^^^^^^^^^^^^^^^^^^^^^^^ diff --git a/llvm/include/llvm/IR/Intrinsics.td b/llvm/include/llvm/IR/Intrinsics.td index a14e9dedef8c9e..bf0ca1687a34d4 100644 --- a/llvm/include/llvm/IR/Intrinsics.td +++ b/llvm/include/llvm/IR/Intrinsics.td @@ -1025,6 +1025,7 @@ let IntrProperties = [IntrNoMem, IntrSpeculatable, IntrWillReturn] in { def int_powi : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>, llvm_anyint_ty]>; def int_sin : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>]>; def int_cos : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>]>; + def int_tan : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>]>; def int_pow : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>, LLVMMatchType<0>]>; def int_log : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>]>; `````````` </details> https://github.com/llvm/llvm-project/pull/90276 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits