from:"Finn Plummer via cfe\-commits"

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-08 Thread Finn Plummer via cfe-commits



@@ -428,6 +431,7 @@ bool SPIRVInstructionSelector::spvSelect(Register ResVReg,
 
   case TargetOpcode::G_INTRINSIC:
   case TargetOpcode::G_INTRINSIC_W_SIDE_EFFECTS:
+  case TargetOpcode::G_INTRINSIC_CONVERGENT:

inbelic wrote:

Afaict, not within this section of the codebase, it is used elsewhere though eg 
`GlobalISel`. But I think @Keenuts would be able to better elaborate.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-08 Thread Finn Plummer via cfe-commits



@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, waveReadLaneAt)

inbelic wrote:

If we use `_` separated then the generated names will use `.` to separate them. 
So eg `wave_read_lane_at` becomes `llvm.dx.wave.read.lane.at`. Which we decided 
was quite confusing due to the confusion with a namespace. I will clean up the 
other intrinsic to cohere to this in a follow up pr.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 01/10] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===-

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-08 Thread Finn Plummer via cfe-commits



@@ -0,0 +1,26 @@
+// RUN: %clang_cc1 -finclude-default-header -triple 
dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only 
-disable-llvm-passes -verify -verify-ignore-unexpected

inbelic wrote:

Good catch. There are no ignored diagnostics and it is not needed.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-08 Thread Finn Plummer via cfe-commits


inbelic wrote:

I will go ahead and merge this tomorrow if there are no other comments.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111576
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic ready_for_review 
https://github.com/llvm/llvm-project/pull/111576
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111576
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex (PR #111576)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic created 
https://github.com/llvm/llvm-project/pull/111576

- add additional lowering for directx backend in CGBuiltin.cpp
- add directx intrinsic to IntrinscsDirectX.td
- add semantic check of arguments in SemaHLSL.cpp
- add mapping to DXIL op in DXIL.td

- add testing of semantics in WaveGetLaneIndex-errors.hlsl
- add testing of dxil lowering in WaveGetLaneIndex.ll

>From 0911b563e582c621993ad908858f2ec12b0bf57e Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Tue, 8 Oct 2024 09:28:43 -0700
Subject: [PATCH] [HLSL][DXIL] Implement WaveGetLaneIndex

- add additional lowering for directx backend in CGBuiltin.cpp
- add directx intrinsic to IntrinscsDirectX.td
- add semantic check of arguments in SemaHLSL.cpp
- add mapping to DXIL op in DXIL.td

- add testing of semantics in WaveGetLaneIndex-errors.hlsl
- add testing of dxil lowering in WaveGetLaneIndex.ll
---
 clang/lib/CodeGen/CGBuiltin.cpp   | 15 ---
 clang/lib/Sema/SemaHLSL.cpp   |  5 +
 .../BuiltIns/WaveGetLaneIndex-errors.hlsl |  6 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/lib/Target/DirectX/DXIL.td   |  9 +
 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll | 10 ++
 6 files changed, 43 insertions(+), 3 deletions(-)
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveGetLaneIndex-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..0d92e5a581cbfd 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18827,9 +18827,18 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
   case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
-return EmitRuntimeCall(CGM.CreateRuntimeFunction(
-llvm::FunctionType::get(IntTy, {}, false), 
"__hlsl_wave_get_lane_index",
-{}, false, true));
+switch (CGM.getTarget().getTriple().getArch()) {
+case llvm::Triple::dxil:
+  return EmitRuntimeCall(Intrinsic::getDeclaration(
+  &CGM.getModule(), Intrinsic::dx_waveGetLaneIndex));
+case llvm::Triple::spirv:
+  return EmitRuntimeCall(CGM.CreateRuntimeFunction(
+  llvm::FunctionType::get(IntTy, {}, false),
+  "__hlsl_wave_get_lane_index", {}, false, true));
+default:
+  llvm_unreachable(
+  "Intrinsic waveGetLaneIndex not supported by target architecture");
+}
   }
   case Builtin::BI__builtin_hlsl_wave_is_first_lane: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index 43cc6c81ae5cb0..7d93f41bb2d7a0 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -1956,6 +1956,11 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned 
BuiltinID, CallExpr *TheCall) {
   return true;
 break;
   }
+  case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
+if (SemaRef.checkArgCount(TheCall, 0))
+  return true;
+break;
+  }
   case Builtin::BI__builtin_elementwise_acos:
   case Builtin::BI__builtin_elementwise_asin:
   case Builtin::BI__builtin_elementwise_atan:
diff --git a/clang/test/SemaHLSL/BuiltIns/WaveGetLaneIndex-errors.hlsl 
b/clang/test/SemaHLSL/BuiltIns/WaveGetLaneIndex-errors.hlsl
new file mode 100644
index 00..94cfd0662b5fc3
--- /dev/null
+++ b/clang/test/SemaHLSL/BuiltIns/WaveGetLaneIndex-errors.hlsl
@@ -0,0 +1,6 @@
+// RUN: %clang_cc1 -finclude-default-header -triple 
dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only 
-disable-llvm-passes -verify -verify-ignore-unexpected
+
+int test_too_many_arg(int x) {
+  return __builtin_hlsl_wave_get_lane_index(x);
+  // expected-error@-1 {{too many arguments to function call, expected 0, have 
1}}
+}
diff --git a/llvm/include/llvm/IR/IntrinsicsDirectX.td 
b/llvm/include/llvm/IR/IntrinsicsDirectX.td
index 555877e7aaf0e5..585cb2b7b300df 100644
--- a/llvm/include/llvm/IR/IntrinsicsDirectX.td
+++ b/llvm/include/llvm/IR/IntrinsicsDirectX.td
@@ -82,6 +82,7 @@ def int_dx_imad : DefaultAttrsIntrinsic<[llvm_anyint_ty], 
[LLVMMatchType<0>, LLV
 def int_dx_umad : DefaultAttrsIntrinsic<[llvm_anyint_ty], [LLVMMatchType<0>, 
LLVMMatchType<0>, LLVMMatchType<0>], [IntrNoMem]>;
 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_waveGetLaneIndex : DefaultAttrsIntrinsic<[llvm_i32_ty], [], 
[IntrConvergent, IntrNoMem]>;
 def int_dx_wave_is_first_lane : DefaultAttrsIntrinsic<[llvm_i1_ty], [], 
[IntrConvergent]>;
 def int_dx_sign : DefaultAttrsIntrinsic<[LLVMScalarOrSameVectorWidth<0, 
llvm_i32_ty>], [llvm_any_ty], [IntrNoMem]>;
 def int_dx_step : Defa

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-15 Thread Finn Plummer via cfe-commits


https://github.com/inbelic created 
https://github.com/llvm/llvm-project/pull/112400

- add clang builtin to Builtins.td
- link builtin in hlsl_intrinsics
- add codegen for spirv intrinsic and two directx intrinsics to retain 
signedness information of the operands in CGBuiltin.cpp
- add semantic analysis in SemaHLSL.cpp
- add lowering of spirv intrinsic to spirv backend in 
SPIRVInstructionSelector.cpp
- add directx intrinsic expansion to WaveActiveOp in 
DXILIntrinsicExpansion.cpp

- add test cases to illustrate passes

>From 952271bd348f9d695a1c655a4fe3d3d37b192da4 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 11 Oct 2024 15:06:15 -0700
Subject: [PATCH] [HLSL] Implement `WaveActiveSum` intrinsic

- add clang builtin to Builtins.td
- link builtin in hlsl_intrinsics
- add codegen for spirv intrinsic and two directx intrinsics to
  retain signedness information of the operands in CGBuiltin.cpp
- add semantic analysis in SemaHLSL.cpp
- add lowering of spirv intrinsic to spirv backend in
  SPIRVInstructionSelector.cpp
- add directx intrinsic expansion to WaveActiveOp in
  DXILIntrinsicExpansion.cpp

- add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 32 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  3 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 21 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 34 +++
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 78 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 13 files changed, 428 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 9ebee81fcb0d3d..2b4fb25c97504a 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4721,6 +4721,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_sum"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void (...)";
+}
+
 def HLSLWaveGetLaneIndex : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_get_lane_index"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td 
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index f4a2d4a3f0656a..e2f7384fb632a6 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -9232,6 +9232,9 @@ def err_typecheck_cond_incompatible_operands : Error<
 def err_typecheck_expect_scalar_or_vector : Error<
   "invalid operand of type %0 where %1 or "
   "a vector of such type is required">;
+def err_typecheck_expect_scalar_or_vector_not_type : Error<
+  "invalid operand of type %0 where %1 or "
+  "a vector of such type is not allowed">;
 def err_typecheck_expect_flt_or_vector : Error<
   "invalid operand of type %0 where floating, complex or "
   "a vector of such types is required">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 2449b90a0e7902..4da012418cf902 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18631,6 +18631,23 @@ static Intrinsic::ID 
getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   return RT.getUDotIntrinsic();
 }
 
+// Return wave active sum that corresponds to the QT scalar type
+static Intrinsic::ID getWaveActiveSumIntrinsic(llvm::Triple::ArchType Arch,
+   CGHLSLRuntime &RT, QualType QT) 
{
+  switch (Arch) {
+  case llvm::Triple::spirv:
+return llvm::Intrinsic::spv_wave_active_sum;
+  case llvm::Triple::dxil: {
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::dx_wave_active_usum;
+return llvm::Intrinsic::dx_wave_active_sum;
+  }
+  default:
+llvm_unreachable("Intrinsic WaveActiveSum"
+ " not supported by target architecture");
+  }
+}
+
 Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
 const CallExpr *E,
 ReturnValueSlot ReturnValue) {
@@ -18866,6 +18883

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-15 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From c541955941a11b57efc624b87f50ce61f1b4c26a Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 11 Oct 2024 15:06:15 -0700
Subject: [PATCH] [HLSL] Implement `WaveActiveSum` intrinsic

- add clang builtin to Builtins.td
- link builtin in hlsl_intrinsics
- add codegen for spirv intrinsic and two directx intrinsics to
  retain signedness information of the operands in CGBuiltin.cpp
- add semantic analysis in SemaHLSL.cpp
- add lowering of spirv intrinsic to spirv backend in
  SPIRVInstructionSelector.cpp
- add directx intrinsic expansion to WaveActiveOp in
  DXILIntrinsicExpansion.cpp

- add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 32 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 21 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 34 +++
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 78 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 13 files changed, 427 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 9ebee81fcb0d3d..2b4fb25c97504a 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4721,6 +4721,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_sum"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void (...)";
+}
+
 def HLSLWaveGetLaneIndex : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_get_lane_index"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td 
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index f4a2d4a3f0656a..e2f7384fb632a6 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -9232,6 +9232,9 @@ def err_typecheck_cond_incompatible_operands : Error<
 def err_typecheck_expect_scalar_or_vector : Error<
   "invalid operand of type %0 where %1 or "
   "a vector of such type is required">;
+def err_typecheck_expect_scalar_or_vector_not_type : Error<
+  "invalid operand of type %0 where %1 or "
+  "a vector of such type is not allowed">;
 def err_typecheck_expect_flt_or_vector : Error<
   "invalid operand of type %0 where floating, complex or "
   "a vector of such types is required">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 2449b90a0e7902..4da012418cf902 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18631,6 +18631,23 @@ static Intrinsic::ID 
getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   return RT.getUDotIntrinsic();
 }
 
+// Return wave active sum that corresponds to the QT scalar type
+static Intrinsic::ID getWaveActiveSumIntrinsic(llvm::Triple::ArchType Arch,
+   CGHLSLRuntime &RT, QualType QT) 
{
+  switch (Arch) {
+  case llvm::Triple::spirv:
+return llvm::Intrinsic::spv_wave_active_sum;
+  case llvm::Triple::dxil: {
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::dx_wave_active_usum;
+return llvm::Intrinsic::dx_wave_active_sum;
+  }
+  default:
+llvm_unreachable("Intrinsic WaveActiveSum"
+ " not supported by target architecture");
+  }
+}
+
 Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
 const CallExpr *E,
 ReturnValueSlot ReturnValue) {
@@ -18866,6 +18883,23 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_sum: {
+// Due to the use of variadic arguments, explicitly retreive argument
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{Op

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-15 Thread Finn Plummer via cfe-commits


inbelic wrote:

Dependent on https://github.com/llvm/llvm-project/pull/112058 and 
https://github.com/llvm/llvm-project/pull/111010.

https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][HLSL] Add GroupMemoryBarrierWithGroupSync intrinsic (PR #111883)

2024-10-15 Thread Finn Plummer via cfe-commits



@@ -0,0 +1,6 @@
+// RUN: %clang_cc1 -finclude-default-header -triple 
dxil-pc-shadermodel6.6-library %s -emit-llvm-only -disable-llvm-passes -verify 
-verify-ignore-unexpected

inbelic wrote:

Will forward this comment here about `-verify-ignore-unexpected`: 
https://github.com/llvm/llvm-project/pull/111209#discussion_r1793906097

https://github.com/llvm/llvm-project/pull/111883
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][HLSL] Add GroupMemoryBarrierWithGroupSync intrinsic (PR #111883)

2024-10-15 Thread Finn Plummer via cfe-commits



@@ -2018,6 +2018,11 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned 
BuiltinID, CallExpr *TheCall) {
   return true;
 break;
   }
+  case Builtin::BI__builtin_hlsl_group_memory_barrier_with_group_sync: {
+if (SemaRef.checkArgCountAtMost(TheCall, 0))

inbelic wrote:

Is there a reason to not use `SemaRef.checkArgCount(TheCall, 0)`? I think that 
would be more descriptive.

https://github.com/llvm/llvm-project/pull/111883
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][HLSL] Add GroupMemoryBarrierWithGroupSync intrinsic (PR #111883)

2024-10-15 Thread Finn Plummer via cfe-commits


https://github.com/inbelic approved this pull request.

LGTM, just a couple nits.

https://github.com/llvm/llvm-project/pull/111883
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][HLSL] Add GroupMemoryBarrierWithGroupSync intrinsic (PR #111883)

2024-10-15 Thread Finn Plummer via cfe-commits



@@ -4830,6 +4830,12 @@ def HLSLRadians : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLGroupMemoryBarrierWithGroupSync: LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_group_memory_barrier_with_group_sync"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";

inbelic wrote:

As with `WaveIsFirstLane` I don't think we need `...` and can just do `()`

https://github.com/llvm/llvm-project/pull/111883
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][HLSL] Add GroupMemoryBarrierWithGroupSync intrinsic (PR #111883)

2024-10-15 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111883
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-09 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 01/11] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===-

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-09 Thread Finn Plummer via cfe-commits



@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, waveReadLaneAt)

inbelic wrote:

Updated to `wave_readlaneat`.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-09 Thread Finn Plummer via cfe-commits



@@ -18827,9 +18827,21 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
   case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
-return EmitRuntimeCall(CGM.CreateRuntimeFunction(
-llvm::FunctionType::get(IntTy, {}, false), 
"__hlsl_wave_get_lane_index",
-{}, false, true));
+// Since we don't define a SPIR-V intrinsic for the SPIR-V built-in from
+// SPIRVBuiltins.td, manually get the matching name for the DirectX
+// intrinsic and the demangled builtin name

inbelic wrote:

The SPIR-V side has already been implemented using a builtin defined in 
`SPIRVBuiltins.td`. I will improve the comment to be more clear.

https://github.com/llvm/llvm-project/pull/111576
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-09 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111576

>From 167718e352554e167c4690a4962234ba782bceff Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Tue, 8 Oct 2024 09:28:43 -0700
Subject: [PATCH 1/3] [HLSL][DXIL] Implement WaveGetLaneIndex

- add additional lowering for directx backend in CGBuiltin.cpp
- add directx intrinsic to IntrinscsDirectX.td
- add semantic check of arguments in SemaHLSL.cpp
- add mapping to DXIL op in DXIL.td

- add testing of semantics in WaveGetLaneIndex-errors.hlsl
- add testing of dxil lowering in WaveGetLaneIndex.ll
---
 clang/lib/CodeGen/CGBuiltin.cpp   | 15 +++---
 clang/lib/Sema/SemaHLSL.cpp   |  5 +
 .../builtins/wave_get_lane_index_simple.hlsl  | 20 +--
 .../BuiltIns/WaveGetLaneIndex-errors.hlsl |  6 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/lib/Target/DirectX/DXIL.td   |  9 +
 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll | 10 ++
 7 files changed, 57 insertions(+), 9 deletions(-)
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveGetLaneIndex-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..c83b2c5a2ca047 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18827,9 +18827,18 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
   case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
-return EmitRuntimeCall(CGM.CreateRuntimeFunction(
-llvm::FunctionType::get(IntTy, {}, false), 
"__hlsl_wave_get_lane_index",
-{}, false, true));
+switch (CGM.getTarget().getTriple().getArch()) {
+case llvm::Triple::dxil:
+  return EmitRuntimeCall(Intrinsic::getDeclaration(
+  &CGM.getModule(), Intrinsic::dx_waveGetLaneIndex));
+case llvm::Triple::spirv:
+  return EmitRuntimeCall(CGM.CreateRuntimeFunction(
+  llvm::FunctionType::get(IntTy, {}, false),
+  "__hlsl_wave_get_lane_index", {}, false, true));
+default:
+  llvm_unreachable(
+  "Intrinsic WaveGetLaneIndex not supported by target architecture");
+}
   }
   case Builtin::BI__builtin_hlsl_wave_is_first_lane: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index 43cc6c81ae5cb0..7d93f41bb2d7a0 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -1956,6 +1956,11 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned 
BuiltinID, CallExpr *TheCall) {
   return true;
 break;
   }
+  case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
+if (SemaRef.checkArgCount(TheCall, 0))
+  return true;
+break;
+  }
   case Builtin::BI__builtin_elementwise_acos:
   case Builtin::BI__builtin_elementwise_asin:
   case Builtin::BI__builtin_elementwise_atan:
diff --git a/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl 
b/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
index 8f52d81091c180..e76d35a86e512a 100644
--- a/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
@@ -1,14 +1,22 @@
 // RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
-// RUN:   spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | 
FileCheck %s
+// RUN:   spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | 
FileCheck %s \
+// RUN:   --check-prefixes=CHECK,CHECK-SPIRV
+// RUN: %clang_cc1 -finclude-default-header \
+// RUN:   -triple dxil-pc-shadermodel6.3-library %s \
+// RUN:   -emit-llvm -disable-llvm-passes -o - | FileCheck %s \
+// RUN:   --check-prefixes=CHECK,CHECK-DXIL
 
-// CHECK: define spir_func noundef i32 @_Z6test_1v() [[A0:#[0-9]+]] {
-// CHECK: %[[CI:[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call i32 @__hlsl_wave_get_lane_index() [ "convergencectrl"(token 
%[[CI]]) ]
-uint test_1() {
+// CHECK-SPIRV: define spir_func noundef i32 @{{.*test_1.*}}() [[A0:#[0-9]+]] {
+// CHECK-DXIL: define noundef i32 @{{.*test_1.*}}() [[A0:#[0-9]+]] {
+// CHECK-SPIRV: %[[CI:[0-9]+]] = call token 
@llvm.experimental.convergence.entry()
+// CHECK-SPIRV: call i32 @__hlsl_wave_get_lane_index() [ 
"convergencectrl"(token %[[CI]]) ]
+// CHECK-DXIL: call i32 @llvm.dx.waveGetLaneIndex()
+int test_1() {
   return WaveGetLaneIndex();
 }
 
-// CHECK: declare i32 @__hlsl_wave_get_lane_index() [[A1:#[0-9]+]]
+// CHECK-SPIRV: declare i32 @__hlsl_wave_get_lane_index() [[A1:#[0-9]+]]
+// CHECK-DXIL: declare i32 @llvm.dx.waveGetLaneIndex() [[A1:#[0-9]+]]
 
 // CHECK-DAG: attributes [[A0]] = { {{.*}}convergent{{.*}} }
 // CHECK-DAG: attributes [[A1]] = { {{.*}}convergent{{.*}} }
diff --git a/clang/test/SemaHL

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-09 Thread Finn Plummer via cfe-commits



@@ -0,0 +1,6 @@
+// RUN: %clang_cc1 -finclude-default-header -triple 
dxil-pc-shadermodel6.6-library %s -fnative-half-type -emit-llvm-only 
-disable-llvm-passes -verify -verify-ignore-unexpected

inbelic wrote:

The flag was not needed, so have removed it.

https://github.com/llvm/llvm-project/pull/111576
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-09 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111209

>From 398dcb5c1a354c12d4d43af152f6d3e14f001274 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 2 Oct 2024 13:30:33 -0700
Subject: [PATCH 1/2] [HLSL] Implementation the `degrees` intrinsic

- add degrees builtin
- link degrees api in hlsl_intrinsics.h
- add degrees intrinsic to IntrinsicDirectX.td
- add degrees intrinsic to IntrinsicSPIRV.td
- add lowering from clang builtin to dx/spv intrinsics in
  CGBuiltin.cpp
- add semantic checks to SemaHLSL.cpp
- add expansion of directx intrinsic to llvm fmul for DirectX in
  DXILIntrinsicExpansion.cpp
- add mapping to spir-v intrinsic in SPIRVInstructionSelector.cpp

- add test coverage:
- degrees.hlsl -> check hlsl lowering to dx/spv degrees intrinsics
- degrees-errors.hlsl/half-float-only-errors ->
check semantic warnings
- hlsl-intrinsics/degrees.ll -> check lowering of spir-v
degrees intrinsic to SPIR-V backend
- DirectX/degrees.ll -> check expansion and scalarization of
  directx degrees intrinsic to fmul
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 11 +++
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 30 
 clang/lib/Sema/SemaHLSL.cpp   |  1 +
 clang/test/CodeGenHLSL/builtins/degrees.hlsl  | 64 +
 .../SemaHLSL/BuiltIns/degrees-errors.hlsl | 26 +++
 .../BuiltIns/half-float-only-errors.hlsl  |  1 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 12 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp |  2 +
 llvm/test/CodeGen/DirectX/degrees.ll  | 54 +++
 .../CodeGen/SPIRV/hlsl-intrinsics/degrees.ll  | 68 +++
 llvm/test/CodeGen/SPIRV/opencl/degrees.ll | 50 ++
 15 files changed, 328 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/degrees.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/degrees-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/opencl/degrees.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 5b504666365b32..44142298987f12 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4739,6 +4739,12 @@ def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDegrees : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_elementwise_degrees"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_dot"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c864714182e019..0f824fc2c4431a 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18740,6 +18740,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 CGM.getHLSLRuntime().getNormalizeIntrinsic(), ArrayRef{X},
 nullptr, "hlsl.normalize");
   }
+  case Builtin::BI__builtin_hlsl_elementwise_degrees: {
+Value *X = EmitScalarExpr(E->getArg(0));
+
+assert(E->getArg(0)->getType()->hasFloatingRepresentation() &&
+"degree operand must have a float representation");
+
+return Builder.CreateIntrinsic(
+/*ReturnType=*/X->getType(),
+CGM.getHLSLRuntime().getDegreesIntrinsic(),
+ArrayRef{X}, nullptr, "hlsl.degrees");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_frac: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 if (!E->getArg(0)->getType()->hasFloatingRepresentation())
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..966111a9600017 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -74,6 +74,7 @@ class CGHLSLRuntime {
 
   GENERATE_HLSL_INTRINSIC_FUNCTION(All, all)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Any, any)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(Degrees, degrees)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Frac, frac)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Length, length)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Lerp, lerp)
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index d28f204e352de5..07004a48f96128 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -766,6 +766,36 @@ uint64_t3 countbits(uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_elementwise_popcount)
 uint64_t4 countbits(uint64_t4);
 
+//===

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-09 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111576
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/9] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===---

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-11 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 01/14] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===-

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-11 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 01/14] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===-

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111209

>From 398dcb5c1a354c12d4d43af152f6d3e14f001274 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 2 Oct 2024 13:30:33 -0700
Subject: [PATCH] [HLSL] Implementation the `degrees` intrinsic

- add degrees builtin
- link degrees api in hlsl_intrinsics.h
- add degrees intrinsic to IntrinsicDirectX.td
- add degrees intrinsic to IntrinsicSPIRV.td
- add lowering from clang builtin to dx/spv intrinsics in
  CGBuiltin.cpp
- add semantic checks to SemaHLSL.cpp
- add expansion of directx intrinsic to llvm fmul for DirectX in
  DXILIntrinsicExpansion.cpp
- add mapping to spir-v intrinsic in SPIRVInstructionSelector.cpp

- add test coverage:
- degrees.hlsl -> check hlsl lowering to dx/spv degrees intrinsics
- degrees-errors.hlsl/half-float-only-errors ->
check semantic warnings
- hlsl-intrinsics/degrees.ll -> check lowering of spir-v
degrees intrinsic to SPIR-V backend
- DirectX/degrees.ll -> check expansion and scalarization of
  directx degrees intrinsic to fmul
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 11 +++
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 30 
 clang/lib/Sema/SemaHLSL.cpp   |  1 +
 clang/test/CodeGenHLSL/builtins/degrees.hlsl  | 64 +
 .../SemaHLSL/BuiltIns/degrees-errors.hlsl | 26 +++
 .../BuiltIns/half-float-only-errors.hlsl  |  1 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 12 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp |  2 +
 llvm/test/CodeGen/DirectX/degrees.ll  | 54 +++
 .../CodeGen/SPIRV/hlsl-intrinsics/degrees.ll  | 68 +++
 llvm/test/CodeGen/SPIRV/opencl/degrees.ll | 50 ++
 15 files changed, 328 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/degrees.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/degrees-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/opencl/degrees.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 5b504666365b32..44142298987f12 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4739,6 +4739,12 @@ def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDegrees : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_elementwise_degrees"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_dot"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c864714182e019..0f824fc2c4431a 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18740,6 +18740,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 CGM.getHLSLRuntime().getNormalizeIntrinsic(), ArrayRef{X},
 nullptr, "hlsl.normalize");
   }
+  case Builtin::BI__builtin_hlsl_elementwise_degrees: {
+Value *X = EmitScalarExpr(E->getArg(0));
+
+assert(E->getArg(0)->getType()->hasFloatingRepresentation() &&
+"degree operand must have a float representation");
+
+return Builder.CreateIntrinsic(
+/*ReturnType=*/X->getType(),
+CGM.getHLSLRuntime().getDegreesIntrinsic(),
+ArrayRef{X}, nullptr, "hlsl.degrees");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_frac: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 if (!E->getArg(0)->getType()->hasFloatingRepresentation())
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..966111a9600017 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -74,6 +74,7 @@ class CGHLSLRuntime {
 
   GENERATE_HLSL_INTRINSIC_FUNCTION(All, all)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Any, any)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(Degrees, degrees)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Frac, frac)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Length, length)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Lerp, lerp)
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index d28f204e352de5..07004a48f96128 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -766,6 +766,36 @@ uint64_t3 countbits(uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_elementwise_popcount)
 uint64_t4 countbits(uint64_t4);
 
+//===

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111576

>From 167718e352554e167c4690a4962234ba782bceff Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Tue, 8 Oct 2024 09:28:43 -0700
Subject: [PATCH] [HLSL][DXIL] Implement WaveGetLaneIndex

- add additional lowering for directx backend in CGBuiltin.cpp
- add directx intrinsic to IntrinscsDirectX.td
- add semantic check of arguments in SemaHLSL.cpp
- add mapping to DXIL op in DXIL.td

- add testing of semantics in WaveGetLaneIndex-errors.hlsl
- add testing of dxil lowering in WaveGetLaneIndex.ll
---
 clang/lib/CodeGen/CGBuiltin.cpp   | 15 +++---
 clang/lib/Sema/SemaHLSL.cpp   |  5 +
 .../builtins/wave_get_lane_index_simple.hlsl  | 20 +--
 .../BuiltIns/WaveGetLaneIndex-errors.hlsl |  6 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/lib/Target/DirectX/DXIL.td   |  9 +
 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll | 10 ++
 7 files changed, 57 insertions(+), 9 deletions(-)
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveGetLaneIndex-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..c83b2c5a2ca047 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18827,9 +18827,18 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
   case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
-return EmitRuntimeCall(CGM.CreateRuntimeFunction(
-llvm::FunctionType::get(IntTy, {}, false), 
"__hlsl_wave_get_lane_index",
-{}, false, true));
+switch (CGM.getTarget().getTriple().getArch()) {
+case llvm::Triple::dxil:
+  return EmitRuntimeCall(Intrinsic::getDeclaration(
+  &CGM.getModule(), Intrinsic::dx_waveGetLaneIndex));
+case llvm::Triple::spirv:
+  return EmitRuntimeCall(CGM.CreateRuntimeFunction(
+  llvm::FunctionType::get(IntTy, {}, false),
+  "__hlsl_wave_get_lane_index", {}, false, true));
+default:
+  llvm_unreachable(
+  "Intrinsic WaveGetLaneIndex not supported by target architecture");
+}
   }
   case Builtin::BI__builtin_hlsl_wave_is_first_lane: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index 43cc6c81ae5cb0..7d93f41bb2d7a0 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -1956,6 +1956,11 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned 
BuiltinID, CallExpr *TheCall) {
   return true;
 break;
   }
+  case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
+if (SemaRef.checkArgCount(TheCall, 0))
+  return true;
+break;
+  }
   case Builtin::BI__builtin_elementwise_acos:
   case Builtin::BI__builtin_elementwise_asin:
   case Builtin::BI__builtin_elementwise_atan:
diff --git a/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl 
b/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
index 8f52d81091c180..e76d35a86e512a 100644
--- a/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
@@ -1,14 +1,22 @@
 // RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
-// RUN:   spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | 
FileCheck %s
+// RUN:   spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | 
FileCheck %s \
+// RUN:   --check-prefixes=CHECK,CHECK-SPIRV
+// RUN: %clang_cc1 -finclude-default-header \
+// RUN:   -triple dxil-pc-shadermodel6.3-library %s \
+// RUN:   -emit-llvm -disable-llvm-passes -o - | FileCheck %s \
+// RUN:   --check-prefixes=CHECK,CHECK-DXIL
 
-// CHECK: define spir_func noundef i32 @_Z6test_1v() [[A0:#[0-9]+]] {
-// CHECK: %[[CI:[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call i32 @__hlsl_wave_get_lane_index() [ "convergencectrl"(token 
%[[CI]]) ]
-uint test_1() {
+// CHECK-SPIRV: define spir_func noundef i32 @{{.*test_1.*}}() [[A0:#[0-9]+]] {
+// CHECK-DXIL: define noundef i32 @{{.*test_1.*}}() [[A0:#[0-9]+]] {
+// CHECK-SPIRV: %[[CI:[0-9]+]] = call token 
@llvm.experimental.convergence.entry()
+// CHECK-SPIRV: call i32 @__hlsl_wave_get_lane_index() [ 
"convergencectrl"(token %[[CI]]) ]
+// CHECK-DXIL: call i32 @llvm.dx.waveGetLaneIndex()
+int test_1() {
   return WaveGetLaneIndex();
 }
 
-// CHECK: declare i32 @__hlsl_wave_get_lane_index() [[A1:#[0-9]+]]
+// CHECK-SPIRV: declare i32 @__hlsl_wave_get_lane_index() [[A1:#[0-9]+]]
+// CHECK-DXIL: declare i32 @llvm.dx.waveGetLaneIndex() [[A1:#[0-9]+]]
 
 // CHECK-DAG: attributes [[A0]] = { {{.*}}convergent{{.*}} }
 // CHECK-DAG: attributes [[A1]] = { {{.*}}convergent{{.*}} }
diff --git a/clang/test/SemaHLSL/B

[clang] [llvm] [HLSL][DXIL] Implement WaveGetLaneIndex Intrinsic (PR #111576)

2024-10-08 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111576

>From 167718e352554e167c4690a4962234ba782bceff Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Tue, 8 Oct 2024 09:28:43 -0700
Subject: [PATCH 1/2] [HLSL][DXIL] Implement WaveGetLaneIndex

- add additional lowering for directx backend in CGBuiltin.cpp
- add directx intrinsic to IntrinscsDirectX.td
- add semantic check of arguments in SemaHLSL.cpp
- add mapping to DXIL op in DXIL.td

- add testing of semantics in WaveGetLaneIndex-errors.hlsl
- add testing of dxil lowering in WaveGetLaneIndex.ll
---
 clang/lib/CodeGen/CGBuiltin.cpp   | 15 +++---
 clang/lib/Sema/SemaHLSL.cpp   |  5 +
 .../builtins/wave_get_lane_index_simple.hlsl  | 20 +--
 .../BuiltIns/WaveGetLaneIndex-errors.hlsl |  6 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/lib/Target/DirectX/DXIL.td   |  9 +
 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll | 10 ++
 7 files changed, 57 insertions(+), 9 deletions(-)
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveGetLaneIndex-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveGetLaneIndex.ll

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..c83b2c5a2ca047 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18827,9 +18827,18 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
   case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
-return EmitRuntimeCall(CGM.CreateRuntimeFunction(
-llvm::FunctionType::get(IntTy, {}, false), 
"__hlsl_wave_get_lane_index",
-{}, false, true));
+switch (CGM.getTarget().getTriple().getArch()) {
+case llvm::Triple::dxil:
+  return EmitRuntimeCall(Intrinsic::getDeclaration(
+  &CGM.getModule(), Intrinsic::dx_waveGetLaneIndex));
+case llvm::Triple::spirv:
+  return EmitRuntimeCall(CGM.CreateRuntimeFunction(
+  llvm::FunctionType::get(IntTy, {}, false),
+  "__hlsl_wave_get_lane_index", {}, false, true));
+default:
+  llvm_unreachable(
+  "Intrinsic WaveGetLaneIndex not supported by target architecture");
+}
   }
   case Builtin::BI__builtin_hlsl_wave_is_first_lane: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index 43cc6c81ae5cb0..7d93f41bb2d7a0 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -1956,6 +1956,11 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned 
BuiltinID, CallExpr *TheCall) {
   return true;
 break;
   }
+  case Builtin::BI__builtin_hlsl_wave_get_lane_index: {
+if (SemaRef.checkArgCount(TheCall, 0))
+  return true;
+break;
+  }
   case Builtin::BI__builtin_elementwise_acos:
   case Builtin::BI__builtin_elementwise_asin:
   case Builtin::BI__builtin_elementwise_atan:
diff --git a/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl 
b/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
index 8f52d81091c180..e76d35a86e512a 100644
--- a/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/wave_get_lane_index_simple.hlsl
@@ -1,14 +1,22 @@
 // RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
-// RUN:   spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | 
FileCheck %s
+// RUN:   spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | 
FileCheck %s \
+// RUN:   --check-prefixes=CHECK,CHECK-SPIRV
+// RUN: %clang_cc1 -finclude-default-header \
+// RUN:   -triple dxil-pc-shadermodel6.3-library %s \
+// RUN:   -emit-llvm -disable-llvm-passes -o - | FileCheck %s \
+// RUN:   --check-prefixes=CHECK,CHECK-DXIL
 
-// CHECK: define spir_func noundef i32 @_Z6test_1v() [[A0:#[0-9]+]] {
-// CHECK: %[[CI:[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call i32 @__hlsl_wave_get_lane_index() [ "convergencectrl"(token 
%[[CI]]) ]
-uint test_1() {
+// CHECK-SPIRV: define spir_func noundef i32 @{{.*test_1.*}}() [[A0:#[0-9]+]] {
+// CHECK-DXIL: define noundef i32 @{{.*test_1.*}}() [[A0:#[0-9]+]] {
+// CHECK-SPIRV: %[[CI:[0-9]+]] = call token 
@llvm.experimental.convergence.entry()
+// CHECK-SPIRV: call i32 @__hlsl_wave_get_lane_index() [ 
"convergencectrl"(token %[[CI]]) ]
+// CHECK-DXIL: call i32 @llvm.dx.waveGetLaneIndex()
+int test_1() {
   return WaveGetLaneIndex();
 }
 
-// CHECK: declare i32 @__hlsl_wave_get_lane_index() [[A1:#[0-9]+]]
+// CHECK-SPIRV: declare i32 @__hlsl_wave_get_lane_index() [[A1:#[0-9]+]]
+// CHECK-DXIL: declare i32 @llvm.dx.waveGetLaneIndex() [[A1:#[0-9]+]]
 
 // CHECK-DAG: attributes [[A0]] = { {{.*}}convergent{{.*}} }
 // CHECK-DAG: attributes [[A1]] = { {{.*}}convergent{{.*}} }
diff --git a/clang/test/SemaHL

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-04 Thread Finn Plummer via cfe-commits



@@ -2653,6 +2653,21 @@ bool SPIRVInstructionSelector::selectIntrinsic(Register 
ResVReg,
 .addUse(GR.getSPIRVTypeID(ResType))
 .addUse(GR.getOrCreateConstInt(3, I, IntTy, TII));
   }
+  case Intrinsic::spv_wave_read_lane_at: {
+assert(I.getNumOperands() == 4);
+assert(I.getOperand(2).isReg());
+assert(I.getOperand(3).isReg());
+
+// Defines the execution scope currently 2 for group, see scope table
+SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
+return BuildMI(BB, I, I.getDebugLoc(),
+   TII.get(SPIRV::OpGroupNonUniformShuffle))
+.addDef(ResVReg)
+.addUse(GR.getSPIRVTypeID(ResType))
+.addUse(I.getOperand(2).getReg())
+.addUse(I.getOperand(3).getReg())
+.addUse(GR.getOrCreateConstInt(2, I, IntTy, TII));

inbelic wrote:

Right, it should facilitate communication within the wave not between waves. 
Thanks.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/3] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===---

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/3] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===---

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implementation the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic created 
https://github.com/llvm/llvm-project/pull/111209

- add degrees builtin
- link degrees api in hlsl_intrinsics.h
- add degrees intrinsic to IntrinsicDirectX.td
- add degrees intrinsic to IntrinsicSPIRV.td
- add lowering from clang builtin to dx/spv intrinsics in CGBuiltin.cpp
- add semantic checks to SemaHLSL.cpp
- add expansion of directx intrinsic to llvm fmul for DirectX in 
DXILIntrinsicExpansion.cpp
- add mapping to spir-v intrinsic in SPIRVInstructionSelector.cpp

- add test coverage:
- degrees.hlsl -> check hlsl lowering to dx/spv degrees intrinsics
- degrees-errors.hlsl/half-float-only-errors -> check semantic warnings
- hlsl-intrinsics/degrees.ll -> check lowering of spir-v degrees 
intrinsic to SPIR-V backend
- DirectX/degrees.ll -> check expansion and scalarization of directx 
degrees intrinsic to fmul

>From 04fa155a8f45c2fdb30b9233f0bd7d4783045a55 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 2 Oct 2024 13:30:33 -0700
Subject: [PATCH] [HLSL] Implementation the `degrees` intrinsic

- add degrees builtin
- link degrees api in hlsl_intrinsics.h
- add degrees intrinsic to IntrinsicDirectX.td
- add degrees intrinsic to IntrinsicSPIRV.td
- add lowering from clang builtin to dx/spv intrinsics in
  CGBuiltin.cpp
- add semantic checks to SemaHLSL.cpp
- add expansion of directx intrinsic to llvm fmul for DirectX in
  DXILIntrinsicExpansion.cpp
- add mapping to spir-v intrinsic in SPIRVInstructionSelector.cpp

- add test coverage:
- degrees.hlsl -> check hlsl lowering to dx/spv degrees intrinsics
- degrees-errors.hlsl/half-float-only-errors ->
check semantic warnings
- hlsl-intrinsics/degrees.ll -> check lowering of spir-v
degrees intrinsic to SPIR-V backend
- DirectX/degrees.ll -> check expansion and scalarization of
  directx degrees intrinsic to fmul
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 11 +++
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 30 
 clang/lib/Sema/SemaHLSL.cpp   |  1 +
 clang/test/CodeGenHLSL/builtins/degrees.hlsl  | 64 +
 .../SemaHLSL/BuiltIns/degrees-errors.hlsl | 37 ++
 .../BuiltIns/half-float-only-errors.hlsl  |  1 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 12 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 22 ++
 llvm/test/CodeGen/DirectX/degrees.ll  | 54 +++
 .../CodeGen/SPIRV/hlsl-intrinsics/degrees.ll  | 68 +++
 14 files changed, 309 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/degrees.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/degrees-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/degrees.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..910d90cb3fbc7e 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4709,6 +4709,12 @@ def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDegrees : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_elementwise_degrees"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_dot"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..ecaff0f83b76f3 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18715,6 +18715,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 CGM.getHLSLRuntime().getNormalizeIntrinsic(), ArrayRef{X},
 nullptr, "hlsl.normalize");
   }
+  case Builtin::BI__builtin_hlsl_elementwise_degrees: {
+Value *X = EmitScalarExpr(E->getArg(0));
+
+assert(E->getArg(0)->getType()->hasFloatingRepresentation() &&
+"degree operand must have a float representation");
+
+return Builder.CreateIntrinsic(
+/*ReturnType=*/X->getType(),
+CGM.getHLSLRuntime().getDegreesIntrinsic(),
+ArrayRef{X}, nullptr, "hlsl.degrees");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_frac: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 if (!E->getArg(0)->getType()->hasFloatingRepresentation())
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..966111a9600017 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -74,6 +74,7 @@ class CGHLS

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111209
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits



@@ -1643,6 +1646,23 @@ bool SPIRVInstructionSelector::selectLength(Register 
ResVReg,
   .constrainAllUses(TII, TRI, RBI);
 }
 
+bool SPIRVInstructionSelector::selectDegrees(Register ResVReg,
+ const SPIRVType *ResType,
+ MachineInstr &I) const {
+
+  assert(I.getNumOperands() == 3);

inbelic wrote:

Correct, but I believe the first 2 operands of the spir-v instruction are 
usually used for the `ReturnType` and the `Result` register.

https://github.com/llvm/llvm-project/pull/111209
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits



@@ -2625,6 +2645,8 @@ bool SPIRVInstructionSelector::selectIntrinsic(Register 
ResVReg,
 return selectFmix(ResVReg, ResType, I);
   case Intrinsic::spv_length:
 return selectLength(ResVReg, ResType, I);
+  case Intrinsic::spv_degrees:
+return selectDegrees(ResVReg, ResType, I);

inbelic wrote:

Yep, you are right and I now realize I should also support the option to use 
the opencl lowering path which can be done using `selectExtInst`.

https://github.com/llvm/llvm-project/pull/111209
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/7] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===---

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits



@@ -82,5 +82,6 @@ let TargetPrefix = "spv" in {
 [llvm_anyint_ty, LLVMScalarOrSameVectorWidth<0, LLVMVectorElementType<0>>],
 [IntrNoMem, Commutative] >;
   def int_spv_wave_is_first_lane : DefaultAttrsIntrinsic<[llvm_i1_ty], [], 
[IntrConvergent]>;
+  def int_spv_wave_read_lane_at : DefaultAttrsIntrinsic<[llvm_any_ty], 
[LLVMMatchType<0>, llvm_i32_ty], [IntrConvergent]>;

inbelic wrote:

Added a commit to change the name to `waveReadLaneAt` as described in below 
comment.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits



@@ -83,6 +83,7 @@ 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_is_first_lane : DefaultAttrsIntrinsic<[llvm_i1_ty], [], 
[IntrConvergent]>;
+def int_dx_wave_read_lane_at : DefaultAttrsIntrinsic<[llvm_any_ty], 
[LLVMMatchType<0>, llvm_i32_ty], [IntrConvergent]>;

inbelic wrote:

My understanding was that `IntrNoMem` denotes there are no side-effects in the 
operation and that cross-lane communication would be a side-effect.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/6] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===---

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits



@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");

inbelic wrote:

I think `__hlsl_wave_get_lane_index` is the odd one out. The other intrinsics 
follow the pattern of `hlsl.name`. Having changed to using one word 
`waveReadLaneAt` I think we can keep it consistent naming with 
`hlsl.waveReadLaneAt`.

I can change the name to `hlsl.waveGetLaneIndex` in the clean-up pr.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111209

>From 461d5133ea1ae1fcdffd036a926243ee3dce5abb Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 2 Oct 2024 13:30:33 -0700
Subject: [PATCH] [HLSL] Implementation the `degrees` intrinsic

- add degrees builtin
- link degrees api in hlsl_intrinsics.h
- add degrees intrinsic to IntrinsicDirectX.td
- add degrees intrinsic to IntrinsicSPIRV.td
- add lowering from clang builtin to dx/spv intrinsics in
  CGBuiltin.cpp
- add semantic checks to SemaHLSL.cpp
- add expansion of directx intrinsic to llvm fmul for DirectX in
  DXILIntrinsicExpansion.cpp
- add mapping to spir-v intrinsic in SPIRVInstructionSelector.cpp

- add test coverage:
- degrees.hlsl -> check hlsl lowering to dx/spv degrees intrinsics
- degrees-errors.hlsl/half-float-only-errors ->
check semantic warnings
- hlsl-intrinsics/degrees.ll -> check lowering of spir-v
degrees intrinsic to SPIR-V backend
- DirectX/degrees.ll -> check expansion and scalarization of
  directx degrees intrinsic to fmul
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 11 +++
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 30 
 clang/lib/Sema/SemaHLSL.cpp   |  1 +
 clang/test/CodeGenHLSL/builtins/degrees.hlsl  | 64 +
 .../SemaHLSL/BuiltIns/degrees-errors.hlsl | 37 ++
 .../BuiltIns/half-float-only-errors.hlsl  |  1 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 12 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp |  2 +
 llvm/test/CodeGen/DirectX/degrees.ll  | 54 +++
 .../CodeGen/SPIRV/hlsl-intrinsics/degrees.ll  | 68 +++
 llvm/test/CodeGen/SPIRV/opencl/degrees.ll | 50 ++
 15 files changed, 339 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/degrees.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/degrees-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/opencl/degrees.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 5b504666365b32..44142298987f12 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4739,6 +4739,12 @@ def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDegrees : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_elementwise_degrees"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_dot"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c864714182e019..0f824fc2c4431a 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18740,6 +18740,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 CGM.getHLSLRuntime().getNormalizeIntrinsic(), ArrayRef{X},
 nullptr, "hlsl.normalize");
   }
+  case Builtin::BI__builtin_hlsl_elementwise_degrees: {
+Value *X = EmitScalarExpr(E->getArg(0));
+
+assert(E->getArg(0)->getType()->hasFloatingRepresentation() &&
+"degree operand must have a float representation");
+
+return Builder.CreateIntrinsic(
+/*ReturnType=*/X->getType(),
+CGM.getHLSLRuntime().getDegreesIntrinsic(),
+ArrayRef{X}, nullptr, "hlsl.degrees");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_frac: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 if (!E->getArg(0)->getType()->hasFloatingRepresentation())
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..966111a9600017 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -74,6 +74,7 @@ class CGHLSLRuntime {
 
   GENERATE_HLSL_INTRINSIC_FUNCTION(All, all)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Any, any)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(Degrees, degrees)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Frac, frac)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Length, length)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Lerp, lerp)
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index d28f204e352de5..07004a48f96128 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -766,6 +766,36 @@ uint64_t3 countbits(uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_elementwise_popcount)
 uint64_t4 countbits(uint64_t4);
 
+//===-

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits


inbelic wrote:

Rebasing onto commit that allows the use of `selectExtInst` with both 
`CL::degrees` and `GL::Degrees`. Adding additional testcase for this as well.

https://github.com/llvm/llvm-project/pull/111209
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/8] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===---

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-07 Thread Finn Plummer via cfe-commits



@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);

inbelic wrote:

Fwiw, this is also true for the `select` intrinsic and we use the arbitrary 
type template.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic ready_for_review 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From 0bdb522d97752859b0b09c3ba7c815865b87 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 18 Oct 2024 10:49:18 -0700
Subject: [PATCH 1/2] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to retain
signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp

  - add test cases to illustrate passes

Note that this defines the dx intrinsics but does not implement
the DirectX lowering to DXIL. This will be implemented in a second pr
when the dependent pr merges.
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 33 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 11 files changed, 326 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..7205f10b01c2b9 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_sum"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void (...)";
+}
+
 def HLSLWaveGetLaneIndex : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_get_lane_index"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td 
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index c458a62d9be48c..95093723c2f495 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -9232,6 +9232,9 @@ def err_typecheck_expect_scalar_or_vector : Error<
   "a vector of such type is required">;
 def err_typecheck_expect_any_scalar_or_vector : Error<
   "invalid operand of type %0 where a scalar or vector is required">;
+def err_typecheck_expect_scalar_or_vector_not_type : Error<
+  "invalid operand of type %0 where %1 or "
+  "a vector of such type is not allowed">;
 def err_typecheck_expect_flt_or_vector : Error<
   "invalid operand of type %0 where floating, complex or "
   "a vector of such types is required">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 12f99d9f1178a9..46646492b30d31 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18638,6 +18638,23 @@ static Intrinsic::ID 
getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   return RT.getUDotIntrinsic();
 }
 
+// Return wave active sum that corresponds to the QT scalar type
+static Intrinsic::ID getWaveActiveSumIntrinsic(llvm::Triple::ArchType Arch,
+   CGHLSLRuntime &RT, QualType QT) 
{
+  switch (Arch) {
+  case llvm::Triple::spirv:
+return llvm::Intrinsic::spv_wave_active_sum;
+  case llvm::Triple::dxil: {
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::dx_wave_active_usum;
+return llvm::Intrinsic::dx_wave_active_sum;
+  }
+  default:
+llvm_unreachable("Intrinsic WaveActiveSum"
+ " not supported by target architecture");
+  }
+}
+
 Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
 const CallExpr *E,
 ReturnValueSlot ReturnValue) {
@@ -18883,6 +18900,23 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_sum: {
+// Due to the use of variadic arguments, explicitly retreive argument
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType()}, false);
+Intrinsic::ID IID = getWaveActiveSumIntrinsic(
+getT

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveMax` intrinsic (PR #112991)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112991

>From b18e40978cae2c0d9ba9aeb61cf7294809b75012 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 18 Oct 2024 14:41:57 -0700
Subject: [PATCH 1/2] [HLSL][SPIRV] Implement `WaveActiveMax` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to retain
signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp

  - add test cases to illustrate passes

Note that this defines the dx intrinsics but does not implement
the DirectX lowering to DXIL. This will be implemented in a second pr
when the dependent pr merges.
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 36 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveMax.hlsl   | 48 +
 .../BuiltIns/WaveActiveMax-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  2 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 34 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveMax.ll| 51 ++
 11 files changed, 340 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveMax.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveMax-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveMax.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..ae2068c7a153f7 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveMax : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_max"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLWaveGetLaneIndex : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_get_lane_index"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td 
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index c458a62d9be48c..95093723c2f495 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -9232,6 +9232,9 @@ def err_typecheck_expect_scalar_or_vector : Error<
   "a vector of such type is required">;
 def err_typecheck_expect_any_scalar_or_vector : Error<
   "invalid operand of type %0 where a scalar or vector is required">;
+def err_typecheck_expect_scalar_or_vector_not_type : Error<
+  "invalid operand of type %0 where %1 or "
+  "a vector of such type is not allowed">;
 def err_typecheck_expect_flt_or_vector : Error<
   "invalid operand of type %0 where floating, complex or "
   "a vector of such types is required">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 12f99d9f1178a9..dd17b31f3d7168 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18638,6 +18638,25 @@ static Intrinsic::ID 
getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   return RT.getUDotIntrinsic();
 }
 
+// Return wave active max that corresponds to the QT scalar type
+static Intrinsic::ID getWaveActiveMaxIntrinsic(llvm::Triple::ArchType Arch,
+   CGHLSLRuntime &RT, QualType QT) 
{
+  switch (Arch) {
+  case llvm::Triple::spirv:
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::spv_wave_active_umax;
+return llvm::Intrinsic::spv_wave_active_max;
+  case llvm::Triple::dxil: {
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::dx_wave_active_umax;
+return llvm::Intrinsic::dx_wave_active_max;
+  }
+  default:
+llvm_unreachable("Intrinsic WaveActiveMax"
+ " not supported by target architecture");
+  }
+}
+
 Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
 const CallExpr *E,
 ReturnValueSlot ReturnValue) {
@@ -18883,6 +18902,23 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_max: {
+// Due to the use of variadic arguments, explicitly retreive argument
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{O

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveMax` intrinsic (PR #112991)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic created 
https://github.com/llvm/llvm-project/pull/112991

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to retain 
signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in 
SPIRVInstructionSelector.cpp

  - add test cases to illustrate passes

Note that this defines the dx intrinsics but does not implement the DirectX 
lowering to DXIL. This will be implemented in a second pr when the dependent pr 
merges.

>From b18e40978cae2c0d9ba9aeb61cf7294809b75012 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 18 Oct 2024 14:41:57 -0700
Subject: [PATCH] [HLSL][SPIRV] Implement `WaveActiveMax` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to retain
signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp

  - add test cases to illustrate passes

Note that this defines the dx intrinsics but does not implement
the DirectX lowering to DXIL. This will be implemented in a second pr
when the dependent pr merges.
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 36 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveMax.hlsl   | 48 +
 .../BuiltIns/WaveActiveMax-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  2 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 34 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveMax.ll| 51 ++
 11 files changed, 340 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveMax.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveMax-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveMax.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..ae2068c7a153f7 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveMax : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_max"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLWaveGetLaneIndex : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_get_lane_index"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td 
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index c458a62d9be48c..95093723c2f495 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -9232,6 +9232,9 @@ def err_typecheck_expect_scalar_or_vector : Error<
   "a vector of such type is required">;
 def err_typecheck_expect_any_scalar_or_vector : Error<
   "invalid operand of type %0 where a scalar or vector is required">;
+def err_typecheck_expect_scalar_or_vector_not_type : Error<
+  "invalid operand of type %0 where %1 or "
+  "a vector of such type is not allowed">;
 def err_typecheck_expect_flt_or_vector : Error<
   "invalid operand of type %0 where floating, complex or "
   "a vector of such types is required">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 12f99d9f1178a9..dd17b31f3d7168 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18638,6 +18638,25 @@ static Intrinsic::ID 
getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   return RT.getUDotIntrinsic();
 }
 
+// Return wave active max that corresponds to the QT scalar type
+static Intrinsic::ID getWaveActiveMaxIntrinsic(llvm::Triple::ArchType Arch,
+   CGHLSLRuntime &RT, QualType QT) 
{
+  switch (Arch) {
+  case llvm::Triple::spirv:
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::spv_wave_active_umax;
+return llvm::Intrinsic::spv_wave_active_max;
+  case llvm::Triple::dxil: {
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::dx_wave_active_umax;
+return llvm::Intrinsic::dx_wave_active_max;
+  }
+  default:
+llvm_unreachable("Intrinsic WaveActiveMax"
+ " not supported by target architecture");
+  }
+}
+
 Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
 const CallExpr *E,

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From aa2ae997f66bf6cd43fc36cbe69c6731914ce9f0 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 18 Oct 2024 10:49:18 -0700
Subject: [PATCH] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to retain
signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp

  - add test cases to illustrate passes

Note that this defines the dx intrinsics but does not implement
the DirectX lowering to DXIL. This will implemented in a secon pr when
the dependent pr merges.
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 33 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 11 files changed, 326 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..7205f10b01c2b9 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_sum"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void (...)";
+}
+
 def HLSLWaveGetLaneIndex : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_get_lane_index"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td 
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index c458a62d9be48c..95093723c2f495 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -9232,6 +9232,9 @@ def err_typecheck_expect_scalar_or_vector : Error<
   "a vector of such type is required">;
 def err_typecheck_expect_any_scalar_or_vector : Error<
   "invalid operand of type %0 where a scalar or vector is required">;
+def err_typecheck_expect_scalar_or_vector_not_type : Error<
+  "invalid operand of type %0 where %1 or "
+  "a vector of such type is not allowed">;
 def err_typecheck_expect_flt_or_vector : Error<
   "invalid operand of type %0 where floating, complex or "
   "a vector of such types is required">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 12f99d9f1178a9..46646492b30d31 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18638,6 +18638,23 @@ static Intrinsic::ID 
getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   return RT.getUDotIntrinsic();
 }
 
+// Return wave active sum that corresponds to the QT scalar type
+static Intrinsic::ID getWaveActiveSumIntrinsic(llvm::Triple::ArchType Arch,
+   CGHLSLRuntime &RT, QualType QT) 
{
+  switch (Arch) {
+  case llvm::Triple::spirv:
+return llvm::Intrinsic::spv_wave_active_sum;
+  case llvm::Triple::dxil: {
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::dx_wave_active_usum;
+return llvm::Intrinsic::dx_wave_active_sum;
+  }
+  default:
+llvm_unreachable("Intrinsic WaveActiveSum"
+ " not supported by target architecture");
+  }
+}
+
 Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
 const CallExpr *E,
 ReturnValueSlot ReturnValue) {
@@ -18883,6 +18900,23 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_sum: {
+// Due to the use of variadic arguments, explicitly retreive argument
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType()}, false);
+Intrinsic::ID IID = getWaveActiveSumIntrinsic(
+getTarget().

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-18 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-16 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From 896b3adb4f2820f1ccc651cf98e2bade2f972dc4 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 14:44:09 -0700
Subject: [PATCH 1/2] [NFC][DXIL] add ability to alias to an intrinsic

   - allows an intrinsic to be aliased for another one and provide
 additional arguments when being replaced
---
 llvm/lib/Target/DirectX/DXIL.td|  3 +++
 llvm/lib/Target/DirectX/DXILOpLowering.cpp | 29 +++-
 llvm/utils/TableGen/DXILEmitter.cpp| 31 ++
 3 files changed, 62 insertions(+), 1 deletion(-)

diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 147b32b1ca9903..0f8bab7656b7ec 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -321,6 +321,9 @@ class DXILOp {
 
   // Versioned attributes of operation
   list attributes = [];
+
+  // List of valid aliases
+  list aliases = [];
 }
 
 // Concrete definitions of DXIL Operations
diff --git a/llvm/lib/Target/DirectX/DXILOpLowering.cpp 
b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
index 99df4850872078..60123fa412208a 100644
--- a/llvm/lib/Target/DirectX/DXILOpLowering.cpp
+++ b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
@@ -40,6 +40,19 @@ static bool isVectorArgExpansion(Function &F) {
   return false;
 }
 
+static SmallVector getAliasArgs(Function &F, IRBuilder<> &Builder) {
+  switch (F.getIntrinsicID()) {
+  case Intrinsic::dx_wave_active_sum:
+return getWaveActiveOpArgs<0, true>(F, Builder);
+  case Intrinsic::dx_wave_active_usum:
+return getWaveActiveOpArgs<0, false>(F, Builder);
+  default:
+break;
+  }
+  SmallVector Args;
+  return Args;
+}
+
 static SmallVector populateOperands(Value *Arg, IRBuilder<> &Builder) 
{
   SmallVector ExtractedElements;
   auto *VecArg = dyn_cast(Arg->getType());
@@ -106,7 +119,9 @@ class OpLowerer {
   }
 
   [[nodiscard]]
-  bool replaceFunctionWithOp(Function &F, dxil::OpCode DXILOp) {
+  bool replaceFunctionWithOp(
+  Function &F, dxil::OpCode DXILOp,
+  std::optional> AliasArgs = std::nullopt) {
 bool IsVectorArgExpansion = isVectorArgExpansion(F);
 return replaceFunction(F, [&](CallInst *CI) -> Error {
   SmallVector Args;
@@ -117,6 +132,11 @@ class OpLowerer {
   } else
 Args.append(CI->arg_begin(), CI->arg_end());
 
+  // Append any given alias arguments
+  if (AliasArgs) {
+Args.append(AliasArgs->begin(), AliasArgs->end());
+  }
+
   Expected OpCall =
   OpBuilder.tryCreateOp(DXILOp, Args, CI->getName(), 
F.getReturnType());
   if (Error E = OpCall.takeError())
@@ -476,6 +496,13 @@ class OpLowerer {
   case Intrin: 
\
 HasErrors |= replaceFunctionWithOp(F, OpCode); 
\
 break;
+#include "DXILOperation.inc"
+#define DXIL_OP_INTRINSIC_ALIAS(OpCode, Intrin)
\
+  case Intrin: {   
\
+SmallVector AliasArgs = getAliasArgs(F, OpBuilder.getIRB());  
\
+HasErrors |= replaceFunctionWithOp(F, OpCode, AliasArgs);  
\
+break; 
\
+  }
 #include "DXILOperation.inc"
   case Intrinsic::dx_handle_fromBinding:
 HasErrors |= lowerHandleFromBinding(F);
diff --git a/llvm/utils/TableGen/DXILEmitter.cpp 
b/llvm/utils/TableGen/DXILEmitter.cpp
index 06bf7a0c0a8372..6045256aff23ab 100644
--- a/llvm/utils/TableGen/DXILEmitter.cpp
+++ b/llvm/utils/TableGen/DXILEmitter.cpp
@@ -45,6 +45,7 @@ struct DXILOperationDesc {
   SmallVector AttrRecs;
   StringRef Intrinsic; // The llvm intrinsic map to OpName. Default is "" which
// means no map exists
+  SmallVector IntrinsicAliases;
   SmallVector
   ShaderStages; // shader stages to which this applies, empty for all.
   int OverloadParamIndex; // Index of parameter with overload type.
@@ -168,6 +169,15 @@ DXILOperationDesc::DXILOperationDesc(const Record *R) {
   Intrinsic = DefName.substr(4);
 }
   }
+
+  Recs = R->getValueAsListOfDefs("aliases");
+
+  for (const Record *CR : Recs) {
+auto DefName = CR->getName();
+assert(DefName.starts_with("int_") && "invalid intrinsic name");
+// Remove the int_ from intrinsic name.
+IntrinsicAliases.push_back(DefName.substr(4));
+  }
 }
 
 /// Return a string representation of OverloadKind enum that maps to
@@ -388,6 +398,26 @@ static void 
emitDXILIntrinsicMap(ArrayRef Ops,
   OS << "#endif\n\n";
 }
 
+/// Emit map of DXIL operation to LLVM or DirectX intrinsic alias
+/// \param A vector of DXIL Ops
+/// \param Output stream
+static void emitDXILIntrinsicAliasMap(ArrayRef Ops,
+  raw_ostream &OS) {
+  OS << "#ifdef DXIL_OP_INTRINSIC_ALIAS\n";
+  OS << "\n";
+  for (const auto &

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-16 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From 0843af11c647e839704eded62c65d87ec1b2b730 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:27:12 -0700
Subject: [PATCH 1/3] [NFC][DXIL] Allow extra args to replaceFunctionWithOp

- allow passing down extra arguments when lowering a directx
  intrinsic to its dxil op code

- this allows us to not create a custom intrinsic for the
  targeted intermediate intrinsic that it lowers to and does not
  require us to write a complete custom lowering
---
 llvm/lib/Target/DirectX/DXILOpLowering.cpp | 9 -
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/llvm/lib/Target/DirectX/DXILOpLowering.cpp 
b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
index 99df4850872078..9feadcb79a5e86 100644
--- a/llvm/lib/Target/DirectX/DXILOpLowering.cpp
+++ b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
@@ -106,7 +106,9 @@ class OpLowerer {
   }
 
   [[nodiscard]]
-  bool replaceFunctionWithOp(Function &F, dxil::OpCode DXILOp) {
+  bool replaceFunctionWithOp(
+  Function &F, dxil::OpCode DXILOp,
+  std::optional> ExtraArgs = std::nullopt) {
 bool IsVectorArgExpansion = isVectorArgExpansion(F);
 return replaceFunction(F, [&](CallInst *CI) -> Error {
   SmallVector Args;
@@ -117,6 +119,11 @@ class OpLowerer {
   } else
 Args.append(CI->arg_begin(), CI->arg_end());
 
+  // Append any given alias arguments
+  if (ExtraArgs) {
+Args.append(ExtraArgs->begin(), ExtraArgs->end());
+  }
+
   Expected OpCall =
   OpBuilder.tryCreateOp(DXILOp, Args, CI->getName(), 
F.getReturnType());
   if (Error E = OpCall.takeError())

>From b68a5616c0bf23857ca1b2dba7dc3a80d6934624 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:31:37 -0700
Subject: [PATCH 2/3] [NFC][DXIL] Add `WaveActiveOp` dxil operation

- lowering target for `WaveActive*` ops
---
 llvm/lib/Target/DirectX/DXIL.td | 9 +
 1 file changed, 9 insertions(+)

diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 147b32b1ca9903..fcae23fadf9b94 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -793,6 +793,15 @@ def CreateHandleFromBinding : DXILOp<218, 
createHandleFromBinding> {
   let stages = [Stages];
 }
 
+def WaveActiveOp : DXILOp<119, waveActiveOp> {
+  let Doc = "returns the result of the operation across waves";
+  let arguments = [OverloadTy, Int8Ty, Int8Ty];
+  let result = OverloadTy;
+  let overloads = [Overloads];
+  let stages = [Stages];
+  let attributes = [Attributes];
+}
+
 def WaveIsFirstLane :  DXILOp<110, waveIsFirstLane> {
   let Doc = "returns 1 for the first lane in the wave";
   let LLVMIntrinsic = int_dx_wave_is_first_lane;

>From fc7d25fe93ce03b1e6789eb2b82d33294e1d86b8 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:30:56 -0700
Subject: [PATCH 3/3] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to
retain signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add directx intrinsic expansion to WaveActiveOp in
DXILIntrinsicExpansion.cpp

  - add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 llvm/lib/Target/DirectX/DXILOpLowering.cpp| 22 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 33 +++
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 78 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 13 files changed, 426 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..7205f10b01c2b9 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LANG"> {

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-16 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From 0843af11c647e839704eded62c65d87ec1b2b730 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:27:12 -0700
Subject: [PATCH 1/3] [NFC][DXIL] Allow extra args to replaceFunctionWithOp

- allow passing down extra arguments when lowering a directx
  intrinsic to its dxil op code

- this allows us to not create a custom intrinsic for the
  targeted intermediate intrinsic that it lowers to and does not
  require us to write a complete custom lowering
---
 llvm/lib/Target/DirectX/DXILOpLowering.cpp | 9 -
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/llvm/lib/Target/DirectX/DXILOpLowering.cpp 
b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
index 99df4850872078..9feadcb79a5e86 100644
--- a/llvm/lib/Target/DirectX/DXILOpLowering.cpp
+++ b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
@@ -106,7 +106,9 @@ class OpLowerer {
   }
 
   [[nodiscard]]
-  bool replaceFunctionWithOp(Function &F, dxil::OpCode DXILOp) {
+  bool replaceFunctionWithOp(
+  Function &F, dxil::OpCode DXILOp,
+  std::optional> ExtraArgs = std::nullopt) {
 bool IsVectorArgExpansion = isVectorArgExpansion(F);
 return replaceFunction(F, [&](CallInst *CI) -> Error {
   SmallVector Args;
@@ -117,6 +119,11 @@ class OpLowerer {
   } else
 Args.append(CI->arg_begin(), CI->arg_end());
 
+  // Append any given alias arguments
+  if (ExtraArgs) {
+Args.append(ExtraArgs->begin(), ExtraArgs->end());
+  }
+
   Expected OpCall =
   OpBuilder.tryCreateOp(DXILOp, Args, CI->getName(), 
F.getReturnType());
   if (Error E = OpCall.takeError())

>From b68a5616c0bf23857ca1b2dba7dc3a80d6934624 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:31:37 -0700
Subject: [PATCH 2/3] [NFC][DXIL] Add `WaveActiveOp` dxil operation

- lowering target for `WaveActive*` ops
---
 llvm/lib/Target/DirectX/DXIL.td | 9 +
 1 file changed, 9 insertions(+)

diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 147b32b1ca9903..fcae23fadf9b94 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -793,6 +793,15 @@ def CreateHandleFromBinding : DXILOp<218, 
createHandleFromBinding> {
   let stages = [Stages];
 }
 
+def WaveActiveOp : DXILOp<119, waveActiveOp> {
+  let Doc = "returns the result of the operation across waves";
+  let arguments = [OverloadTy, Int8Ty, Int8Ty];
+  let result = OverloadTy;
+  let overloads = [Overloads];
+  let stages = [Stages];
+  let attributes = [Attributes];
+}
+
 def WaveIsFirstLane :  DXILOp<110, waveIsFirstLane> {
   let Doc = "returns 1 for the first lane in the wave";
   let LLVMIntrinsic = int_dx_wave_is_first_lane;

>From f298a3291cd2b62a1b15862f4e776c32e870b055 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:30:56 -0700
Subject: [PATCH 3/3] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to
retain signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add directx intrinsic expansion to WaveActiveOp in
DXILIntrinsicExpansion.cpp

  - add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 llvm/lib/Target/DirectX/DXILOpLowering.cpp| 17 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 33 +++
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 78 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 13 files changed, 421 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..7205f10b01c2b9 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LANG"> {

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-16 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From 0843af11c647e839704eded62c65d87ec1b2b730 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:27:12 -0700
Subject: [PATCH 1/3] [NFC][DXIL] Allow extra args to replaceFunctionWithOp

- allow passing down extra arguments when lowering a directx
  intrinsic to its dxil op code

- this allows us to not create a custom intrinsic for the
  targeted intermediate intrinsic that it lowers to and does not
  require us to write a complete custom lowering
---
 llvm/lib/Target/DirectX/DXILOpLowering.cpp | 9 -
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/llvm/lib/Target/DirectX/DXILOpLowering.cpp 
b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
index 99df4850872078..9feadcb79a5e86 100644
--- a/llvm/lib/Target/DirectX/DXILOpLowering.cpp
+++ b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
@@ -106,7 +106,9 @@ class OpLowerer {
   }
 
   [[nodiscard]]
-  bool replaceFunctionWithOp(Function &F, dxil::OpCode DXILOp) {
+  bool replaceFunctionWithOp(
+  Function &F, dxil::OpCode DXILOp,
+  std::optional> ExtraArgs = std::nullopt) {
 bool IsVectorArgExpansion = isVectorArgExpansion(F);
 return replaceFunction(F, [&](CallInst *CI) -> Error {
   SmallVector Args;
@@ -117,6 +119,11 @@ class OpLowerer {
   } else
 Args.append(CI->arg_begin(), CI->arg_end());
 
+  // Append any given alias arguments
+  if (ExtraArgs) {
+Args.append(ExtraArgs->begin(), ExtraArgs->end());
+  }
+
   Expected OpCall =
   OpBuilder.tryCreateOp(DXILOp, Args, CI->getName(), 
F.getReturnType());
   if (Error E = OpCall.takeError())

>From b68a5616c0bf23857ca1b2dba7dc3a80d6934624 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:31:37 -0700
Subject: [PATCH 2/3] [NFC][DXIL] Add `WaveActiveOp` dxil operation

- lowering target for `WaveActive*` ops
---
 llvm/lib/Target/DirectX/DXIL.td | 9 +
 1 file changed, 9 insertions(+)

diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 147b32b1ca9903..fcae23fadf9b94 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -793,6 +793,15 @@ def CreateHandleFromBinding : DXILOp<218, 
createHandleFromBinding> {
   let stages = [Stages];
 }
 
+def WaveActiveOp : DXILOp<119, waveActiveOp> {
+  let Doc = "returns the result of the operation across waves";
+  let arguments = [OverloadTy, Int8Ty, Int8Ty];
+  let result = OverloadTy;
+  let overloads = [Overloads];
+  let stages = [Stages];
+  let attributes = [Attributes];
+}
+
 def WaveIsFirstLane :  DXILOp<110, waveIsFirstLane> {
   let Doc = "returns 1 for the first lane in the wave";
   let LLVMIntrinsic = int_dx_wave_is_first_lane;

>From ec4a8deb12e87d9afa61d8529d6477b38f9f5fc4 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:30:56 -0700
Subject: [PATCH 3/3] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to
retain signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add lowering of directx intrinsics to a WaveActiveOp dxil op
in DXILOpLowering.cpp

  - add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 llvm/lib/Target/DirectX/DXILOpLowering.cpp| 22 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 33 +++
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 78 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 13 files changed, 426 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..7205f10b01c2b9 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LAN

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-16 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From 0843af11c647e839704eded62c65d87ec1b2b730 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:27:12 -0700
Subject: [PATCH 1/5] [NFC][DXIL] Allow extra args to replaceFunctionWithOp

- allow passing down extra arguments when lowering a directx
  intrinsic to its dxil op code

- this allows us to not create a custom intrinsic for the
  targeted intermediate intrinsic that it lowers to and does not
  require us to write a complete custom lowering
---
 llvm/lib/Target/DirectX/DXILOpLowering.cpp | 9 -
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/llvm/lib/Target/DirectX/DXILOpLowering.cpp 
b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
index 99df4850872078..9feadcb79a5e86 100644
--- a/llvm/lib/Target/DirectX/DXILOpLowering.cpp
+++ b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
@@ -106,7 +106,9 @@ class OpLowerer {
   }
 
   [[nodiscard]]
-  bool replaceFunctionWithOp(Function &F, dxil::OpCode DXILOp) {
+  bool replaceFunctionWithOp(
+  Function &F, dxil::OpCode DXILOp,
+  std::optional> ExtraArgs = std::nullopt) {
 bool IsVectorArgExpansion = isVectorArgExpansion(F);
 return replaceFunction(F, [&](CallInst *CI) -> Error {
   SmallVector Args;
@@ -117,6 +119,11 @@ class OpLowerer {
   } else
 Args.append(CI->arg_begin(), CI->arg_end());
 
+  // Append any given alias arguments
+  if (ExtraArgs) {
+Args.append(ExtraArgs->begin(), ExtraArgs->end());
+  }
+
   Expected OpCall =
   OpBuilder.tryCreateOp(DXILOp, Args, CI->getName(), 
F.getReturnType());
   if (Error E = OpCall.takeError())

>From b68a5616c0bf23857ca1b2dba7dc3a80d6934624 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:31:37 -0700
Subject: [PATCH 2/5] [NFC][DXIL] Add `WaveActiveOp` dxil operation

- lowering target for `WaveActive*` ops
---
 llvm/lib/Target/DirectX/DXIL.td | 9 +
 1 file changed, 9 insertions(+)

diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 147b32b1ca9903..fcae23fadf9b94 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -793,6 +793,15 @@ def CreateHandleFromBinding : DXILOp<218, 
createHandleFromBinding> {
   let stages = [Stages];
 }
 
+def WaveActiveOp : DXILOp<119, waveActiveOp> {
+  let Doc = "returns the result of the operation across waves";
+  let arguments = [OverloadTy, Int8Ty, Int8Ty];
+  let result = OverloadTy;
+  let overloads = [Overloads];
+  let stages = [Stages];
+  let attributes = [Attributes];
+}
+
 def WaveIsFirstLane :  DXILOp<110, waveIsFirstLane> {
   let Doc = "returns 1 for the first lane in the wave";
   let LLVMIntrinsic = int_dx_wave_is_first_lane;

>From ec4a8deb12e87d9afa61d8529d6477b38f9f5fc4 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:30:56 -0700
Subject: [PATCH 3/5] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to
retain signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add lowering of directx intrinsics to a WaveActiveOp dxil op
in DXILOpLowering.cpp

  - add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 llvm/lib/Target/DirectX/DXILOpLowering.cpp| 22 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 33 +++
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 78 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 13 files changed, 426 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..7205f10b01c2b9 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4749,6 +4749,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LAN

[clang] [llvm] [HLSL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-16 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From 0843af11c647e839704eded62c65d87ec1b2b730 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:27:12 -0700
Subject: [PATCH 1/3] [NFC][DXIL] Allow extra args to replaceFunctionWithOp

- allow passing down extra arguments when lowering a directx
  intrinsic to its dxil op code

- this allows us to not create a custom intrinsic for the
  targeted intermediate intrinsic that it lowers to and does not
  require us to write a complete custom lowering
---
 llvm/lib/Target/DirectX/DXILOpLowering.cpp | 9 -
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/llvm/lib/Target/DirectX/DXILOpLowering.cpp 
b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
index 99df4850872078..9feadcb79a5e86 100644
--- a/llvm/lib/Target/DirectX/DXILOpLowering.cpp
+++ b/llvm/lib/Target/DirectX/DXILOpLowering.cpp
@@ -106,7 +106,9 @@ class OpLowerer {
   }
 
   [[nodiscard]]
-  bool replaceFunctionWithOp(Function &F, dxil::OpCode DXILOp) {
+  bool replaceFunctionWithOp(
+  Function &F, dxil::OpCode DXILOp,
+  std::optional> ExtraArgs = std::nullopt) {
 bool IsVectorArgExpansion = isVectorArgExpansion(F);
 return replaceFunction(F, [&](CallInst *CI) -> Error {
   SmallVector Args;
@@ -117,6 +119,11 @@ class OpLowerer {
   } else
 Args.append(CI->arg_begin(), CI->arg_end());
 
+  // Append any given alias arguments
+  if (ExtraArgs) {
+Args.append(ExtraArgs->begin(), ExtraArgs->end());
+  }
+
   Expected OpCall =
   OpBuilder.tryCreateOp(DXILOp, Args, CI->getName(), 
F.getReturnType());
   if (Error E = OpCall.takeError())

>From b68a5616c0bf23857ca1b2dba7dc3a80d6934624 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:31:37 -0700
Subject: [PATCH 2/3] [NFC][DXIL] Add `WaveActiveOp` dxil operation

- lowering target for `WaveActive*` ops
---
 llvm/lib/Target/DirectX/DXIL.td | 9 +
 1 file changed, 9 insertions(+)

diff --git a/llvm/lib/Target/DirectX/DXIL.td b/llvm/lib/Target/DirectX/DXIL.td
index 147b32b1ca9903..fcae23fadf9b94 100644
--- a/llvm/lib/Target/DirectX/DXIL.td
+++ b/llvm/lib/Target/DirectX/DXIL.td
@@ -793,6 +793,15 @@ def CreateHandleFromBinding : DXILOp<218, 
createHandleFromBinding> {
   let stages = [Stages];
 }
 
+def WaveActiveOp : DXILOp<119, waveActiveOp> {
+  let Doc = "returns the result of the operation across waves";
+  let arguments = [OverloadTy, Int8Ty, Int8Ty];
+  let result = OverloadTy;
+  let overloads = [Overloads];
+  let stages = [Stages];
+  let attributes = [Attributes];
+}
+
 def WaveIsFirstLane :  DXILOp<110, waveIsFirstLane> {
   let Doc = "returns 1 for the first lane in the wave";
   let LLVMIntrinsic = int_dx_wave_is_first_lane;

>From 751f95388a7213ea96682d83fda00aca9fa5e27d Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 16 Oct 2024 15:30:56 -0700
Subject: [PATCH 3/3] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic

  - add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to
retain signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add lowering of directx intrinsics to a WaveActiveOp dxil op
in DXILOpLowering.cpp

  - add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 .../clang/Basic/DiagnosticSemaKinds.td|  3 +
 clang/lib/CodeGen/CGBuiltin.cpp   | 34 +++
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 99 +++
 clang/lib/Sema/SemaHLSL.cpp   | 31 ++
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   | 48 +
 .../BuiltIns/WaveActiveSum-errors.hlsl| 28 ++
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 llvm/lib/Target/DirectX/DXILConstants.h   | 12 +++
 llvm/lib/Target/DirectX/DXILOpLowering.cpp| 24 +
 .../DirectX/DirectXTargetTransformInfo.cpp|  2 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 33 +++
 .../test/CodeGen/DirectX/WaveActiveSum-vec.ll | 34 +++
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 78 +++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll| 41 
 16 files changed, 476 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum-vec.ll
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 382fb6b7a3c031..7205f10b01c2b9 100644
--- a/clang/include/cla

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-10 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111209

>From 398dcb5c1a354c12d4d43af152f6d3e14f001274 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 2 Oct 2024 13:30:33 -0700
Subject: [PATCH 1/2] [HLSL] Implementation the `degrees` intrinsic

- add degrees builtin
- link degrees api in hlsl_intrinsics.h
- add degrees intrinsic to IntrinsicDirectX.td
- add degrees intrinsic to IntrinsicSPIRV.td
- add lowering from clang builtin to dx/spv intrinsics in
  CGBuiltin.cpp
- add semantic checks to SemaHLSL.cpp
- add expansion of directx intrinsic to llvm fmul for DirectX in
  DXILIntrinsicExpansion.cpp
- add mapping to spir-v intrinsic in SPIRVInstructionSelector.cpp

- add test coverage:
- degrees.hlsl -> check hlsl lowering to dx/spv degrees intrinsics
- degrees-errors.hlsl/half-float-only-errors ->
check semantic warnings
- hlsl-intrinsics/degrees.ll -> check lowering of spir-v
degrees intrinsic to SPIR-V backend
- DirectX/degrees.ll -> check expansion and scalarization of
  directx degrees intrinsic to fmul
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 11 +++
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 30 
 clang/lib/Sema/SemaHLSL.cpp   |  1 +
 clang/test/CodeGenHLSL/builtins/degrees.hlsl  | 64 +
 .../SemaHLSL/BuiltIns/degrees-errors.hlsl | 26 +++
 .../BuiltIns/half-float-only-errors.hlsl  |  1 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 12 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp |  2 +
 llvm/test/CodeGen/DirectX/degrees.ll  | 54 +++
 .../CodeGen/SPIRV/hlsl-intrinsics/degrees.ll  | 68 +++
 llvm/test/CodeGen/SPIRV/opencl/degrees.ll | 50 ++
 15 files changed, 328 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/degrees.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/degrees-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/opencl/degrees.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 5b504666365b32..44142298987f12 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4739,6 +4739,12 @@ def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDegrees : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_elementwise_degrees"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_dot"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c864714182e019..0f824fc2c4431a 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18740,6 +18740,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 CGM.getHLSLRuntime().getNormalizeIntrinsic(), ArrayRef{X},
 nullptr, "hlsl.normalize");
   }
+  case Builtin::BI__builtin_hlsl_elementwise_degrees: {
+Value *X = EmitScalarExpr(E->getArg(0));
+
+assert(E->getArg(0)->getType()->hasFloatingRepresentation() &&
+"degree operand must have a float representation");
+
+return Builder.CreateIntrinsic(
+/*ReturnType=*/X->getType(),
+CGM.getHLSLRuntime().getDegreesIntrinsic(),
+ArrayRef{X}, nullptr, "hlsl.degrees");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_frac: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 if (!E->getArg(0)->getType()->hasFloatingRepresentation())
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..966111a9600017 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -74,6 +74,7 @@ class CGHLSLRuntime {
 
   GENERATE_HLSL_INTRINSIC_FUNCTION(All, all)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Any, any)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(Degrees, degrees)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Frac, frac)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Length, length)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Lerp, lerp)
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index d28f204e352de5..07004a48f96128 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -766,6 +766,36 @@ uint64_t3 countbits(uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_elementwise_popcount)
 uint64_t4 countbits(uint64_t4);
 
+//===

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-10 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111209

>From 398dcb5c1a354c12d4d43af152f6d3e14f001274 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 2 Oct 2024 13:30:33 -0700
Subject: [PATCH 1/3] [HLSL] Implementation the `degrees` intrinsic

- add degrees builtin
- link degrees api in hlsl_intrinsics.h
- add degrees intrinsic to IntrinsicDirectX.td
- add degrees intrinsic to IntrinsicSPIRV.td
- add lowering from clang builtin to dx/spv intrinsics in
  CGBuiltin.cpp
- add semantic checks to SemaHLSL.cpp
- add expansion of directx intrinsic to llvm fmul for DirectX in
  DXILIntrinsicExpansion.cpp
- add mapping to spir-v intrinsic in SPIRVInstructionSelector.cpp

- add test coverage:
- degrees.hlsl -> check hlsl lowering to dx/spv degrees intrinsics
- degrees-errors.hlsl/half-float-only-errors ->
check semantic warnings
- hlsl-intrinsics/degrees.ll -> check lowering of spir-v
degrees intrinsic to SPIR-V backend
- DirectX/degrees.ll -> check expansion and scalarization of
  directx degrees intrinsic to fmul
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 11 +++
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 30 
 clang/lib/Sema/SemaHLSL.cpp   |  1 +
 clang/test/CodeGenHLSL/builtins/degrees.hlsl  | 64 +
 .../SemaHLSL/BuiltIns/degrees-errors.hlsl | 26 +++
 .../BuiltIns/half-float-only-errors.hlsl  |  1 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/DirectX/DXILIntrinsicExpansion.cpp | 12 
 .../Target/SPIRV/SPIRVInstructionSelector.cpp |  2 +
 llvm/test/CodeGen/DirectX/degrees.ll  | 54 +++
 .../CodeGen/SPIRV/hlsl-intrinsics/degrees.ll  | 68 +++
 llvm/test/CodeGen/SPIRV/opencl/degrees.ll | 50 ++
 15 files changed, 328 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/degrees.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/degrees-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/degrees.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/opencl/degrees.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 5b504666365b32..44142298987f12 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4739,6 +4739,12 @@ def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDegrees : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_elementwise_degrees"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_dot"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c864714182e019..0f824fc2c4431a 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18740,6 +18740,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 CGM.getHLSLRuntime().getNormalizeIntrinsic(), ArrayRef{X},
 nullptr, "hlsl.normalize");
   }
+  case Builtin::BI__builtin_hlsl_elementwise_degrees: {
+Value *X = EmitScalarExpr(E->getArg(0));
+
+assert(E->getArg(0)->getType()->hasFloatingRepresentation() &&
+"degree operand must have a float representation");
+
+return Builder.CreateIntrinsic(
+/*ReturnType=*/X->getType(),
+CGM.getHLSLRuntime().getDegreesIntrinsic(),
+ArrayRef{X}, nullptr, "hlsl.degrees");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_frac: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 if (!E->getArg(0)->getType()->hasFloatingRepresentation())
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..966111a9600017 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -74,6 +74,7 @@ class CGHLSLRuntime {
 
   GENERATE_HLSL_INTRINSIC_FUNCTION(All, all)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Any, any)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(Degrees, degrees)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Frac, frac)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Length, length)
   GENERATE_HLSL_INTRINSIC_FUNCTION(Lerp, lerp)
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index d28f204e352de5..07004a48f96128 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -766,6 +766,36 @@ uint64_t3 countbits(uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_elementwise_popcount)
 uint64_t4 countbits(uint64_t4);
 
+//===

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic (PR #111010)

2024-10-10 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 01/11] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===-

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 70089645ec5cf62b491a56df96ec46f4328fbc11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/2] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 16 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 155 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..dff56af9282e9d 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,22 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(CGM.CreateRuntimeFunction(FT, name, {}, false, 
true),
+   ArrayRef{OpExpr, OpIndex}, 
"hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===---

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/111010

>From 358ea8278b4e10a094d342aa9b3b1571120a1478 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Thu, 3 Oct 2024 11:43:51 -0700
Subject: [PATCH 1/2] [HLSL] Implement `WaveReadLaneAt` intrinsic

- create a clang built-in in Builtins.td
- add semantic checking in SemaHLSL.cpp
- link the WaveReadLaneAt api in hlsl_intrinsics.h
- add lowering to spirv backend op GroupNonUniformShuffle
  with Scope = 2 (Group) in SPIRVInstructionSelector.cpp

- add tests for HLSL intrinsic lowering to spirv intrinsic in
  WaveReadLaneAt.hlsl
- add tests for sema checks in WaveReadLaneAt-errors.hlsl
- add spir-v backend tests in WaveReadLaneAt.ll
---
 clang/include/clang/Basic/Builtins.td |  6 +++
 clang/lib/CodeGen/CGBuiltin.cpp   | 17 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  7 
 clang/lib/Sema/SemaHLSL.cpp   | 20 ++
 .../CodeGenHLSL/builtins/WaveReadLaneAt.hlsl  | 40 +++
 .../BuiltIns/WaveReadLaneAt-errors.hlsl   | 21 ++
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 15 +++
 .../SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll   | 28 +
 10 files changed, 156 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveReadLaneAt.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveReadLaneAt-errors.hlsl
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveReadLaneAt.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 8090119e512fbb..eec9acd4d27d7d 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4703,6 +4703,12 @@ def HLSLWaveIsFirstLane : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "bool()";
 }
 
+def HLSLWaveReadLaneAt : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_read_lane_at"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void(...)";
+}
+
 def HLSLClamp : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_clamp"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index da3eca73bfb575..2cd851475b6dd9 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18835,6 +18835,23 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveIsFirstLaneIntrinsic();
 return EmitRuntimeCall(Intrinsic::getDeclaration(&CGM.getModule(), ID));
   }
+  case Builtin::BI__builtin_hlsl_wave_read_lane_at: {
+// Due to the use of variadic arguments we must explicitly retreive them 
and
+// create our function type.
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Value *OpIndex = EmitScalarExpr(E->getArg(1));
+llvm::FunctionType *FT = llvm::FunctionType::get(
+OpExpr->getType(), ArrayRef{OpExpr->getType(), OpIndex->getType()},
+false);
+
+// Get overloaded name
+std::string name =
+Intrinsic::getName(CGM.getHLSLRuntime().getWaveReadLaneAtIntrinsic(),
+   ArrayRef{OpExpr->getType()}, &CGM.getModule());
+return EmitRuntimeCall(
+CGM.CreateRuntimeFunction(FT, name, {}, false, true),
+ArrayRef{OpExpr, OpIndex}, "hlsl.wave.read.lane.at");
+  }
   case Builtin::BI__builtin_hlsl_elementwise_sign: {
 Value *Op0 = EmitScalarExpr(E->getArg(0));
 llvm::Type *Xty = Op0->getType();
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index a8aabca7348ffb..a639ce2d784f4a 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -87,6 +87,7 @@ class CGHLSLRuntime {
   GENERATE_HLSL_INTRINSIC_FUNCTION(SDot, sdot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(UDot, udot)
   GENERATE_HLSL_INTRINSIC_FUNCTION(WaveIsFirstLane, wave_is_first_lane)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(WaveReadLaneAt, wave_read_lane_at)
 
   
//===--===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h 
b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
index 810a16d75f0228..a7bdc353ae71bf 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -2015,6 +2015,13 @@ _HLSL_AVAILABILITY(shadermodel, 6.0)
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_is_first_lane)
 __attribute__((convergent)) bool WaveIsFirstLane();
 
+// \brief Returns the value of the expression for the given lane index within
+// the specified wave.
+template 
+_HLSL_AVAILABILITY(shadermodel, 6.0)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_wave_read_lane_at)
+__attribute__((convergent)) T WaveReadLaneAt(T, int32_t);
+
 
//===--

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic ready_for_review 
https://github.com/llvm/llvm-project/pull/111209
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement the `degrees` intrinsic (PR #111209)

2024-10-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111209
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits



@@ -2653,6 +2653,21 @@ bool SPIRVInstructionSelector::selectIntrinsic(Register 
ResVReg,
 .addUse(GR.getSPIRVTypeID(ResType))
 .addUse(GR.getOrCreateConstInt(3, I, IntTy, TII));
   }
+  case Intrinsic::spv_wave_read_lane_at: {
+assert(I.getNumOperands() == 4);
+assert(I.getOperand(2).isReg());
+assert(I.getOperand(3).isReg());
+
+// Defines the execution scope currently 2 for group, see scope table
+SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
+return BuildMI(BB, I, I.getDebugLoc(),
+   TII.get(SPIRV::OpGroupNonUniformShuffle))
+.addDef(ResVReg)
+.addUse(GR.getSPIRVTypeID(ResType))
+.addUse(I.getOperand(2).getReg())
+.addUse(I.getOperand(3).getReg())
+.addUse(GR.getOrCreateConstInt(2, I, IntTy, TII));

inbelic wrote:

IIUC, an HLSL wave corresponds to the SPIR-V workgroup scope, which is denoted 
as 2.

https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic ready_for_review 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL] Implement `WaveReadLaneAt` intrinsic for spirv backend (PR #111010)

2024-10-03 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/111010
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-22 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/113382

>From 68c16dc2e21d3a78a388fdb883551f32b902db11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 18 Oct 2024 15:48:29 -0700
Subject: [PATCH 1/2] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic

  - add codegen for llvm builtin to spirv/directx intrinsic
in CGBuiltin.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add lowering of directx intrinsic to dxil op in DXIL.td

  - add test cases to illustrate passes
  - add test case for semantic analysis
---
 clang/lib/CodeGen/CGBuiltin.cpp   |  7 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 .../builtins/WaveActiveCountBits.hlsl | 22 +++
 .../BuiltIns/WaveActiveCountBits-errors.hlsl  | 18 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 llvm/lib/Target/DirectX/DXIL.td   |  9 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 37 +++
 .../CodeGen/DirectX/WaveActiveCountBits.ll| 10 +
 .../hlsl-intrinsics/WaveActiveCountBits.ll| 19 ++
 10 files changed, 125 insertions(+)
 create mode 100755 clang/test/CodeGenHLSL/builtins/WaveActiveCountBits.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveCountBits-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveCountBits.ll
 create mode 100644 
llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveCountBits.ll

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 28f28c70b5ae52..f178499156eb1f 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18879,6 +18879,13 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_count_bits: {
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveActiveCountBitsIntrinsic();
+return EmitRuntimeCall(
+Intrinsic::getOrInsertDeclaration(&CGM.getModule(), ID),
+ArrayRef{OpExpr});
+  }
   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 ff7df41b5c62e7..ada57b070783c6 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(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/WaveActiveCountBits.hlsl 
b/clang/test/CodeGenHLSL/builtins/WaveActiveCountBits.hlsl
new file mode 100755
index 00..3e1f8fcaace9c2
--- /dev/null
+++ b/clang/test/CodeGenHLSL/builtins/WaveActiveCountBits.hlsl
@@ -0,0 +1,22 @@
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -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 -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.
+
+// CHECK-LABEL: test_bool
+int test_bool(bool expr) {
+  // CHECK-SPIRV: %[[#entry_tok:]] = call token 
@llvm.experimental.convergence.entry()
+  // CHECK-SPIRV:  %[[RET:.*]] = call spir_func i32 
@llvm.spv.wave.active.countbits(i1 %{{.*}}) [ "convergencectrl"(token 
%[[#entry_tok]]) ]
+  // CHECK-DXIL:  %[[RET:.*]] = call i32 @llvm.dx.wave.active.countbits(i1 
%{{.*}})
+  // CHECK:  ret i32 %[[RET]]
+  return WaveActiveCountBits(expr);
+}
+
+// CHECK-DXIL: declare i32 @llvm.dx.wave.active.countbits(i1) #[[#attr:]]
+// CHECK-SPIRV: declare i32 @llvm.spv.wave.active.countbits(i1) #[[#attr:]]
+
+// CHECK: attributes #[[#attr]] = {{{.*}} convergent {{.*}}}
diff --git a/clang/test/SemaHLSL/BuiltIns/WaveActiveCountBits-errors.hlsl 
b/clang/test/SemaHLSL/BuiltIns/WaveActiveCountBits-errors.hlsl
new file mode 100644
index 00..02f45eb30b377a
--- /dev/null
+++ b/clang/test/SemaHLSL/BuiltIns/WaveActiveCountBits-errors.hlsl
@@ -0,0 +1,18 @@
+// RUN: %clang_cc1 -finclude-default-header -triple 
dxil-pc-shadermodel6.6-library %s -emit-llvm-only -disable-llvm-passes -verify
+
+int test_too_few_arg() {
+  return __

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-22 Thread Finn Plummer via cfe-commits


https://github.com/inbelic created 
https://github.com/llvm/llvm-project/pull/113382

  - add codegen for llvm builtin to spirv/directx intrinsic in CGBuiltin.cpp
  - add lowering of spirv intrinsic to spirv backend in 
SPIRVInstructionSelector.cpp
  - add lowering of directx intrinsic to dxil op in DXIL.td

  - add test cases to illustrate passes
  - add test case for semantic analysis

>From 68c16dc2e21d3a78a388fdb883551f32b902db11 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 18 Oct 2024 15:48:29 -0700
Subject: [PATCH] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic

  - add codegen for llvm builtin to spirv/directx intrinsic
in CGBuiltin.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add lowering of directx intrinsic to dxil op in DXIL.td

  - add test cases to illustrate passes
  - add test case for semantic analysis
---
 clang/lib/CodeGen/CGBuiltin.cpp   |  7 
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 .../builtins/WaveActiveCountBits.hlsl | 22 +++
 .../BuiltIns/WaveActiveCountBits-errors.hlsl  | 18 +
 llvm/include/llvm/IR/IntrinsicsDirectX.td |  1 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |  1 +
 llvm/lib/Target/DirectX/DXIL.td   |  9 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp | 37 +++
 .../CodeGen/DirectX/WaveActiveCountBits.ll| 10 +
 .../hlsl-intrinsics/WaveActiveCountBits.ll| 19 ++
 10 files changed, 125 insertions(+)
 create mode 100755 clang/test/CodeGenHLSL/builtins/WaveActiveCountBits.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveCountBits-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveCountBits.ll
 create mode 100644 
llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveCountBits.ll

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 28f28c70b5ae52..f178499156eb1f 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18879,6 +18879,13 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_count_bits: {
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+Intrinsic::ID ID = CGM.getHLSLRuntime().getWaveActiveCountBitsIntrinsic();
+return EmitRuntimeCall(
+Intrinsic::getOrInsertDeclaration(&CGM.getModule(), ID),
+ArrayRef{OpExpr});
+  }
   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 ff7df41b5c62e7..ada57b070783c6 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(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/WaveActiveCountBits.hlsl 
b/clang/test/CodeGenHLSL/builtins/WaveActiveCountBits.hlsl
new file mode 100755
index 00..3e1f8fcaace9c2
--- /dev/null
+++ b/clang/test/CodeGenHLSL/builtins/WaveActiveCountBits.hlsl
@@ -0,0 +1,22 @@
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -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 -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.
+
+// CHECK-LABEL: test_bool
+int test_bool(bool expr) {
+  // CHECK-SPIRV: %[[#entry_tok:]] = call token 
@llvm.experimental.convergence.entry()
+  // CHECK-SPIRV:  %[[RET:.*]] = call spir_func i32 
@llvm.spv.wave.active.countbits(i1 %{{.*}}) [ "convergencectrl"(token 
%[[#entry_tok]]) ]
+  // CHECK-DXIL:  %[[RET:.*]] = call i32 @llvm.dx.wave.active.countbits(i1 
%{{.*}})
+  // CHECK:  ret i32 %[[RET]]
+  return WaveActiveCountBits(expr);
+}
+
+// CHECK-DXIL: declare i32 @llvm.dx.wave.active.countbits(i1) #[[#attr:]]
+// CHECK-SPIRV: declare i32 @llvm.spv.wave.active.countbits(i1) #[[#attr:]]
+
+// CHECK: attributes #[[#attr]] = {{{.*}} convergent {{.*}}}
diff --git a/clang/test/SemaHLSL/BuiltIns/WaveActiveCountBits-errors.hlsl 
b/clang/test/SemaHLSL/BuiltIns/WaveActiveCountBits-errors.hlsl
new file mode 100644
index 0

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-22 Thread Finn Plummer via cfe-commits


https://github.com/inbelic ready_for_review 
https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-22 Thread Finn Plummer via cfe-commits


inbelic wrote:

Failing testcase is unrelated. Will need to rebase on review changes.

https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-22 Thread Finn Plummer via cfe-commits



@@ -1762,6 +1765,36 @@ bool SPIRVInstructionSelector::selectSign(Register 
ResVReg,
   return Result;
 }
 
+bool SPIRVInstructionSelector::selectWaveActiveCountBits(
+Register ResVReg, const SPIRVType *ResType, MachineInstr &I) const {
+  assert(I.getNumOperands() == 3);
+  assert(I.getOperand(2).isReg());
+  MachineBasicBlock &BB = *I.getParent();
+
+  Register BallotReg = MRI->createVirtualRegister(&SPIRV::IDRegClass);
+  SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
+  SPIRVType *BallotType = GR.getOrCreateSPIRVVectorType(IntTy, 4, I, TII);
+
+  bool Result =
+  BuildMI(BB, I, I.getDebugLoc(), TII.get(SPIRV::OpGroupNonUniformBallot))
+  .addDef(BallotReg)
+  .addUse(GR.getSPIRVTypeID(BallotType))
+  .addUse(GR.getOrCreateConstInt(SPIRV::Scope::Subgroup, I, IntTy, 
TII))
+  .addUse(I.getOperand(2).getReg());
+
+  Result |=
+  BuildMI(BB, I, I.getDebugLoc(),
+  TII.get(SPIRV::OpGroupNonUniformBallotBitCount))
+  .addDef(ResVReg)
+  .addUse(GR.getSPIRVTypeID(ResType))
+  .addUse(GR.getOrCreateConstInt(SPIRV::Scope::Subgroup, I, IntTy, 
TII))
+  .addImm(0)

inbelic wrote:

Will update to use the `SPIRV::OperationGroup` enum

https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed` intrinsic (PR #113623)

2024-10-24 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/113623
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-24 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-24 Thread Finn Plummer via cfe-commits



@@ -1762,6 +1765,36 @@ bool SPIRVInstructionSelector::selectSign(Register 
ResVReg,
   return Result;
 }
 
+bool SPIRVInstructionSelector::selectWaveActiveCountBits(
+Register ResVReg, const SPIRVType *ResType, MachineInstr &I) const {
+  assert(I.getNumOperands() == 3);
+  assert(I.getOperand(2).isReg());
+  MachineBasicBlock &BB = *I.getParent();
+
+  Register BallotReg = MRI->createVirtualRegister(&SPIRV::IDRegClass);
+  SPIRVType *IntTy = GR.getOrCreateSPIRVIntegerType(32, I, TII);
+  SPIRVType *BallotType = GR.getOrCreateSPIRVVectorType(IntTy, 4, I, TII);
+
+  bool Result =
+  BuildMI(BB, I, I.getDebugLoc(), TII.get(SPIRV::OpGroupNonUniformBallot))
+  .addDef(BallotReg)
+  .addUse(GR.getSPIRVTypeID(BallotType))
+  .addUse(GR.getOrCreateConstInt(SPIRV::Scope::Subgroup, I, IntTy, 
TII))
+  .addUse(I.getOperand(2).getReg());
+
+  Result |=
+  BuildMI(BB, I, I.getDebugLoc(),
+  TII.get(SPIRV::OpGroupNonUniformBallotBitCount))
+  .addDef(ResVReg)
+  .addUse(GR.getSPIRVTypeID(ResType))
+  .addUse(GR.getOrCreateConstInt(SPIRV::Scope::Subgroup, I, IntTy, 
TII))
+  .addImm(0)

inbelic wrote:

Yep, have it added for when the other review comments are addressed.

https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed` intrinsic (PR #113623)

2024-10-24 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/113623

>From 46822507e0b70e91cf6f4f0295d257978652bfbc Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 23 Oct 2024 22:59:15 +
Subject: [PATCH 1/2] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed`
 intrinsic

- create a clang built-in in Builtins.td
- link dot4add_i8packed in hlsl_intrinsics.h
- add lowering to spirv backend through expansion of operation as
  OPSDot is missing up to SPIRV 1.6 in SPIRVInstructionSelector.cpp
- add dot4add_i8packed intrinsic to IntrinsicsDirectX.td and mapping
  to DXIL.td op Dot4AddI8Packed

- add tests for HLSL intrinsic lowering to dx/spv intrinsic in
  dot4add_i8packed.hlsl
- add tests for sema checks in dot4add_i8packed-errors.hlsl
- add test of spir-v lowering in SPIRV/dot4add_i8packed.ll
- add test to dxil lowering in DirectX/dot4add_i8packed.ll
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 12 ++-
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 10 +++
 .../builtins/dot4add_i8packed.hlsl| 17 
 .../BuiltIns/dot4add_i8packed-errors.hlsl | 28 +++
 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 | 84 +++
 llvm/test/CodeGen/DirectX/dot4add_i8packed.ll | 10 +++
 .../SPIRV/hlsl-intrinsics/dot4add_i8packed.ll | 48 +++
 12 files changed, 227 insertions(+), 1 deletion(-)
 create mode 100644 clang/test/CodeGenHLSL/builtins/dot4add_i8packed.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/dot4add_i8packed-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/dot4add_i8packed.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/dot4add_i8packed.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 90475a361bb8f8..eb6b07e8858602 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4792,6 +4792,12 @@ def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDot4AddI8Packed : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_dot4add_i8packed"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "int(unsigned int, unsigned int, int)";
+}
+
 def HLSLFrac : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_frac"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 28f28c70b5ae52..3eceafff49b776 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18722,7 +18722,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 /*ReturnType=*/T0->getScalarType(),
 getDotProductIntrinsic(CGM.getHLSLRuntime(), VecTy0->getElementType()),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.dot");
-  } break;
+  }
+  case Builtin::BI__builtin_hlsl_dot4add_i8packed: {
+Value *A = EmitScalarExpr(E->getArg(0));
+Value *B = EmitScalarExpr(E->getArg(1));
+Value *C = EmitScalarExpr(E->getArg(2));
+
+Intrinsic::ID ID = CGM.getHLSLRuntime().getDot4AddI8PackedIntrinsic();
+return Builder.CreateIntrinsic(
+/*ReturnType=*/C->getType(), ID,
+ArrayRef{A, B, C}, nullptr, "hlsl.dot4add.i8packed");
+  }
   case Builtin::BI__builtin_hlsl_lerp: {
 Value *X = EmitScalarExpr(E->getArg(0));
 Value *Y = EmitScalarExpr(E->getArg(1));
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index ff7df41b5c62e7..8b1141375106cc 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(Dot4AddI8Packed, dot4add_i8packed)
   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 30dce60b3ff702..d10bfcbeed97ea 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -894,6 +894,16 @@ uint64_t dot(uint64_t3, uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
 uint64_t dot(uint64_t4, uint64_t4);
 
+//===--===//
+// dot4add builtins
+//===--===//
+
+/// \fn int dot4add_i8packed(uint A, uint B, int C)
+
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.4)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot4a

[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed` intrinsic (PR #113623)

2024-10-29 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/113623

>From 81dfa26a941f7a0926a3126fe3ebbb4d2a67cec1 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 23 Oct 2024 22:59:15 +
Subject: [PATCH 1/3] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed`
 intrinsic

- create a clang built-in in Builtins.td
- link dot4add_i8packed in hlsl_intrinsics.h
- add lowering to spirv backend through expansion of operation as
  OPSDot is missing up to SPIRV 1.6 in SPIRVInstructionSelector.cpp
- add dot4add_i8packed intrinsic to IntrinsicsDirectX.td and mapping
  to DXIL.td op Dot4AddI8Packed

- add tests for HLSL intrinsic lowering to dx/spv intrinsic in
  dot4add_i8packed.hlsl
- add tests for sema checks in dot4add_i8packed-errors.hlsl
- add test of spir-v lowering in SPIRV/dot4add_i8packed.ll
- add test to dxil lowering in DirectX/dot4add_i8packed.ll
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 12 ++-
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 10 +++
 .../builtins/dot4add_i8packed.hlsl| 17 
 .../BuiltIns/dot4add_i8packed-errors.hlsl | 28 +++
 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 | 84 +++
 llvm/test/CodeGen/DirectX/dot4add_i8packed.ll | 10 +++
 .../SPIRV/hlsl-intrinsics/dot4add_i8packed.ll | 48 +++
 12 files changed, 227 insertions(+), 1 deletion(-)
 create mode 100644 clang/test/CodeGenHLSL/builtins/dot4add_i8packed.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/dot4add_i8packed-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/dot4add_i8packed.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/dot4add_i8packed.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 90475a361bb8f8..eb6b07e8858602 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4792,6 +4792,12 @@ def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDot4AddI8Packed : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_dot4add_i8packed"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "int(unsigned int, unsigned int, int)";
+}
+
 def HLSLFrac : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_frac"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 28f28c70b5ae52..13ed0f99da9815 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18722,7 +18722,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 /*ReturnType=*/T0->getScalarType(),
 getDotProductIntrinsic(CGM.getHLSLRuntime(), VecTy0->getElementType()),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.dot");
-  } break;
+  }
+  case Builtin::BI__builtin_hlsl_dot4add_i8packed: {
+Value *A = EmitScalarExpr(E->getArg(0));
+Value *B = EmitScalarExpr(E->getArg(1));
+Value *C = EmitScalarExpr(E->getArg(2));
+
+Intrinsic::ID ID = CGM.getHLSLRuntime().getDot4AddI8PackedIntrinsic();
+return Builder.CreateIntrinsic(
+/*ReturnType=*/C->getType(), ID, ArrayRef{A, B, C}, nullptr,
+"hlsl.dot4add.i8packed");
+  }
   case Builtin::BI__builtin_hlsl_lerp: {
 Value *X = EmitScalarExpr(E->getArg(0));
 Value *Y = EmitScalarExpr(E->getArg(1));
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index ff7df41b5c62e7..8b1141375106cc 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(Dot4AddI8Packed, dot4add_i8packed)
   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 30dce60b3ff702..d10bfcbeed97ea 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -894,6 +894,16 @@ uint64_t dot(uint64_t3, uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
 uint64_t dot(uint64_t4, uint64_t4);
 
+//===--===//
+// dot4add builtins
+//===--===//
+
+/// \fn int dot4add_i8packed(uint A, uint B, int C)
+
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.4)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot4a

[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-29 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-29 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/112400
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-10-29 Thread Finn Plummer via cfe-commits



@@ -820,3 +820,12 @@ def WaveGetLaneIndex : DXILOp<111, waveGetLaneIndex> {
   let stages = [Stages];
   let attributes = [Attributes];
 }
+
+def WaveAllBitCount : DXILOp<135, waveAllOp> {
+  let Doc = "returns the count of bits set to 1 across the wave";
+  let LLVMIntrinsic = int_dx_wave_active_countbits;
+  let arguments = [Int1Ty];
+  let result = Int32Ty;
+  let stages = [Stages];
+  let attributes = [Attributes];

inbelic wrote:

@pow2clk Just want to ping to double check these attributes are okay. 
`ReadNone` is not in `hctdb.py` but it would seem to be applicable. Should 
`IsWave` be added now or will we add them all at once in a separate commit?

https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [HLSL][SPIRV] Implement `WaveActiveSum` intrinsic (PR #112400)

2024-10-29 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/112400

>From cb507ccb6abbd39198011f5213371bd22405fb30 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Fri, 18 Oct 2024 10:49:18 -0700
Subject: [PATCH] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic

- add clang builtin to Builtins.td
  - link builtin in hlsl_intrinsics
  - add codegen for spirv intrinsic and two directx intrinsics to retain
signedness information of the operands in CGBuiltin.cpp
  - add semantic analysis in SemaHLSL.cpp
  - add lowering of spirv intrinsic to spirv backend in
SPIRVInstructionSelector.cpp
  - add lowering of directx intrinsics to WaveActiveOp dxil op in
DXIL.td

  - add test cases to illustrate passes
---
 clang/include/clang/Basic/Builtins.td |   6 +
 .../clang/Basic/DiagnosticSemaKinds.td|   3 +
 clang/lib/CodeGen/CGBuiltin.cpp   |  34 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  |  99 
 clang/lib/Sema/SemaHLSL.cpp   |  31 
 .../CodeGenHLSL/builtins/WaveActiveSum.hlsl   |  45 ++
 .../BuiltIns/WaveActiveSum-errors.hlsl|  28 
 llvm/include/llvm/IR/IntrinsicsDirectX.td |   2 +
 llvm/include/llvm/IR/IntrinsicsSPIRV.td   |   1 +
 llvm/lib/Target/DirectX/DXIL.td   |  30 
 .../DirectX/DirectXTargetTransformInfo.cpp|   2 +
 .../Target/SPIRV/SPIRVInstructionSelector.cpp |  30 
 llvm/test/CodeGen/DirectX/WaveActiveSum.ll| 143 ++
 .../SPIRV/hlsl-intrinsics/WaveActiveSum.ll|  41 +
 14 files changed, 495 insertions(+)
 create mode 100644 clang/test/CodeGenHLSL/builtins/WaveActiveSum.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/WaveActiveSum-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/WaveActiveSum.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/WaveActiveSum.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 9bd67e0cefebc3..ee97f0b2326fb8 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4750,6 +4750,12 @@ def HLSLWaveActiveCountBits : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "unsigned int(bool)";
 }
 
+def HLSLWaveActiveSum : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_wave_active_sum"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "void (...)";
+}
+
 def HLSLWaveGetLaneIndex : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_wave_get_lane_index"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td 
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index 34ff49d7238a7f..b84577a38518ff 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -9236,6 +9236,9 @@ def err_typecheck_expect_scalar_or_vector : Error<
   "a vector of such type is required">;
 def err_typecheck_expect_any_scalar_or_vector : Error<
   "invalid operand of type %0 where a scalar or vector is required">;
+def err_typecheck_expect_scalar_or_vector_not_type : Error<
+  "invalid operand of type %0 where %1 or "
+  "a vector of such type is not allowed">;
 def err_typecheck_expect_flt_or_vector : Error<
   "invalid operand of type %0 where floating, complex or "
   "a vector of such types is required">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 65d7f5c54a1913..beb4e66272d44b 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18715,6 +18715,23 @@ static Intrinsic::ID 
getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   return RT.getUDotIntrinsic();
 }
 
+// Return wave active sum that corresponds to the QT scalar type
+static Intrinsic::ID getWaveActiveSumIntrinsic(llvm::Triple::ArchType Arch,
+   CGHLSLRuntime &RT, QualType QT) 
{
+  switch (Arch) {
+  case llvm::Triple::spirv:
+return llvm::Intrinsic::spv_wave_active_sum;
+  case llvm::Triple::dxil: {
+if (QT->isUnsignedIntegerType())
+  return llvm::Intrinsic::dx_wave_active_usum;
+return llvm::Intrinsic::dx_wave_active_sum;
+  }
+  default:
+llvm_unreachable("Intrinsic WaveActiveSum"
+ " not supported by target architecture");
+  }
+}
+
 Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
 const CallExpr *E,
 ReturnValueSlot ReturnValue) {
@@ -18960,6 +18977,23 @@ case Builtin::BI__builtin_hlsl_elementwise_isinf: {
 /*ReturnType=*/Op0->getType(), CGM.getHLSLRuntime().getStepIntrinsic(),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.step");
   }
+  case Builtin::BI__builtin_hlsl_wave_active_sum: {
+// Due to the use of variadic arguments, explicitly retreive argument
+Value *OpExpr = EmitScalarExpr(E->getArg(0));
+llvm::FunctionType *FT = llvm::Func

[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed` intrinsic (PR #113623)

2024-10-29 Thread Finn Plummer via cfe-commits


https://github.com/inbelic updated 
https://github.com/llvm/llvm-project/pull/113623

>From 81dfa26a941f7a0926a3126fe3ebbb4d2a67cec1 Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Wed, 23 Oct 2024 22:59:15 +
Subject: [PATCH 1/4] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed`
 intrinsic

- create a clang built-in in Builtins.td
- link dot4add_i8packed in hlsl_intrinsics.h
- add lowering to spirv backend through expansion of operation as
  OPSDot is missing up to SPIRV 1.6 in SPIRVInstructionSelector.cpp
- add dot4add_i8packed intrinsic to IntrinsicsDirectX.td and mapping
  to DXIL.td op Dot4AddI8Packed

- add tests for HLSL intrinsic lowering to dx/spv intrinsic in
  dot4add_i8packed.hlsl
- add tests for sema checks in dot4add_i8packed-errors.hlsl
- add test of spir-v lowering in SPIRV/dot4add_i8packed.ll
- add test to dxil lowering in DirectX/dot4add_i8packed.ll
---
 clang/include/clang/Basic/Builtins.td |  6 ++
 clang/lib/CodeGen/CGBuiltin.cpp   | 12 ++-
 clang/lib/CodeGen/CGHLSLRuntime.h |  1 +
 clang/lib/Headers/hlsl/hlsl_intrinsics.h  | 10 +++
 .../builtins/dot4add_i8packed.hlsl| 17 
 .../BuiltIns/dot4add_i8packed-errors.hlsl | 28 +++
 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 | 84 +++
 llvm/test/CodeGen/DirectX/dot4add_i8packed.ll | 10 +++
 .../SPIRV/hlsl-intrinsics/dot4add_i8packed.ll | 48 +++
 12 files changed, 227 insertions(+), 1 deletion(-)
 create mode 100644 clang/test/CodeGenHLSL/builtins/dot4add_i8packed.hlsl
 create mode 100644 clang/test/SemaHLSL/BuiltIns/dot4add_i8packed-errors.hlsl
 create mode 100644 llvm/test/CodeGen/DirectX/dot4add_i8packed.ll
 create mode 100644 llvm/test/CodeGen/SPIRV/hlsl-intrinsics/dot4add_i8packed.ll

diff --git a/clang/include/clang/Basic/Builtins.td 
b/clang/include/clang/Basic/Builtins.td
index 90475a361bb8f8..eb6b07e8858602 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -4792,6 +4792,12 @@ def HLSLDotProduct : LangBuiltin<"HLSL_LANG"> {
   let Prototype = "void(...)";
 }
 
+def HLSLDot4AddI8Packed : LangBuiltin<"HLSL_LANG"> {
+  let Spellings = ["__builtin_hlsl_dot4add_i8packed"];
+  let Attributes = [NoThrow, Const];
+  let Prototype = "int(unsigned int, unsigned int, int)";
+}
+
 def HLSLFrac : LangBuiltin<"HLSL_LANG"> {
   let Spellings = ["__builtin_hlsl_elementwise_frac"];
   let Attributes = [NoThrow, Const];
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 28f28c70b5ae52..13ed0f99da9815 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -18722,7 +18722,17 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned 
BuiltinID,
 /*ReturnType=*/T0->getScalarType(),
 getDotProductIntrinsic(CGM.getHLSLRuntime(), VecTy0->getElementType()),
 ArrayRef{Op0, Op1}, nullptr, "hlsl.dot");
-  } break;
+  }
+  case Builtin::BI__builtin_hlsl_dot4add_i8packed: {
+Value *A = EmitScalarExpr(E->getArg(0));
+Value *B = EmitScalarExpr(E->getArg(1));
+Value *C = EmitScalarExpr(E->getArg(2));
+
+Intrinsic::ID ID = CGM.getHLSLRuntime().getDot4AddI8PackedIntrinsic();
+return Builder.CreateIntrinsic(
+/*ReturnType=*/C->getType(), ID, ArrayRef{A, B, C}, nullptr,
+"hlsl.dot4add.i8packed");
+  }
   case Builtin::BI__builtin_hlsl_lerp: {
 Value *X = EmitScalarExpr(E->getArg(0));
 Value *Y = EmitScalarExpr(E->getArg(1));
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h 
b/clang/lib/CodeGen/CGHLSLRuntime.h
index ff7df41b5c62e7..8b1141375106cc 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(Dot4AddI8Packed, dot4add_i8packed)
   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 30dce60b3ff702..d10bfcbeed97ea 100644
--- a/clang/lib/Headers/hlsl/hlsl_intrinsics.h
+++ b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -894,6 +894,16 @@ uint64_t dot(uint64_t3, uint64_t3);
 _HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
 uint64_t dot(uint64_t4, uint64_t4);
 
+//===--===//
+// dot4add builtins
+//===--===//
+
+/// \fn int dot4add_i8packed(uint A, uint B, int C)
+
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.4)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot4a

[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `dot4add_i8packed` intrinsic (PR #113623)

2024-10-29 Thread Finn Plummer via cfe-commits



@@ -1694,6 +1698,84 @@ bool SPIRVInstructionSelector::selectIntegerDot(Register 
ResVReg,
   return Result;
 }
 
+// Since pre-1.6 SPIRV has no DotProductInput4x8BitPacked implementation,
+// extract the elements of the packed inputs, multiply them and add the result
+// to the accumulator.

inbelic wrote:

Okay sounds good thanks. In terms of when the operation is not available, do 
you agree with the current expansion implementation?

https://github.com/llvm/llvm-project/pull/113623
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-11-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic edited 
https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

2024-11-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic deleted 
https://github.com/llvm/llvm-project/pull/113382
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] Inbelic/as double (PR #114847)

2024-11-04 Thread Finn Plummer via cfe-commits


https://github.com/inbelic created 
https://github.com/llvm/llvm-project/pull/114847

None

>From f340a6f0421693bd3489adc1c68983dfae9646dd Mon Sep 17 00:00:00 2001
From: Finn Plummer 
Date: Mon, 4 Nov 2024 17:38:36 +
Subject: [PATCH 1/4] [NFC][Scalarizer][TargetTransformInfo] Add
 `isVectorIntrinsicWithOverloadTypeAtArg`

This changes allows target intrinsic to specify overloaded types.

This change will let us add scalarization for `asdouble`:
---
 llvm/include/llvm/Analysis/TargetTransformInfo.h   | 14 ++
 .../llvm/Analysis/TargetTransformInfoImpl.h|  6 ++
 llvm/include/llvm/CodeGen/BasicTTIImpl.h   |  6 ++
 llvm/lib/Analysis/TargetTransformInfo.cpp  |  6 ++
 .../Target/DirectX/DirectXTargetTransformInfo.cpp  |  8 
 .../Target/DirectX/DirectXTargetTransformInfo.h|  3 +++
 llvm/lib/Transforms/Scalar/Scalarizer.cpp  |  9 ++---
 7 files changed, 49 insertions(+), 3 deletions(-)

diff --git a/llvm/include/llvm/Analysis/TargetTransformInfo.h 
b/llvm/include/llvm/Analysis/TargetTransformInfo.h
index 0459941fe05cdc..796b4011d71c0c 100644
--- a/llvm/include/llvm/Analysis/TargetTransformInfo.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfo.h
@@ -896,6 +896,10 @@ class TargetTransformInfo {
   bool isTargetIntrinsicWithScalarOpAtArg(Intrinsic::ID ID,
   unsigned ScalarOpdIdx) const;
 
+  bool isVectorIntrinsicWithOverloadTypeAtArg(Intrinsic::ID ID,
+  unsigned ScalarOpdIdx,
+  bool Default) const;
+
   /// Estimate the overhead of scalarizing an instruction. Insert and Extract
   /// are set if the demanded result elements need to be inserted and/or
   /// extracted from vectors.
@@ -1969,6 +1973,9 @@ class TargetTransformInfo::Concept {
   virtual bool isTargetIntrinsicTriviallyScalarizable(Intrinsic::ID ID) = 0;
   virtual bool isTargetIntrinsicWithScalarOpAtArg(Intrinsic::ID ID,
   unsigned ScalarOpdIdx) = 0;
+  virtual bool isVectorIntrinsicWithOverloadTypeAtArg(Intrinsic::ID ID,
+  unsigned ScalarOpdIdx,
+  bool Default) = 0;
   virtual InstructionCost getScalarizationOverhead(VectorType *Ty,
const APInt &DemandedElts,
bool Insert, bool Extract,
@@ -2530,6 +2537,13 @@ class TargetTransformInfo::Model final : public 
TargetTransformInfo::Concept {
 return Impl.isTargetIntrinsicWithScalarOpAtArg(ID, ScalarOpdIdx);
   }
 
+  bool isVectorIntrinsicWithOverloadTypeAtArg(Intrinsic::ID ID,
+  unsigned ScalarOpdIdx,
+  bool Default) override {
+return Impl.isVectorIntrinsicWithOverloadTypeAtArg(ID, ScalarOpdIdx,
+   Default);
+  }
+
   InstructionCost getScalarizationOverhead(VectorType *Ty,
const APInt &DemandedElts,
bool Insert, bool Extract,
diff --git a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h 
b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
index dbdfb4d8cdfa32..42d1082cf4d9eb 100644
--- a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
@@ -392,6 +392,12 @@ class TargetTransformInfoImplBase {
 return false;
   }
 
+  bool isVectorIntrinsicWithOverloadTypeAtArg(Intrinsic::ID ID,
+  unsigned ScalarOpdIdx,
+  bool Default) const {
+return Default;
+  }
+
   InstructionCost getScalarizationOverhead(VectorType *Ty,
const APInt &DemandedElts,
bool Insert, bool Extract,
diff --git a/llvm/include/llvm/CodeGen/BasicTTIImpl.h 
b/llvm/include/llvm/CodeGen/BasicTTIImpl.h
index db3b5cddd7c1c3..b2841e778947dd 100644
--- a/llvm/include/llvm/CodeGen/BasicTTIImpl.h
+++ b/llvm/include/llvm/CodeGen/BasicTTIImpl.h
@@ -798,6 +798,12 @@ class BasicTTIImplBase : public 
TargetTransformInfoImplCRTPBase {
 return false;
   }
 
+  bool isVectorIntrinsicWithOverloadTypeAtArg(Intrinsic::ID ID,
+  unsigned ScalarOpdIdx,
+  bool Default) const {
+return Default;
+  }
+
   /// Helper wrapper for the DemandedElts variant of getScalarizationOverhead.
   InstructionCost getScalarizationOverhead(VectorType *InTy, bool Insert,
bool Extract,
diff --git a/llvm/lib/Analysis/TargetTransformInfo.cpp 
b/llvm/lib/Analysis/TargetTransformInfo.cpp
index a47462b61e03b2

1 2 3 4 >

1 - 100 of 388 matches

Mail list logo