[clang] [llvm] [HLSL] Implement WaveActiveAnyTrue intrinsic (PR #115902)
Ashley Coleman via cfe-commits
cfe-commits at lists.llvm.org
Thu Nov 14 15:55:37 PST 2024
https://github.com/V-FEXrt updated https://github.com/llvm/llvm-project/pull/115902
>From 845256b2ed971a4e42f7f871e8b51e711486261a Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Mon, 11 Nov 2024 16:34:23 -0700
Subject: [PATCH 01/11] [HLSL] Implement WaveActiveAnyTrue intrinsic
---
clang/include/clang/Basic/Builtins.td | 6 +++++
clang/lib/CodeGen/CGBuiltin.cpp | 13 ++++++++++
clang/lib/CodeGen/CGHLSLRuntime.h | 1 +
clang/lib/Headers/hlsl/hlsl_intrinsics.h | 9 +++++++
clang/lib/Sema/SemaHLSL.cpp | 6 +++++
.../builtins/WaveActiveAnyTrue.hlsl | 17 +++++++++++++
.../BuiltIns/WaveActiveAnyTrue-errors.hlsl | 21 ++++++++++++++++
llvm/include/llvm/IR/IntrinsicsDirectX.td | 1 +
llvm/include/llvm/IR/IntrinsicsSPIRV.td | 1 +
llvm/lib/Target/DirectX/DXIL.td | 10 ++++++++
.../Target/SPIRV/SPIRVInstructionSelector.cpp | 24 +++++++++++++++++++
.../test/CodeGen/DirectX/WaveActiveAnyTrue.ll | 10 ++++++++
.../hlsl-intrinsics/WaveActiveAnyTrue.ll | 17 +++++++++++++
13 files changed, 136 insertions(+)
create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl
create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
diff --git a/clang/include/clang/Basic/Builtins.td b/clang/include/clang/Basic/Builtins.td
index 9bd67e0cefebc3..496b6739724295 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4744,6 +4744,12 @@ def HLSLAny : LangBuiltin<"HLSL_LANG"> {
let Prototype = "bool(...)";
}
+def HLSLWaveActiveAnyTrue : LangBuiltin<"HLSL_LANG"> {
+ let Spellings = ["__builtin_hlsl_wave_active_any_true"];
+ let Attributes = [NoThrow, Const];
+ let Prototype = "bool(bool)";
+}
+
def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
let Spellings = ["__builtin_hlsl_wave_active_count_bits"];
let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 65d7f5c54a1913..c4a6a9abee63cc 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18960,6 +18960,19 @@ 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_any_true: {
+ // Assert Op->getType() == Bool
+
+ // FIXME: PR Question: Can this be simpler? Looks like Int1Ty isn't predefined
+ IntegerType* Int1Ty = llvm::Type::getInt1Ty(CGM.getTypes().getLLVMContext());
+ Value *Op = EmitScalarExpr(E->getArg(0));
+ assert(Op->getType() == Int1Ty && "wave_active_any_true operand must be a bool");
+
+ // FIXME: PR Question: Re Style SingleRef vs {SingleRef} vs ArrayRef{SingleRef}
+ llvm::FunctionType *FT = llvm::FunctionType::get(Int1Ty, {Int1Ty}, /*isVarArg=*/false);
+ llvm::StringRef Name = Intrinsic::getName(CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic());
+ return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, Name, {}, /*Local=*/false, /*AssumeConvergent=*/true), {Op}, "hlsl.wave.activeanytrue");
+ }
case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
// We don't define a SPIR-V intrinsic, instead it is a SPIR-V built-in
// defined in SPIRVBuiltins.td. So instead we manually get the matching name
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h b/clang/lib/CodeGen/CGHLSLRuntime.h
index cd533cad84e9fb..c3b689ea44fcb8 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -89,6 +89,7 @@ class CGHLSLRuntime {
GENERATE_HLSL_INTRINSIC_FUNCTION(FDot, fdot)
GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
+ GENERATE_HLSL_INTRINSIC_FUNCTION(WaveActiveAnyTrue, wave_activeanytrue)
GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_readlane)
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 8ade4b27f360fb..90991e95d6565a 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2096,6 +2096,15 @@ float4 trunc(float4);
// Wave* builtins
//===----------------------------------------------------------------------===//
+/// \brief Returns true if the expression is true in any active lane in the
+/// current wave.
+///
+/// \param Val The boolean expression to evaluate.
+/// \return True if the expression is true in any lane.
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_active_any_true)
+__attribute__((convergent)) bool WaveActiveAnyTrue(bool Val);
+
/// \brief Counts the number of boolean variables which evaluate to true across
/// all active lanes in the current wave.
///
diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index a472538236e2d9..e12a4ce1e75bd6 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -2066,6 +2066,12 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
return true;
break;
}
+ case Builtin::BI__builtin_hlsl_wave_active_any_true: {
+ if (SemaRef.checkArgCount(TheCall, 1))
+ return true;
+
+ break;
+ }
case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
if (SemaRef.checkArgCount(TheCall, 2))
return true;
diff --git a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
new file mode 100644
index 00000000000000..d8233c45e0ff2d
--- /dev/null
+++ b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
@@ -0,0 +1,17 @@
+// RUN: %clang_cc1 -std=hlsl2021 -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 -std=hlsl2021 -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: 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.activeanytrue(i1 %{{[a-zA-Z0-9]+}}) [ "convergencectrl"(token %[[#entry_tok0]]) ]
+ // CHECK-DXIL: %[[RET:.*]] = call i1 @llvm.dx.wave.activeanytrue(i1 %{{[a-zA-Z0-9]+}})
+ // CHECK: ret i1 %[[RET]]
+ return WaveActiveAnyTrue(p1);
+}
diff --git a/clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl b/clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl
new file mode 100644
index 00000000000000..48fda7b571fca7
--- /dev/null
+++ b/clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-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_any_true();
+ // expected-error at -1 {{too few arguments to function call, expected 1, have 0}}
+}
+
+bool test_too_many_arg(bool p0) {
+ return __builtin_hlsl_wave_active_any_true(p0, p0);
+ // expected-error at -1 {{too many arguments to function call, expected 1, have 2}}
+}
+
+struct Foo
+{
+ int a;
+};
+
+bool test_type_check_2(Foo p0) {
+ return __builtin_hlsl_wave_active_any_true(p0);
+ // expected-error at -1 {{no viable conversion from 'Foo' to 'bool'}}
+}
diff --git a/llvm/include/llvm/IR/IntrinsicsDirectX.td b/llvm/include/llvm/IR/IntrinsicsDirectX.td
index e30d37f69f781e..7c21dc8f41f169 100644
--- a/llvm/include/llvm/IR/IntrinsicsDirectX.td
+++ b/llvm/include/llvm/IR/IntrinsicsDirectX.td
@@ -86,6 +86,7 @@ def int_dx_normalize : DefaultAttrsIntrinsic<[LLVMMatchType<0>], [llvm_anyfloat_
def int_dx_rsqrt : DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>], [IntrNoMem]>;
def int_dx_wave_getlaneindex : DefaultAttrsIntrinsic<[llvm_i32_ty], [], [IntrConvergent, IntrNoMem]>;
def int_dx_wave_is_first_lane : DefaultAttrsIntrinsic<[llvm_i1_ty], [], [IntrConvergent]>;
+def int_dx_wave_activeanytrue : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
def int_dx_wave_readlane : DefaultAttrsIntrinsic<[llvm_any_ty], [LLVMMatchType<0>, llvm_i32_ty], [IntrConvergent, IntrNoMem]>;
def int_dx_sign : DefaultAttrsIntrinsic<[LLVMScalarOrSameVectorWidth<0, llvm_i32_ty>], [llvm_any_ty], [IntrNoMem]>;
def int_dx_step : DefaultAttrsIntrinsic<[LLVMMatchType<0>], [llvm_anyfloat_ty, LLVMMatchType<0>], [IntrNoMem]>;
diff --git a/llvm/include/llvm/IR/IntrinsicsSPIRV.td b/llvm/include/llvm/IR/IntrinsicsSPIRV.td
index 6df2eb156a0774..c5e93aa4cfe01d 100644
--- a/llvm/include/llvm/IR/IntrinsicsSPIRV.td
+++ b/llvm/include/llvm/IR/IntrinsicsSPIRV.td
@@ -85,6 +85,7 @@ let TargetPrefix = "spv" in {
[IntrNoMem, Commutative] >;
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]>;
+ def int_spv_wave_activeanytrue : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
def int_spv_sign : DefaultAttrsIntrinsic<[LLVMScalarOrSameVectorWidth<0, llvm_i32_ty>], [llvm_any_ty], [IntrNoMem]>;
def int_spv_radians : DefaultAttrsIntrinsic<[LLVMMatchType<0>], [llvm_anyfloat_ty], [IntrNoMem]>;
diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 68ae5de06423c2..9657d249f9170a 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -202,6 +202,7 @@ defset list<DXILOpClass> OpClasses = {
def unpack4x8 : DXILOpClass;
def viewID : DXILOpClass;
def waveActiveAllEqual : DXILOpClass;
+ def waveActiveAnyTrue : DXILOpClass;
def waveActiveBallot : DXILOpClass;
def waveActiveBit : DXILOpClass;
def waveActiveOp : DXILOpClass;
@@ -803,6 +804,15 @@ def CreateHandleFromBinding : DXILOp<218, createHandleFromBinding> {
let stages = [Stages<DXIL1_6, [all_stages]>];
}
+def WaveActiveAnyTrue : DXILOp<113, waveActiveAnyTrue> {
+ let Doc = "returns true if the expression is true in any of the active lanes in the current wave";
+ let LLVMIntrinsic = int_dx_wave_activeanytrue;
+ let arguments = [Int1Ty];
+ let result = Int1Ty;
+ let stages = [Stages<DXIL1_0, [all_stages]>];
+ let attributes = [Attributes<DXIL1_0, [ReadNone]>];
+}
+
def WaveIsFirstLane : DXILOp<110, waveIsFirstLane> {
let Doc = "returns 1 for the first lane in the wave";
let LLVMIntrinsic = int_dx_wave_is_first_lane;
diff --git a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
index d9377fe4b91a1a..8d4faedd1c6210 100644
--- a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
+++ b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
@@ -230,6 +230,9 @@ class SPIRVInstructionSelector : public InstructionSelector {
bool selectSpvThreadId(Register ResVReg, const SPIRVType *ResType,
MachineInstr &I) const;
+ bool selectWaveActiveAnyTrue(Register ResVReg, const SPIRVType *ResType,
+ MachineInstr &I) const;
+
bool selectWaveReadLaneAt(Register ResVReg, const SPIRVType *ResType,
MachineInstr &I) const;
@@ -1762,6 +1765,25 @@ bool SPIRVInstructionSelector::selectSign(Register ResVReg,
return Result;
}
+bool SPIRVInstructionSelector::selectWaveActiveAnyTrue(Register ResVReg,
+ const SPIRVType *ResType,
+ MachineInstr &I) const {
+ assert(I.getNumOperands() == 3);
+ assert(I.getOperand(2).isReg());
+
+ // IntTy is used to define the execution scope, set to 3 to denote a
+ // cross-lane interaction equivalent to a SPIR-V subgroup.
+ MachineBasicBlock &BB = *I.getParent();
+ SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
+
+ return BuildMI(BB, I, I.getDebugLoc(),
+ TII.get(SPIRV::OpGroupNonUniformAny))
+ .addDef(ResVReg)
+ .addUse(GR.getSPIRVTypeID(ResType))
+ .addUse(GR.getOrCreateConstInt(3, I, IntTy, TII))
+ .addUse(I.getOperand(2).getReg());
+}
+
bool SPIRVInstructionSelector::selectWaveReadLaneAt(Register ResVReg,
const SPIRVType *ResType,
MachineInstr &I) const {
@@ -2567,6 +2589,8 @@ bool SPIRVInstructionSelector::selectIntrinsic(Register ResVReg,
.addUse(GR.getSPIRVTypeID(ResType))
.addUse(GR.getOrCreateConstInt(3, I, IntTy, TII));
}
+ case Intrinsic::spv_wave_activeanytrue:
+ return selectWaveActiveAnyTrue(ResVReg, ResType, I);
case Intrinsic::spv_wave_readlane:
return selectWaveReadLaneAt(ResVReg, ResType, I);
case Intrinsic::spv_step:
diff --git a/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll b/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
new file mode 100644
index 00000000000000..cc97eb2eedcd60
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
@@ -0,0 +1,10 @@
+; RUN: opt -S -dxil-op-lower -mtriple=dxil-pc-shadermodel6.3-compute %s | FileCheck %s
+
+define noundef i1 @wave_aat_simple(i1 noundef %p1) {
+entry:
+; CHECK: call i1 @dx.op.waveActiveAnyTrue(i32 113, i1 %p1)
+ %ret = call i1 @llvm.dx.wave.activeanytrue(i1 %p1)
+ ret i1 %ret
+}
+
+declare i1 @llvm.dx.wave.activeanytrue(i1)
diff --git a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
new file mode 100644
index 00000000000000..8b56b60f1da4da
--- /dev/null
+++ b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
@@ -0,0 +1,17 @@
+; RUN: llc -verify-machineinstrs -O0 -mtriple=spirv32v1.3-vulkan-unknown %s -o - | FileCheck %s
+; RUN: %if spirv-tools %{ llc -O0 -mtriple=spirv32v1.3-vulkan-unknown %s -o - -filetype=obj | spirv-val %}
+
+; CHECK: %[[#bool:]] = OpTypeBool
+; CHECK: %[[#uint:]] = OpTypeInt 32 0
+; CHECK: %[[#scope:]] = OpConstant %[[#uint]] 3
+
+; CHECK-LABEL: Begin function test_wave_aat
+define i1 @test_wave_aat(i1 %p1) {
+entry:
+; CHECK: %[[#param:]] = OpFunctionParameter %[[#bool]]
+; CHECK: %[[#ret:]] = OpGroupNonUniformAny %[[#bool]] %[[#scope]] %[[#param]]
+ %ret = call i1 @llvm.spv.wave.activeanytrue(i1 %p1)
+ ret i1 %ret
+}
+
+declare i1 @llvm.spv.wave.activeanytrue(i1)
>From 3b63a3a3135026114ca25143484a7d0187cbc020 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Tue, 12 Nov 2024 09:38:34 -0700
Subject: [PATCH 02/11] formatting
---
clang/lib/CodeGen/CGBuiltin.cpp | 25 +++++++++++++------
clang/lib/Sema/SemaHLSL.cpp | 6 ++---
.../Target/SPIRV/SPIRVInstructionSelector.cpp | 7 +++---
3 files changed, 23 insertions(+), 15 deletions(-)
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c4a6a9abee63cc..23b89c446213da 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18963,15 +18963,24 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
case Builtin::BI__builtin_hlsl_wave_active_any_true: {
// Assert Op->getType() == Bool
- // FIXME: PR Question: Can this be simpler? Looks like Int1Ty isn't predefined
- IntegerType* Int1Ty = llvm::Type::getInt1Ty(CGM.getTypes().getLLVMContext());
+ // FIXME: PR Question: Can this be simpler? Looks like Int1Ty isn't
+ // predefined
+ IntegerType *Int1Ty =
+ llvm::Type::getInt1Ty(CGM.getTypes().getLLVMContext());
Value *Op = EmitScalarExpr(E->getArg(0));
- assert(Op->getType() == Int1Ty && "wave_active_any_true operand must be a bool");
-
- // FIXME: PR Question: Re Style SingleRef vs {SingleRef} vs ArrayRef{SingleRef}
- llvm::FunctionType *FT = llvm::FunctionType::get(Int1Ty, {Int1Ty}, /*isVarArg=*/false);
- llvm::StringRef Name = Intrinsic::getName(CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic());
- return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, Name, {}, /*Local=*/false, /*AssumeConvergent=*/true), {Op}, "hlsl.wave.activeanytrue");
+ assert(Op->getType() == Int1Ty &&
+ "wave_active_any_true operand must be a bool");
+
+ // FIXME: PR Question: Re Style SingleRef vs {SingleRef} vs
+ // ArrayRef{SingleRef}
+ llvm::FunctionType *FT =
+ llvm::FunctionType::get(Int1Ty, {Int1Ty}, /*isVarArg=*/false);
+ llvm::StringRef Name = Intrinsic::getName(
+ CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic());
+ return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, Name, {},
+ /*Local=*/false,
+ /*AssumeConvergent=*/true),
+ {Op}, "hlsl.wave.activeanytrue");
}
case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
// We don't define a SPIR-V intrinsic, instead it is a SPIR-V built-in
diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index e12a4ce1e75bd6..0377f46dcf1991 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -2067,10 +2067,10 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
break;
}
case Builtin::BI__builtin_hlsl_wave_active_any_true: {
- if (SemaRef.checkArgCount(TheCall, 1))
- return true;
+ if (SemaRef.checkArgCount(TheCall, 1))
+ return true;
- break;
+ break;
}
case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
if (SemaRef.checkArgCount(TheCall, 2))
diff --git a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
index 8d4faedd1c6210..47538da068ec0a 100644
--- a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
+++ b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
@@ -1766,8 +1766,8 @@ bool SPIRVInstructionSelector::selectSign(Register ResVReg,
}
bool SPIRVInstructionSelector::selectWaveActiveAnyTrue(Register ResVReg,
- const SPIRVType *ResType,
- MachineInstr &I) const {
+ const SPIRVType *ResType,
+ MachineInstr &I) const {
assert(I.getNumOperands() == 3);
assert(I.getOperand(2).isReg());
@@ -1776,8 +1776,7 @@ bool SPIRVInstructionSelector::selectWaveActiveAnyTrue(Register ResVReg,
MachineBasicBlock &BB = *I.getParent();
SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
- return BuildMI(BB, I, I.getDebugLoc(),
- TII.get(SPIRV::OpGroupNonUniformAny))
+ return BuildMI(BB, I, I.getDebugLoc(), TII.get(SPIRV::OpGroupNonUniformAny))
.addDef(ResVReg)
.addUse(GR.getSPIRVTypeID(ResType))
.addUse(GR.getOrCreateConstInt(3, I, IntTy, TII))
>From 37fbd29b7032de3172cf4e372ab0e4674ee76e96 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Tue, 12 Nov 2024 09:41:29 -0700
Subject: [PATCH 03/11] cleanup
---
clang/lib/CodeGen/CGBuiltin.cpp | 6 ------
1 file changed, 6 deletions(-)
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 23b89c446213da..cc1995cd2382c8 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18961,18 +18961,12 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
ArrayRef<Value *>{Op0, Op1}, nullptr, "hlsl.step");
}
case Builtin::BI__builtin_hlsl_wave_active_any_true: {
- // Assert Op->getType() == Bool
-
- // FIXME: PR Question: Can this be simpler? Looks like Int1Ty isn't
- // predefined
IntegerType *Int1Ty =
llvm::Type::getInt1Ty(CGM.getTypes().getLLVMContext());
Value *Op = EmitScalarExpr(E->getArg(0));
assert(Op->getType() == Int1Ty &&
"wave_active_any_true operand must be a bool");
- // FIXME: PR Question: Re Style SingleRef vs {SingleRef} vs
- // ArrayRef{SingleRef}
llvm::FunctionType *FT =
llvm::FunctionType::get(Int1Ty, {Int1Ty}, /*isVarArg=*/false);
llvm::StringRef Name = Intrinsic::getName(
>From fcd1949d0e82d39c2808700bbd3b0ecd0bdc199d Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Tue, 12 Nov 2024 16:47:38 -0700
Subject: [PATCH 04/11] Address comments
---
clang/lib/CodeGen/CGBuiltin.cpp | 9 ++++-----
clang/lib/Sema/SemaHLSL.cpp | 6 ------
clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl | 4 ++--
llvm/lib/Target/DirectX/DXIL.td | 3 +--
llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp | 6 ++----
llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll | 2 +-
.../CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll | 4 ++--
7 files changed, 12 insertions(+), 22 deletions(-)
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 88960bffe0bee7..b2261f650616d3 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -19109,20 +19109,19 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
ArrayRef<Value *>{Op0, Op1}, nullptr, "hlsl.step");
}
case Builtin::BI__builtin_hlsl_wave_active_any_true: {
- IntegerType *Int1Ty =
- llvm::Type::getInt1Ty(CGM.getTypes().getLLVMContext());
Value *Op = EmitScalarExpr(E->getArg(0));
- assert(Op->getType() == Int1Ty &&
- "wave_active_any_true operand must be a bool");
+ llvm::Type *Ty = Op->getType();
+ assert(Ty->isIntegerTy(1) && "wave_active_any_true operand must be a bool");
llvm::FunctionType *FT =
- llvm::FunctionType::get(Int1Ty, {Int1Ty}, /*isVarArg=*/false);
+ llvm::FunctionType::get(Ty, {Ty}, /*isVarArg=*/false);
llvm::StringRef Name = Intrinsic::getName(
CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic());
return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, Name, {},
/*Local=*/false,
/*AssumeConvergent=*/true),
{Op}, "hlsl.wave.activeanytrue");
+ }
case Builtin::BI__builtin_hlsl_wave_active_count_bits: {
Value *OpExpr = EmitScalarExpr(E->getArg(0));
Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveActiveCountBitsIntrinsic();
diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index 8aa0b7520b05eb..65b0d9cd65637f 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -2091,12 +2091,6 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
return true;
break;
}
- case Builtin::BI__builtin_hlsl_wave_active_any_true: {
- if (SemaRef.checkArgCount(TheCall, 1))
- return true;
-
- break;
- }
case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
if (SemaRef.checkArgCount(TheCall, 2))
return true;
diff --git a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
index d8233c45e0ff2d..dd7a9dee5c3f1e 100644
--- a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
@@ -1,7 +1,7 @@
-// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -fnative-half-type -triple \
+// 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 -std=hlsl2021 -finclude-default-header -fnative-half-type -triple \
+// 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
diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 9b5e2b5ce1e92a..cb0a0a17492bb5 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -202,7 +202,6 @@ defset list<DXILOpClass> OpClasses = {
def unpack4x8 : DXILOpClass;
def viewID : DXILOpClass;
def waveActiveAllEqual : DXILOpClass;
- def waveActiveAnyTrue : DXILOpClass;
def waveActiveBallot : DXILOpClass;
def waveActiveBit : DXILOpClass;
def waveActiveOp : DXILOpClass;
@@ -854,7 +853,7 @@ def CreateHandleFromBinding : DXILOp<217, createHandleFromBinding> {
let stages = [Stages<DXIL1_6, [all_stages]>];
}
-def WaveActiveAnyTrue : DXILOp<113, waveActiveAnyTrue> {
+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 LLVMIntrinsic = int_dx_wave_activeanytrue;
let arguments = [Int1Ty];
diff --git a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
index 728880a855f1c5..98735c2582ffa2 100644
--- a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
+++ b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
@@ -1929,15 +1929,13 @@ bool SPIRVInstructionSelector::selectWaveActiveAnyTrue(Register ResVReg,
assert(I.getNumOperands() == 3);
assert(I.getOperand(2).isReg());
- // IntTy is used to define the execution scope, set to 3 to denote a
- // cross-lane interaction equivalent to a SPIR-V subgroup.
MachineBasicBlock &BB = *I.getParent();
SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
return BuildMI(BB, I, I.getDebugLoc(), TII.get(SPIRV::OpGroupNonUniformAny))
.addDef(ResVReg)
.addUse(GR.getSPIRVTypeID(ResType))
- .addUse(GR.getOrCreateConstInt(3, I, IntTy, TII))
+ .addUse(GR.getOrCreateConstInt(SPIRV::Scope::Subgroup, I, IntTy, TII))
.addUse(I.getOperand(2).getReg());
}
@@ -1987,7 +1985,7 @@ bool SPIRVInstructionSelector::selectWaveReadLaneAt(Register ResVReg,
TII.get(SPIRV::OpGroupNonUniformShuffle))
.addDef(ResVReg)
.addUse(GR.getSPIRVTypeID(ResType))
- .addUse(GR.getOrCreateConstInt(3, I, IntTy, TII))
+ .addUse(GR.getOrCreateConstInt(SPIRV::Scope::Subgroup, I, IntTy, TII))
.addUse(I.getOperand(2).getReg())
.addUse(I.getOperand(3).getReg());
}
diff --git a/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll b/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
index cc97eb2eedcd60..8f7ae9ea7b370b 100644
--- a/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
+++ b/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
@@ -2,7 +2,7 @@
define noundef i1 @wave_aat_simple(i1 noundef %p1) {
entry:
-; CHECK: call i1 @dx.op.waveActiveAnyTrue(i32 113, i1 %p1)
+; CHECK: call i1 @dx.op.waveAnyTrue(i32 113, i1 %p1)
%ret = call i1 @llvm.dx.wave.activeanytrue(i1 %p1)
ret i1 %ret
}
diff --git a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
index 8b56b60f1da4da..25d195a776ab73 100644
--- a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
+++ b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
@@ -1,5 +1,5 @@
-; RUN: llc -verify-machineinstrs -O0 -mtriple=spirv32v1.3-vulkan-unknown %s -o - | FileCheck %s
-; RUN: %if spirv-tools %{ llc -O0 -mtriple=spirv32v1.3-vulkan-unknown %s -o - -filetype=obj | spirv-val %}
+; RUN: llc -verify-machineinstrs -O0 -mtriple=spirv32v1.5-unknown-unknown %s -o - | FileCheck %s
+; RUN: %if spirv-tools %{ llc -O0 -mtriple=spirv32v1.5-unknown-unknown %s -o - -filetype=obj | spirv-val %}
; CHECK: %[[#bool:]] = OpTypeBool
; CHECK: %[[#uint:]] = OpTypeInt 32 0
>From dd6eff10aa8ec763edbfecb484a3130b4d8704e1 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Wed, 13 Nov 2024 10:02:03 -0700
Subject: [PATCH 05/11] Address comments
---
clang/lib/CodeGen/CGBuiltin.cpp | 11 +++--------
.../test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl | 1 +
.../SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl | 2 +-
3 files changed, 5 insertions(+), 9 deletions(-)
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index b2261f650616d3..ddff01acb134a0 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -19113,14 +19113,9 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
llvm::Type *Ty = Op->getType();
assert(Ty->isIntegerTy(1) && "wave_active_any_true operand must be a bool");
- llvm::FunctionType *FT =
- llvm::FunctionType::get(Ty, {Ty}, /*isVarArg=*/false);
- llvm::StringRef Name = Intrinsic::getName(
- CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic());
- return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, Name, {},
- /*Local=*/false,
- /*AssumeConvergent=*/true),
- {Op}, "hlsl.wave.activeanytrue");
+ Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic();
+ return EmitRuntimeCall(
+ Intrinsic::getOrInsertDeclaration(&CGM.getModule(), ID), {Op});
}
case Builtin::BI__builtin_hlsl_wave_active_count_bits: {
Value *OpExpr = EmitScalarExpr(E->getArg(0));
diff --git a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
index dd7a9dee5c3f1e..c8dc45aa396dec 100644
--- a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
@@ -13,5 +13,6 @@ bool test(bool p1) {
// CHECK-SPIRV: %[[RET:.*]] = call spir_func i1 @llvm.spv.wave.activeanytrue(i1 %{{[a-zA-Z0-9]+}}) [ "convergencectrl"(token %[[#entry_tok0]]) ]
// CHECK-DXIL: %[[RET:.*]] = call i1 @llvm.dx.wave.activeanytrue(i1 %{{[a-zA-Z0-9]+}})
// CHECK: ret i1 %[[RET]]
+ // CHECK: fex
return WaveActiveAnyTrue(p1);
}
diff --git a/clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl b/clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl
index 48fda7b571fca7..875aae06517020 100644
--- a/clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl
+++ b/clang/test/SemaHLSL/BuiltIns/WaveActiveAnyTrue-errors.hlsl
@@ -15,7 +15,7 @@ struct Foo
int a;
};
-bool test_type_check_2(Foo p0) {
+bool test_type_check(Foo p0) {
return __builtin_hlsl_wave_active_any_true(p0);
// expected-error at -1 {{no viable conversion from 'Foo' to 'bool'}}
}
>From 9c43edfa9ea48383748d3e919f37f7337e048465 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Wed, 13 Nov 2024 10:09:52 -0700
Subject: [PATCH 06/11] cleanup
---
clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl | 1 -
1 file changed, 1 deletion(-)
diff --git a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
index c8dc45aa396dec..dd7a9dee5c3f1e 100644
--- a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
@@ -13,6 +13,5 @@ bool test(bool p1) {
// CHECK-SPIRV: %[[RET:.*]] = call spir_func i1 @llvm.spv.wave.activeanytrue(i1 %{{[a-zA-Z0-9]+}}) [ "convergencectrl"(token %[[#entry_tok0]]) ]
// CHECK-DXIL: %[[RET:.*]] = call i1 @llvm.dx.wave.activeanytrue(i1 %{{[a-zA-Z0-9]+}})
// CHECK: ret i1 %[[RET]]
- // CHECK: fex
return WaveActiveAnyTrue(p1);
}
>From 7823ce6bc12b7ab88e4d64608b3c27129d6d01aa Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Thu, 14 Nov 2024 14:59:36 -0700
Subject: [PATCH 07/11] Rename intrinsic
---
clang/lib/CodeGen/CGHLSLRuntime.h | 2 +-
clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl | 4 ++--
llvm/include/llvm/IR/IntrinsicsDirectX.td | 2 +-
llvm/include/llvm/IR/IntrinsicsSPIRV.td | 2 +-
llvm/lib/Target/DirectX/DXIL.td | 2 +-
llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp | 4 ++--
llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll | 6 +++---
.../CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll | 8 ++++----
8 files changed, 15 insertions(+), 15 deletions(-)
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h b/clang/lib/CodeGen/CGHLSLRuntime.h
index e8413a5baa32e9..a8e0ed42b79a35 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -91,7 +91,7 @@ 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(WaveActiveAnyTrue, wave_activeanytrue)
+ GENERATE_HLSL_INTRINSIC_FUNCTION(WaveActiveAnyTrue, wave_any)
GENERATE_HLSL_INTRINSIC_FUNCTION(WaveActiveCountBits, wave_active_countbits)
GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_readlane)
diff --git a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
index dd7a9dee5c3f1e..d657a3f8357b08 100644
--- a/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/WaveActiveAnyTrue.hlsl
@@ -10,8 +10,8 @@
// CHECK-LABEL: 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.activeanytrue(i1 %{{[a-zA-Z0-9]+}}) [ "convergencectrl"(token %[[#entry_tok0]]) ]
- // CHECK-DXIL: %[[RET:.*]] = call i1 @llvm.dx.wave.activeanytrue(i1 %{{[a-zA-Z0-9]+}})
+ // CHECK-SPIRV: %[[RET:.*]] = call spir_func i1 @llvm.spv.wave.any(i1 %{{[a-zA-Z0-9]+}}) [ "convergencectrl"(token %[[#entry_tok0]]) ]
+ // CHECK-DXIL: %[[RET:.*]] = call i1 @llvm.dx.wave.any(i1 %{{[a-zA-Z0-9]+}})
// CHECK: ret i1 %[[RET]]
return WaveActiveAnyTrue(p1);
}
diff --git a/llvm/include/llvm/IR/IntrinsicsDirectX.td b/llvm/include/llvm/IR/IntrinsicsDirectX.td
index 8ed81e9852c863..58e5f44789c9cf 100644
--- a/llvm/include/llvm/IR/IntrinsicsDirectX.td
+++ b/llvm/include/llvm/IR/IntrinsicsDirectX.td
@@ -91,9 +91,9 @@ 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_any : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
def int_dx_wave_getlaneindex : DefaultAttrsIntrinsic<[llvm_i32_ty], [], [IntrConvergent, IntrNoMem]>;
def int_dx_wave_is_first_lane : DefaultAttrsIntrinsic<[llvm_i1_ty], [], [IntrConvergent]>;
-def int_dx_wave_activeanytrue : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
def int_dx_wave_readlane : DefaultAttrsIntrinsic<[llvm_any_ty], [LLVMMatchType<0>, llvm_i32_ty], [IntrConvergent, IntrNoMem]>;
def int_dx_sign : DefaultAttrsIntrinsic<[LLVMScalarOrSameVectorWidth<0, llvm_i32_ty>], [llvm_any_ty], [IntrNoMem]>;
def int_dx_step : DefaultAttrsIntrinsic<[LLVMMatchType<0>], [llvm_anyfloat_ty, LLVMMatchType<0>], [IntrNoMem]>;
diff --git a/llvm/include/llvm/IR/IntrinsicsSPIRV.td b/llvm/include/llvm/IR/IntrinsicsSPIRV.td
index 02f9acc3401555..8dfdcf01e0d757 100644
--- a/llvm/include/llvm/IR/IntrinsicsSPIRV.td
+++ b/llvm/include/llvm/IR/IntrinsicsSPIRV.td
@@ -86,9 +86,9 @@ 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_any : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
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]>;
- def int_spv_wave_activeanytrue : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
def int_spv_sign : DefaultAttrsIntrinsic<[LLVMScalarOrSameVectorWidth<0, llvm_i32_ty>], [llvm_any_ty], [IntrNoMem]>;
def int_spv_radians : DefaultAttrsIntrinsic<[LLVMMatchType<0>], [llvm_anyfloat_ty], [IntrNoMem]>;
def int_spv_group_memory_barrier_with_group_sync : DefaultAttrsIntrinsic<[], [], []>;
diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index cb0a0a17492bb5..04fee894a46b06 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -855,7 +855,7 @@ def CreateHandleFromBinding : DXILOp<217, createHandleFromBinding> {
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 LLVMIntrinsic = int_dx_wave_activeanytrue;
+ let LLVMIntrinsic = int_dx_wave_any;
let arguments = [Int1Ty];
let result = Int1Ty;
let stages = [Stages<DXIL1_0, [all_stages]>];
diff --git a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
index 98735c2582ffa2..23ba5f4b3282ef 100644
--- a/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
+++ b/llvm/lib/Target/SPIRV/SPIRVInstructionSelector.cpp
@@ -2800,6 +2800,8 @@ 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_any:
+ return selectWaveActiveAnyTrue(ResVReg, ResType, I);
case Intrinsic::spv_wave_is_first_lane: {
SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
return BuildMI(BB, I, I.getDebugLoc(),
@@ -2808,8 +2810,6 @@ bool SPIRVInstructionSelector::selectIntrinsic(Register ResVReg,
.addUse(GR.getSPIRVTypeID(ResType))
.addUse(GR.getOrCreateConstInt(3, I, IntTy, TII));
}
- case Intrinsic::spv_wave_activeanytrue:
- return selectWaveActiveAnyTrue(ResVReg, ResType, I);
case Intrinsic::spv_wave_readlane:
return selectWaveReadLaneAt(ResVReg, ResType, I);
case Intrinsic::spv_step:
diff --git a/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll b/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
index 8f7ae9ea7b370b..5adf050a76c98f 100644
--- a/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
+++ b/llvm/test/CodeGen/DirectX/WaveActiveAnyTrue.ll
@@ -1,10 +1,10 @@
; RUN: opt -S -dxil-op-lower -mtriple=dxil-pc-shadermodel6.3-compute %s | FileCheck %s
-define noundef i1 @wave_aat_simple(i1 noundef %p1) {
+define noundef i1 @wave_any_simple(i1 noundef %p1) {
entry:
; CHECK: call i1 @dx.op.waveAnyTrue(i32 113, i1 %p1)
- %ret = call i1 @llvm.dx.wave.activeanytrue(i1 %p1)
+ %ret = call i1 @llvm.dx.wave.any(i1 %p1)
ret i1 %ret
}
-declare i1 @llvm.dx.wave.activeanytrue(i1)
+declare i1 @llvm.dx.wave.any(i1)
diff --git a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
index 25d195a776ab73..39418bd35ff4c3 100644
--- a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
+++ b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
@@ -5,13 +5,13 @@
; CHECK: %[[#uint:]] = OpTypeInt 32 0
; CHECK: %[[#scope:]] = OpConstant %[[#uint]] 3
-; CHECK-LABEL: Begin function test_wave_aat
-define i1 @test_wave_aat(i1 %p1) {
+; CHECK-LABEL: Begin function test_wave_any
+define i1 @test_wave_any(i1 %p1) {
entry:
; CHECK: %[[#param:]] = OpFunctionParameter %[[#bool]]
; CHECK: %[[#ret:]] = OpGroupNonUniformAny %[[#bool]] %[[#scope]] %[[#param]]
- %ret = call i1 @llvm.spv.wave.activeanytrue(i1 %p1)
+ %ret = call i1 @llvm.spv.wave.any(i1 %p1)
ret i1 %ret
}
-declare i1 @llvm.spv.wave.activeanytrue(i1)
+declare i1 @llvm.spv.wave.any(i1)
>From dfeb3276bf4b67022e4b39919a8a2e8e353c9777 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Thu, 14 Nov 2024 15:40:25 -0700
Subject: [PATCH 08/11] address comments
---
llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp | 4 +++-
llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll | 4 ++--
2 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp b/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp
index e8641b3a105dec..17912a398b982f 100644
--- a/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp
+++ b/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp
@@ -630,6 +630,9 @@ void RequirementHandler::initAvailableCapabilities(const SPIRVSubtarget &ST) {
addAvailableCaps({Capability::Shader, Capability::Linkage, Capability::Int8,
Capability::Int16});
+ if (ST.isAtLeastSPIRVVer(VersionTuple(1, 3)))
+ addAvailableCaps({Capability::GroupNonUniformVote});
+
if (ST.isAtLeastSPIRVVer(VersionTuple(1, 6)))
addAvailableCaps({Capability::DotProduct, Capability::DotProductInputAll,
Capability::DotProductInput4x8Bit,
@@ -675,7 +678,6 @@ void RequirementHandler::initAvailableCapabilitiesForOpenCL(
addAvailableCaps({Capability::SubgroupDispatch, Capability::PipeStorage});
if (ST.isAtLeastSPIRVVer(VersionTuple(1, 3)))
addAvailableCaps({Capability::GroupNonUniform,
- Capability::GroupNonUniformVote,
Capability::GroupNonUniformArithmetic,
Capability::GroupNonUniformBallot,
Capability::GroupNonUniformClustered,
diff --git a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
index 39418bd35ff4c3..66bc9619fb74f9 100644
--- a/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
+++ b/llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveAnyTrue.ll
@@ -1,5 +1,5 @@
-; RUN: llc -verify-machineinstrs -O0 -mtriple=spirv32v1.5-unknown-unknown %s -o - | FileCheck %s
-; RUN: %if spirv-tools %{ llc -O0 -mtriple=spirv32v1.5-unknown-unknown %s -o - -filetype=obj | spirv-val %}
+; 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: %[[#bool:]] = OpTypeBool
; CHECK: %[[#uint:]] = OpTypeInt 32 0
>From 66848d80e6da2e433bdc332562c862ef1bf4e3e7 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <me at ashleycoleman.me>
Date: Thu, 14 Nov 2024 16:44:57 -0700
Subject: [PATCH 09/11] Apply suggestions from code review
Co-authored-by: Finn Plummer <50529406+inbelic at users.noreply.github.com>
---
clang/lib/CodeGen/CGBuiltin.cpp | 2 +-
llvm/include/llvm/IR/IntrinsicsDirectX.td | 2 +-
llvm/include/llvm/IR/IntrinsicsSPIRV.td | 2 +-
llvm/lib/Target/DirectX/DXIL.td | 1 -
4 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 0602bcf85e5574..af03d477f13509 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -19123,7 +19123,7 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
case Builtin::BI__builtin_hlsl_wave_active_any_true: {
Value *Op = EmitScalarExpr(E->getArg(0));
llvm::Type *Ty = Op->getType();
- assert(Ty->isIntegerTy(1) && "wave_active_any_true operand must be a bool");
+ assert(Ty->isIntegerTy(1) && "Intrinsic WaveActiveAnyTrue operand must be a bool");
Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic();
return EmitRuntimeCall(
diff --git a/llvm/include/llvm/IR/IntrinsicsDirectX.td b/llvm/include/llvm/IR/IntrinsicsDirectX.td
index 52f3845ecba6df..a5fe1e55f5cc13 100644
--- a/llvm/include/llvm/IR/IntrinsicsDirectX.td
+++ b/llvm/include/llvm/IR/IntrinsicsDirectX.td
@@ -94,7 +94,7 @@ 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_any : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
+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]>;
def int_dx_wave_readlane : DefaultAttrsIntrinsic<[llvm_any_ty], [LLVMMatchType<0>, llvm_i32_ty], [IntrConvergent, IntrNoMem]>;
diff --git a/llvm/include/llvm/IR/IntrinsicsSPIRV.td b/llvm/include/llvm/IR/IntrinsicsSPIRV.td
index 9af8ae70646283..28db6fd8117790 100644
--- a/llvm/include/llvm/IR/IntrinsicsSPIRV.td
+++ b/llvm/include/llvm/IR/IntrinsicsSPIRV.td
@@ -86,7 +86,7 @@ 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_any : DefaultAttrsIntrinsic<[llvm_i1_ty], [llvm_i1_ty], [IntrConvergent]>;
+ 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]>;
def int_spv_sign : DefaultAttrsIntrinsic<[LLVMScalarOrSameVectorWidth<0, llvm_i32_ty>], [llvm_any_ty], [IntrNoMem]>;
diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 5c1891e855d977..2a499b1d9b75c0 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -859,7 +859,6 @@ def WaveActiveAnyTrue : DXILOp<113, waveAnyTrue> {
let arguments = [Int1Ty];
let result = Int1Ty;
let stages = [Stages<DXIL1_0, [all_stages]>];
- let attributes = [Attributes<DXIL1_0, [ReadNone]>];
}
def WaveIsFirstLane : DXILOp<110, waveIsFirstLane> {
>From 883b2a80002b8e0d0c5222bd302c716df642ba76 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Thu, 14 Nov 2024 16:50:08 -0700
Subject: [PATCH 10/11] format
---
clang/lib/CodeGen/CGBuiltin.cpp | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index af03d477f13509..d6e90e7821ceec 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -19123,7 +19123,8 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
case Builtin::BI__builtin_hlsl_wave_active_any_true: {
Value *Op = EmitScalarExpr(E->getArg(0));
llvm::Type *Ty = Op->getType();
- assert(Ty->isIntegerTy(1) && "Intrinsic WaveActiveAnyTrue operand must be a bool");
+ assert(Ty->isIntegerTy(1) &&
+ "Intrinsic WaveActiveAnyTrue operand must be a bool");
Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveActiveAnyTrueIntrinsic();
return EmitRuntimeCall(
>From 4dd242a1abbe2a402e7ed6e4bca522357286f980 Mon Sep 17 00:00:00 2001
From: Ashley Coleman <ascoleman at microsoft.com>
Date: Thu, 14 Nov 2024 16:54:52 -0700
Subject: [PATCH 11/11] Address comments
---
llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp | 15 +++++++--------
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp b/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp
index 3e887a3b03150c..3093521b962fd3 100644
--- a/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp
+++ b/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.cpp
@@ -631,7 +631,13 @@ void RequirementHandler::initAvailableCapabilities(const SPIRVSubtarget &ST) {
Capability::Int16});
if (ST.isAtLeastSPIRVVer(VersionTuple(1, 3)))
- addAvailableCaps({Capability::GroupNonUniformVote});
+ addAvailableCaps({Capability::GroupNonUniform,
+ Capability::GroupNonUniformVote,
+ Capability::GroupNonUniformArithmetic,
+ Capability::GroupNonUniformBallot,
+ Capability::GroupNonUniformClustered,
+ Capability::GroupNonUniformShuffle,
+ Capability::GroupNonUniformShuffleRelative});
if (ST.isAtLeastSPIRVVer(VersionTuple(1, 6)))
addAvailableCaps({Capability::DotProduct, Capability::DotProductInputAll,
@@ -677,13 +683,6 @@ void RequirementHandler::initAvailableCapabilitiesForOpenCL(
if (ST.isAtLeastSPIRVVer(VersionTuple(1, 1)) &&
ST.isAtLeastOpenCLVer(VersionTuple(2, 2)))
addAvailableCaps({Capability::SubgroupDispatch, Capability::PipeStorage});
- if (ST.isAtLeastSPIRVVer(VersionTuple(1, 3)))
- addAvailableCaps({Capability::GroupNonUniform,
- Capability::GroupNonUniformArithmetic,
- Capability::GroupNonUniformBallot,
- Capability::GroupNonUniformClustered,
- Capability::GroupNonUniformShuffle,
- Capability::GroupNonUniformShuffleRelative});
if (ST.isAtLeastSPIRVVer(VersionTuple(1, 4)))
addAvailableCaps({Capability::DenormPreserve, Capability::DenormFlushToZero,
Capability::SignedZeroInfNanPreserve,
More information about the cfe-commits
mailing list