[clang][hlsl] Add tan intrinsic part 1 (#90276)

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:
#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

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

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];

clang/lib/CodeGen/CGBuiltin.cpp

@@ -3822,7 +3822,9 @@ 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"));

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
 //===----------------------------------------------------------------------===//

clang/lib/Sema/SemaChecking.cpp

@@ -3047,6 +3047,7 @@ 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))
@@ -5677,6 +5678,7 @@ 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;

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(

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:

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 );
+}

clang/test/Sema/aarch64-sve-vector-trig-ops.c

@@ -1,18 +1,24 @@
-// RUN: %clang_cc1 -triple aarch64 -target-feature +sve \
-// RUN:   -disable-O0-optnone -o - -fsyntax-only %s -verify
-// REQUIRES: aarch64-registered-target
-
-#include <arm_sve.h>
-
-
-svfloat32_t test_sin_vv_i8mf8(svfloat32_t v) {
-
-  return __builtin_elementwise_sin(v);
-  // expected-error@-1 {{1st argument must be a vector, integer or floating point type}}
-}
-
-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}}
-}
+// RUN: %clang_cc1 -triple aarch64 -target-feature +sve \
+// RUN:   -disable-O0-optnone -o - -fsyntax-only %s -verify
+// REQUIRES: aarch64-registered-target
+
+#include <arm_sve.h>
+
+
+svfloat32_t test_sin_vv_i8mf8(svfloat32_t v) {
+
+  return __builtin_elementwise_sin(v);
+  // expected-error@-1 {{1st argument must be a vector, integer or floating point type}}
+}
+
+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}}
+}

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);

clang/test/Sema/riscv-rvv-vector-trig-ops.c

@@ -1,19 +1,25 @@
-// RUN: %clang_cc1 -triple riscv64 -target-feature +f -target-feature +d \
-// RUN:   -target-feature +v -target-feature +zfh -target-feature +zvfh \
-// RUN:   -disable-O0-optnone -o - -fsyntax-only %s -verify
-// REQUIRES: riscv-registered-target
-
-#include <riscv_vector.h>
-
-
-vfloat32mf2_t test_sin_vv_i8mf8(vfloat32mf2_t v) {
-
-  return __builtin_elementwise_sin(v);
-  // expected-error@-1 {{1st argument must be a vector, integer or floating point type}}
-}
-
-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}}
-}
+// RUN: %clang_cc1 -triple riscv64 -target-feature +f -target-feature +d \
+// RUN:   -target-feature +v -target-feature +zfh -target-feature +zvfh \
+// RUN:   -disable-O0-optnone -o - -fsyntax-only %s -verify
+// REQUIRES: riscv-registered-target
+
+#include <riscv_vector.h>
+
+
+vfloat32mf2_t test_sin_vv_i8mf8(vfloat32mf2_t v) {
+
+  return __builtin_elementwise_sin(v);
+  // expected-error@-1 {{1st argument must be a vector, integer or floating point type}}
+}
+
+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}}
+}

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;

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) {

llvm/docs/LangRef.rst

@@ -15272,6 +15272,43 @@ 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
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
 

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>]>;