[HLSL] Implement WaveActiveAllTrue Intrinsic (#117245)

V-FEXrt · web-flow · commit 41a6e9cfd693 · 2024-12-16T16:13:35.000-08:00
Resolves #99161 - [x] Implement `WaveActiveAllTrue` clang builtin, - [x] Link `WaveActiveAllTrue` clang builtin with `hlsl_intrinsics.h` - [x] Add sema checks for `WaveActiveAllTrue` to `CheckHLSLBuiltinFunctionCall` in `SemaChecking.cpp` - [x] Add codegen for `WaveActiveAllTrue` to `EmitHLSLBuiltinExpr` in `CGBuiltin.cpp` - [x] Add codegen tests to `clang/test/CodeGenHLSL/builtins/WaveActiveAllTrue.hlsl` - [x] Add sema tests to `clang/test/SemaHLSL/BuiltIns/WaveActiveAllTrue-errors.hlsl` - [x] Create the `int_dx_WaveActiveAllTrue` intrinsic in `IntrinsicsDirectX.td` - [x] Create the `DXILOpMapping` of `int_dx_WaveActiveAllTrue` to `114` in `DXIL.td` - [x] Create the `WaveActiveAllTrue.ll` and `WaveActiveAllTrue_errors.ll` tests in `llvm/test/CodeGen/DirectX/` - [x] Create the `int_spv_WaveActiveAllTrue` intrinsic in `IntrinsicsSPIRV.td` - [x] In SPIRVInstructionSelector.cpp create the `WaveActiveAllTrue` lowering and map it to `int_spv_WaveActiveAllTrue` in `SPIRVInstructionSelector::selectIntrinsic`. - [x] Create SPIR-V backend test case in `llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAllTrue.ll`
diff --git a/clang/include/clang/Basic/Builtins.td b/clang/include/clang/Basic/Builtins.td
@@ -4762,6 +4762,12 @@ def HLSLAsDouble : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLWaveActiveAllTrue : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_all_true"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "bool(bool)";
+}
+
 def HLSLWaveActiveAnyTrue : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_active_any_true"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -19436,6 +19436,16 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
         /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
         ArrayRef<Value *>{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_all_true: {
+    Value *Op = EmitScalarExpr(E->getArg(0));
+    llvm::Type *Ty = Op->getType();
+    assert(Ty->isIntegerTy(1) &&
+           "Intrinsic WaveActiveAllTrue operand must be a bool");
+
+    Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveActiveAllTrueIntrinsic();
+    return EmitRuntimeCall(
+        Intrinsic::getOrInsertDeclaration(&CGM.getModule(), ID), {Op});
+  }
   case Builtin::BI__builtin_hlsl_wave_active_any_true: {
     Value *Op = EmitScalarExpr(E->getArg(0));
     assert(Op->getType()->isIntegerTy(1) &&
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -92,6 +92,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Dot4AddI8Packed, dot4add_i8packed)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Dot4AddU8Packed, dot4add_u8packed)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveActiveAllTrue, wave_all)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveActiveAnyTrue, wave_any)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveActiveCountBits, wave_active_countbits)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2241,6 +2241,15 @@ float4 trunc(float4);
 // Wave* builtins
 //===----------------------------------------------------------------------===//
 
+/// \brief Returns true if the expression is true in all active lanes in the
+/// current wave.
+///
+/// \param Val The boolean expression to evaluate.
+/// \return True if the expression is true in all lanes.
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_active_all_true)
+__attribute__((convergent)) bool WaveActiveAllTrue(bool Val);
+
 /// \brief Returns true if the expression is true in any active lane in the
 /// current wave.
 ///
diff --git a/clang/test/CodeGenHLSL/builtins/WaveActiveAllTrue.hlsl b/clang/test/CodeGenHLSL/builtins/WaveActiveAllTrue.hlsl
@@ -0,0 +1,17 @@
+// RUN: %clang_cc1 -finclude-default-header -fnative-half-type -triple \
+// RUN:   dxil-pc-shadermodel6.3-compute %s -emit-llvm -disable-llvm-passes -o - | \
+// RUN:   FileCheck %s --check-prefixes=CHECK,CHECK-DXIL
+// RUN: %clang_cc1 -finclude-default-header -fnative-half-type -triple \
+// RUN:   spirv-pc-vulkan-compute %s -emit-llvm -disable-llvm-passes -o - | \
+// RUN:   FileCheck %s --check-prefixes=CHECK,CHECK-SPIRV
+
+// Test basic lowering to runtime function call for int values.
+
+// CHECK-LABEL: define {{.*}}test
+bool test(bool p1) {
+  // CHECK-SPIRV: %[[#entry_tok0:]] = call token @llvm.experimental.convergence.entry()
+  // CHECK-SPIRV:  %[[RET:.*]] = call spir_func i1 @llvm.spv.wave.all(i1 %{{[a-zA-Z0-9]+}}) [ "convergencectrl"(token %[[#entry_tok0]]) ]
+  // CHECK-DXIL:  %[[RET:.*]] = call i1 @llvm.dx.wave.all(i1 %{{[a-zA-Z0-9]+}})
+  // CHECK:  ret i1 %[[RET]]
+  return WaveActiveAllTrue(p1);
+}
diff --git a/clang/test/SemaHLSL/BuiltIns/WaveActiveAllTrue-errors.hlsl b/clang/test/SemaHLSL/BuiltIns/WaveActiveAllTrue-errors.hlsl
@@ -0,0 +1,21 @@
+// RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -emit-llvm-only -disable-llvm-passes -verify
+
+bool test_too_few_arg() {
+  return __builtin_hlsl_wave_active_all_true();
+  // expected-error@-1 {{too few arguments to function call, expected 1, have 0}}
+}
+
+bool test_too_many_arg(bool p0) {
+  return __builtin_hlsl_wave_active_all_true(p0, p0);
+  // expected-error@-1 {{too many arguments to function call, expected 1, have 2}}
+}
+
+struct Foo
+{
+  int a;
+};
+
+bool test_type_check(Foo p0) {
+  return __builtin_hlsl_wave_active_all_true(p0);
+  // expected-error@-1 {{no viable conversion from 'Foo' to 'bool'}}
+}
diff --git a/llvm/include/llvm/IR/IntrinsicsDirectX.td b/llvm/include/llvm/IR/IntrinsicsDirectX.td
@@ -98,6 +98,7 @@ def int_dx_umad : DefaultAttrsIntrinsic<[llvm_anyint_ty], [LLVMMatchType<0>, LLV
 def int_dx_normalize : DefaultAttrsIntrinsic<[LLVMMatchType<0>], [llvm_anyfloat_ty], [IntrNoMem]>;
 def int_dx_rsqrt  : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>], [IntrNoMem]>;
 def int_dx_wave_active_countbits : DefaultAttrsIntrinsic<[llvm_i32_ty], [llvm_i1_ty], [IntrConvergent, IntrNoMem]>;
+def int_dx_wave_all : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent, IntrNoMem]>;
 def int_dx_wave_any : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent, IntrNoMem]>;
 def int_dx_wave_getlaneindex : DefaultAttrsIntrinsic<[llvm_i32_ty], [], [IntrConvergent, IntrNoMem]>;
 def int_dx_wave_is_first_lane : DefaultAttrsIntrinsic<[llvm_i1_ty], [], [IntrConvergent]>;
diff --git a/llvm/include/llvm/IR/IntrinsicsSPIRV.td b/llvm/include/llvm/IR/IntrinsicsSPIRV.td
@@ -87,6 +87,7 @@ let TargetPrefix = "spv" in {
   def int_spv_dot4add_i8packed : DefaultAttrsIntrinsic<[llvm_i32_ty], [llvm_i32_ty, llvm_i32_ty, llvm_i32_ty], [IntrNoMem]>;
   def int_spv_dot4add_u8packed : DefaultAttrsIntrinsic<[llvm_i32_ty], [llvm_i32_ty, llvm_i32_ty, llvm_i32_ty], [IntrNoMem]>;
   def int_spv_wave_active_countbits : DefaultAttrsIntrinsic<[llvm_i32_ty], [llvm_i1_ty], [IntrConvergent, IntrNoMem]>;
+  def int_spv_wave_all : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent, IntrNoMem]>;
   def int_spv_wave_any : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent, IntrNoMem]>;
   def int_spv_wave_is_first_lane : DefaultAttrsIntrinsic<[llvm_i1_ty], [], [IntrConvergent]>;
   def int_spv_wave_readlane : DefaultAttrsIntrinsic<[llvm_any_ty], [LLVMMatchType<0>, llvm_i32_ty], [IntrConvergent, IntrNoMem]>;
diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
@@ -924,6 +924,14 @@ def CreateHandleFromBinding : DXILOp<217, createHandleFromBinding> {
   let stages = [Stages<DXIL1_6, [all_stages]>];
 }
 
+def WaveActiveAllTrue : DXILOp<114, waveAllTrue> {
+  let Doc = "returns true if the expression is true in all of the active lanes in the current wave";
+  let intrinsics = [ IntrinSelect<int_dx_wave_all> ];
+  let arguments = [Int1Ty];
+  let result = Int1Ty;
+  let stages = [Stages<DXIL1_0, [all_stages]>];
+}
+
 def WaveActiveAnyTrue : DXILOp<113, waveAnyTrue> {
   let Doc = "returns true if the expression is true in any of the active lanes in the current wave";
   let intrinsics = [ IntrinSelect<int_dx_wave_any> ];
diff --git a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
@@ -2959,6 +2959,8 @@ bool SPIRVInstructionSelector::selectIntrinsic(Register ResVReg,
     return selectExtInst(ResVReg, ResType, I, CL::s_clamp, GL::SClamp);
   case Intrinsic::spv_wave_active_countbits:
     return selectWaveActiveCountBits(ResVReg, ResType, I);
+  case Intrinsic::spv_wave_all:
+    return selectWaveOpInst(ResVReg, ResType, I, SPIRV::OpGroupNonUniformAll);
   case Intrinsic::spv_wave_any:
     return selectWaveOpInst(ResVReg, ResType, I, SPIRV::OpGroupNonUniformAny);
   case Intrinsic::spv_wave_is_first_lane:
diff --git a/llvm/test/CodeGen/DirectX/WaveActiveAllTrue.ll b/llvm/test/CodeGen/DirectX/WaveActiveAllTrue.ll
@@ -0,0 +1,10 @@
+; RUN: opt -S -dxil-op-lower -mtriple=dxil-pc-shadermodel6.3-compute %s | FileCheck %s
+
+define noundef i1 @wave_all_simple(i1 noundef %p1) {
+entry:
+; CHECK: call i1 @dx.op.waveAllTrue(i32 114, i1 %p1)
+  %ret = call i1 @llvm.dx.wave.all(i1 %p1)
+  ret i1 %ret
+}
+
+declare i1 @llvm.dx.wave.all(i1)
diff --git a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAllTrue.ll b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAllTrue.ll
@@ -0,0 +1,21 @@
+; RUN: llc -verify-machineinstrs -O0 -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: %[[#bool:]] = OpTypeBool
+; CHECK-DAG: %[[#uint:]] = OpTypeInt 32 0
+; CHECK-DAG: %[[#scope:]] = OpConstant %[[#uint]] 3
+; CHECK-DAG: OpCapability GroupNonUniformVote
+
+; CHECK-LABEL: Begin function test_wave_all
+define i1 @test_wave_all(i1 %p1) #0 {
+entry:
+; CHECK: %[[#param:]] = OpFunctionParameter %[[#bool]]
+; CHECK: %{{.+}} = OpGroupNonUniformAll %[[#bool]] %[[#scope]] %[[#param]]
+  %0 = call token @llvm.experimental.convergence.entry()
+  %ret = call i1 @llvm.spv.wave.all(i1 %p1) [ "convergencectrl"(token %0) ]
+  ret i1 %ret
+}
+
+declare i1 @llvm.spv.wave.all(i1) #0
+
+attributes #0 = { convergent }
