[MLIR][OpenMP] Lowering aligned clause to LLVM IR for SIMD directive #119536

kaviya2510 · 2024-12-11T10:25:36Z

Added lowering support for aligned clause to LLVM IR

github-actions · 2024-12-11T10:25:58Z

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

llvmbot · 2024-12-11T10:26:33Z

@llvm/pr-subscribers-mlir
@llvm/pr-subscribers-flang-openmp

@llvm/pr-subscribers-flang-fir-hlfir

Author: Kaviya Rajendiran (kaviya2510)

Changes

Added lowering support for aligned clause to LLVM IR

Full diff: https://github.com/llvm/llvm-project/pull/119536.diff

7 Files Affected:

(modified) clang/test/OpenMP/irbuilder_simd_aligned.cpp (+3-3)
(modified) flang/lib/Lower/OpenMP/ClauseProcessor.cpp (+2)
(modified) llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp (+2-1)
(modified) llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp (+4-4)
(modified) mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp (+18-5)
(added) mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir (+60)
(modified) mlir/test/Target/LLVMIR/openmp-todo.mlir (-12)

diff --git a/clang/test/OpenMP/irbuilder_simd_aligned.cpp b/clang/test/OpenMP/irbuilder_simd_aligned.cpp
index 1c3dc49b717ed4..721fde6d954958 100644
--- a/clang/test/OpenMP/irbuilder_simd_aligned.cpp
+++ b/clang/test/OpenMP/irbuilder_simd_aligned.cpp
@@ -70,8 +70,11 @@ void simple(float *a, float *b, int *c) {
 // CHECK-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP3:![0-9]+]]
 // CHECK:       for.end:
 // CHECK-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[A_ADDR]], align 8
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP4]], i64 128) ]
 // CHECK-NEXT:    [[TMP5:%.*]] = load ptr, ptr [[P]], align 8
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP5]], i64 64) ]
 // CHECK-NEXT:    [[ARRAYDECAY:%.*]] = getelementptr inbounds [32 x i32], ptr [[D]], i64 0, i64 0
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[ARRAYDECAY]], i64 16) ]
 // CHECK-NEXT:    store i32 3, ptr [[I1]], align 4
 // CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
 // CHECK-NEXT:    store ptr [[I1]], ptr [[TMP6]], align 8
@@ -82,9 +85,6 @@ void simple(float *a, float *b, int *c) {
 // CHECK-NEXT:    [[DOTCOUNT:%.*]] = load i32, ptr [[DOTCOUNT_ADDR]], align 4
 // CHECK-NEXT:    br label [[OMP_LOOP_PREHEADER:%.*]]
 // CHECK:       omp_loop.preheader:
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP4]], i64 128) ]
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP5]], i64 64) ]
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[ARRAYDECAY]], i64 16) ]
 // CHECK-NEXT:    br label [[OMP_LOOP_HEADER:%.*]]
 // CHECK:       omp_loop.header:
 // CHECK-NEXT:    [[OMP_LOOP_IV:%.*]] = phi i32 [ 0, [[OMP_LOOP_PREHEADER]] ], [ [[OMP_LOOP_NEXT:%.*]], [[OMP_LOOP_INC:%.*]] ]
diff --git a/flang/lib/Lower/OpenMP/ClauseProcessor.cpp b/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
index 48c559a78b9bc4..1b226482a8c13c 100644
--- a/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
+++ b/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
@@ -606,6 +606,8 @@ addAlignedClause(lower::AbstractConverter &converter,
   // Do not generate alignment assumption if alignment is less than or equal to
   // 0.
   if (alignment > 0) {
+    // alignment value must be power of 2
+    assert((alignment & (alignment - 1)) == 0 && "alignment is not power of 2");
     auto &objects = std::get<omp::ObjectList>(clause.t);
     if (!objects.empty())
       genObjectList(objects, converter, alignedVars);
diff --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
index 21004e6a15d495..2fb53609e84397 100644
--- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
+++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
@@ -5302,10 +5302,11 @@ void OpenMPIRBuilder::applySimd(CanonicalLoopInfo *CanonicalLoop,
   Loop *L = LI.getLoopFor(CanonicalLoop->getHeader());
   if (AlignedVars.size()) {
     InsertPointTy IP = Builder.saveIP();
-    Builder.SetInsertPoint(CanonicalLoop->getPreheader()->getTerminator());
     for (auto &AlignedItem : AlignedVars) {
       Value *AlignedPtr = AlignedItem.first;
       Value *Alignment = AlignedItem.second;
+      Instruction *loadInst = dyn_cast<Instruction>(AlignedPtr);
+      Builder.SetInsertPoint(loadInst->getNextNode());
       Builder.CreateAlignmentAssumption(F->getDataLayout(),
                                         AlignedPtr, Alignment);
     }
diff --git a/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp b/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
index 630cd03c688012..4ec1b281d4d2d3 100644
--- a/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
+++ b/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
@@ -1993,6 +1993,7 @@ TEST_F(OpenMPIRBuilderTest, ApplySimdCustomAligned) {
   OpenMPIRBuilder OMPBuilder(*M);
   IRBuilder<> Builder(BB);
   const int AlignmentValue = 32;
+  llvm::BasicBlock *sourceBlock = Builder.GetInsertBlock();
   AllocaInst *Alloc1 =
       Builder.CreateAlloca(Builder.getPtrTy(), Builder.getInt64(1));
   LoadInst *Load1 = Builder.CreateLoad(Alloc1->getAllocatedType(), Alloc1);
@@ -2031,13 +2032,12 @@ TEST_F(OpenMPIRBuilderTest, ApplySimdCustomAligned) {
 
   // Check if number of assumption instructions is equal to number of aligned
   // variables
-  BasicBlock *LoopPreheader = CLI->getPreheader();
-  size_t NumAssummptionCallsInPreheader = count_if(
-      *LoopPreheader, [](Instruction &I) { return isa<AssumeInst>(I); });
+  size_t NumAssummptionCallsInPreheader =
+      count_if(*sourceBlock, [](Instruction &I) { return isa<AssumeInst>(I); });
   EXPECT_EQ(NumAssummptionCallsInPreheader, AlignedVars.size());
 
   // Check if variables are correctly aligned
-  for (Instruction &Instr : *LoopPreheader) {
+  for (Instruction &Instr : *sourceBlock) {
     if (!isa<AssumeInst>(Instr))
       continue;
     AssumeInst *AssumeInstruction = cast<AssumeInst>(&Instr);
diff --git a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
index 49fe509800491a..b4ccb77f45a42c 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
@@ -150,10 +150,6 @@ static LogicalResult checkImplementationStatus(Operation &op) {
                           << " operation";
   };
 
-  auto checkAligned = [&todo](auto op, LogicalResult &result) {
-    if (!op.getAlignedVars().empty() || op.getAlignments())
-      result = todo("aligned");
-  };
   auto checkAllocate = [&todo](auto op, LogicalResult &result) {
     if (!op.getAllocateVars().empty() || !op.getAllocatorVars().empty())
       result = todo("allocate");
@@ -271,7 +267,6 @@ static LogicalResult checkImplementationStatus(Operation &op) {
       })
       .Case([&](omp::ParallelOp op) { checkAllocate(op, result); })
       .Case([&](omp::SimdOp op) {
-        checkAligned(op, result);
         checkLinear(op, result);
         checkNontemporal(op, result);
         checkPrivate(op, result);
@@ -2257,6 +2252,24 @@ convertOmpSimd(Operation &opInst, llvm::IRBuilderBase &builder,
 
   llvm::MapVector<llvm::Value *, llvm::Value *> alignedVars;
   llvm::omp::OrderKind order = convertOrderKind(simdOp.getOrder());
+  llvm::BasicBlock *sourceBlock = builder.GetInsertBlock();
+  std::optional<ArrayAttr> alignmentValues = simdOp.getAlignments();
+  mlir::OperandRange operands = simdOp.getAlignedVars();
+  for (size_t i = 0; i < operands.size(); ++i) {
+    llvm::Value *alignment = nullptr;
+    llvm::Value *llvmVal = moduleTranslation.lookupValue(operands[i]);
+    llvm::Type *ty = llvmVal->getType();
+    if (auto intAttr = llvm::dyn_cast<IntegerAttr>((*alignmentValues)[i])) {
+      alignment = builder.getInt64(intAttr.getInt());
+      assert(ty->isPointerTy() && "Invalid type for aligned variable");
+      assert(alignment && "Invalid alignment value");
+      auto curInsert = builder.saveIP();
+      builder.SetInsertPoint(sourceBlock->getTerminator());
+      llvmVal = builder.CreateLoad(ty, llvmVal);
+      builder.restoreIP(curInsert);
+      alignedVars[llvmVal] = alignment;
+    }
+  }
   ompBuilder->applySimd(loopInfo, alignedVars,
                         simdOp.getIfExpr()
                             ? moduleTranslation.lookupValue(simdOp.getIfExpr())
diff --git a/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir b/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir
new file mode 100644
index 00000000000000..234604e4b664ad
--- /dev/null
+++ b/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir
@@ -0,0 +1,60 @@
+// RUN: mlir-translate -mlir-to-llvmir -split-input-file %s | FileCheck %s
+
+//CHECK-LABEL: define void @_QPsimd_aligned_pointer() {
+//CHECK:   %[[A_PTR:.*]] = alloca { ptr, i64, i32, i8, i8, i8, i8 }, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_PTR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_pointer() {
+  %1 = llvm.mlir.constant(1 : i64) : i64
+  %2 = llvm.alloca %1 x !llvm.struct<(ptr, i64, i32, i8, i8, i8, i8)> {bindc_name = "x"} : (i64) -> !llvm.ptr
+  %3 = llvm.alloca %1 x i32 {bindc_name = "i", pinned} : (i64) -> !llvm.ptr
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  %5 = llvm.mlir.constant(10 : i32) : i32
+  %6 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%2 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%4) to (%5) inclusive step (%6) {
+      llvm.store %arg0, %3 : i32, !llvm.ptr
+      omp.yield
+    }
+  }
+  llvm.return
+}
+
+//CHECK-LABEL: define void @_QPsimd_aligned_cptr() {
+//CHECK:   %[[A_CPTR:.*]] = alloca %_QM__fortran_builtinsT__builtin_c_ptr, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_CPTR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_cptr() {
+  %0 = llvm.mlir.constant(1 : i64) : i64
+  %1 = llvm.alloca %0 x !llvm.struct<"_QM__fortran_builtinsT__builtin_c_ptr", (i64)> {bindc_name = "a"} : (i64) -> !llvm.ptr
+  %2 = llvm.mlir.constant(1 : i64) : i64
+  %3 = llvm.alloca %2 x i32 {bindc_name = "i", pinned} : (i64) -> !llvm.ptr
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  %5 = llvm.mlir.constant(10 : i32) : i32
+  %6 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%1 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%4) to (%5) inclusive step (%6) {
+      llvm.store %arg0, %3 : i32, !llvm.ptr
+      omp.yield
+    }
+  }
+  llvm.return
+}
+
+//CHECK-LABEL: define void @_QPsimd_aligned_allocatable() {
+//CHECK:   %[[A_ADDR:.*]] = alloca { ptr, i64, i32, i8, i8, i8, i8, [1 x [3 x i64]] }, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_ADDR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_allocatable() {
+  %0 = llvm.mlir.constant(1 : i64) : i64
+  %1 = llvm.alloca %0 x !llvm.struct<(ptr, i64, i32, i8, i8, i8, i8, array<1 x array<3 x i64>>)> {bindc_name = "a"} : (i64) -> !llvm.ptr
+  %2 = llvm.mlir.constant(1 : i32) : i32
+  %3 = llvm.mlir.constant(10 : i32) : i32
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%1 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%2) to (%3) inclusive step (%4) {
+      omp.yield
+    }
+  }
+  llvm.return
+}
diff --git a/mlir/test/Target/LLVMIR/openmp-todo.mlir b/mlir/test/Target/LLVMIR/openmp-todo.mlir
index c1e0014d1f571f..010f6048706280 100644
--- a/mlir/test/Target/LLVMIR/openmp-todo.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-todo.mlir
@@ -127,18 +127,6 @@ llvm.func @sections_private(%x : !llvm.ptr) {
   llvm.return
 }
 
-// -----
-
-llvm.func @simd_aligned(%lb : i32, %ub : i32, %step : i32, %x : !llvm.ptr) {
-  // expected-error@below {{not yet implemented: Unhandled clause aligned in omp.simd operation}}
-  // expected-error@below {{LLVM Translation failed for operation: omp.simd}}
-  omp.simd aligned(%x : !llvm.ptr -> 32) {
-    omp.loop_nest (%iv) : i32 = (%lb) to (%ub) step (%step) {
-      omp.yield
-    }
-  }
-  llvm.return
-}
 
 // -----

llvmbot · 2024-12-11T10:26:33Z

@llvm/pr-subscribers-clang

Author: Kaviya Rajendiran (kaviya2510)

Changes

Added lowering support for aligned clause to LLVM IR

Full diff: https://github.com/llvm/llvm-project/pull/119536.diff

7 Files Affected:

(modified) clang/test/OpenMP/irbuilder_simd_aligned.cpp (+3-3)
(modified) flang/lib/Lower/OpenMP/ClauseProcessor.cpp (+2)
(modified) llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp (+2-1)
(modified) llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp (+4-4)
(modified) mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp (+18-5)
(added) mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir (+60)
(modified) mlir/test/Target/LLVMIR/openmp-todo.mlir (-12)

diff --git a/clang/test/OpenMP/irbuilder_simd_aligned.cpp b/clang/test/OpenMP/irbuilder_simd_aligned.cpp
index 1c3dc49b717ed4..721fde6d954958 100644
--- a/clang/test/OpenMP/irbuilder_simd_aligned.cpp
+++ b/clang/test/OpenMP/irbuilder_simd_aligned.cpp
@@ -70,8 +70,11 @@ void simple(float *a, float *b, int *c) {
 // CHECK-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP3:![0-9]+]]
 // CHECK:       for.end:
 // CHECK-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[A_ADDR]], align 8
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP4]], i64 128) ]
 // CHECK-NEXT:    [[TMP5:%.*]] = load ptr, ptr [[P]], align 8
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP5]], i64 64) ]
 // CHECK-NEXT:    [[ARRAYDECAY:%.*]] = getelementptr inbounds [32 x i32], ptr [[D]], i64 0, i64 0
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[ARRAYDECAY]], i64 16) ]
 // CHECK-NEXT:    store i32 3, ptr [[I1]], align 4
 // CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
 // CHECK-NEXT:    store ptr [[I1]], ptr [[TMP6]], align 8
@@ -82,9 +85,6 @@ void simple(float *a, float *b, int *c) {
 // CHECK-NEXT:    [[DOTCOUNT:%.*]] = load i32, ptr [[DOTCOUNT_ADDR]], align 4
 // CHECK-NEXT:    br label [[OMP_LOOP_PREHEADER:%.*]]
 // CHECK:       omp_loop.preheader:
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP4]], i64 128) ]
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP5]], i64 64) ]
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[ARRAYDECAY]], i64 16) ]
 // CHECK-NEXT:    br label [[OMP_LOOP_HEADER:%.*]]
 // CHECK:       omp_loop.header:
 // CHECK-NEXT:    [[OMP_LOOP_IV:%.*]] = phi i32 [ 0, [[OMP_LOOP_PREHEADER]] ], [ [[OMP_LOOP_NEXT:%.*]], [[OMP_LOOP_INC:%.*]] ]
diff --git a/flang/lib/Lower/OpenMP/ClauseProcessor.cpp b/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
index 48c559a78b9bc4..1b226482a8c13c 100644
--- a/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
+++ b/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
@@ -606,6 +606,8 @@ addAlignedClause(lower::AbstractConverter &converter,
   // Do not generate alignment assumption if alignment is less than or equal to
   // 0.
   if (alignment > 0) {
+    // alignment value must be power of 2
+    assert((alignment & (alignment - 1)) == 0 && "alignment is not power of 2");
     auto &objects = std::get<omp::ObjectList>(clause.t);
     if (!objects.empty())
       genObjectList(objects, converter, alignedVars);
diff --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
index 21004e6a15d495..2fb53609e84397 100644
--- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
+++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
@@ -5302,10 +5302,11 @@ void OpenMPIRBuilder::applySimd(CanonicalLoopInfo *CanonicalLoop,
   Loop *L = LI.getLoopFor(CanonicalLoop->getHeader());
   if (AlignedVars.size()) {
     InsertPointTy IP = Builder.saveIP();
-    Builder.SetInsertPoint(CanonicalLoop->getPreheader()->getTerminator());
     for (auto &AlignedItem : AlignedVars) {
       Value *AlignedPtr = AlignedItem.first;
       Value *Alignment = AlignedItem.second;
+      Instruction *loadInst = dyn_cast<Instruction>(AlignedPtr);
+      Builder.SetInsertPoint(loadInst->getNextNode());
       Builder.CreateAlignmentAssumption(F->getDataLayout(),
                                         AlignedPtr, Alignment);
     }
diff --git a/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp b/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
index 630cd03c688012..4ec1b281d4d2d3 100644
--- a/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
+++ b/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
@@ -1993,6 +1993,7 @@ TEST_F(OpenMPIRBuilderTest, ApplySimdCustomAligned) {
   OpenMPIRBuilder OMPBuilder(*M);
   IRBuilder<> Builder(BB);
   const int AlignmentValue = 32;
+  llvm::BasicBlock *sourceBlock = Builder.GetInsertBlock();
   AllocaInst *Alloc1 =
       Builder.CreateAlloca(Builder.getPtrTy(), Builder.getInt64(1));
   LoadInst *Load1 = Builder.CreateLoad(Alloc1->getAllocatedType(), Alloc1);
@@ -2031,13 +2032,12 @@ TEST_F(OpenMPIRBuilderTest, ApplySimdCustomAligned) {
 
   // Check if number of assumption instructions is equal to number of aligned
   // variables
-  BasicBlock *LoopPreheader = CLI->getPreheader();
-  size_t NumAssummptionCallsInPreheader = count_if(
-      *LoopPreheader, [](Instruction &I) { return isa<AssumeInst>(I); });
+  size_t NumAssummptionCallsInPreheader =
+      count_if(*sourceBlock, [](Instruction &I) { return isa<AssumeInst>(I); });
   EXPECT_EQ(NumAssummptionCallsInPreheader, AlignedVars.size());
 
   // Check if variables are correctly aligned
-  for (Instruction &Instr : *LoopPreheader) {
+  for (Instruction &Instr : *sourceBlock) {
     if (!isa<AssumeInst>(Instr))
       continue;
     AssumeInst *AssumeInstruction = cast<AssumeInst>(&Instr);
diff --git a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
index 49fe509800491a..b4ccb77f45a42c 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
@@ -150,10 +150,6 @@ static LogicalResult checkImplementationStatus(Operation &op) {
                           << " operation";
   };
 
-  auto checkAligned = [&todo](auto op, LogicalResult &result) {
-    if (!op.getAlignedVars().empty() || op.getAlignments())
-      result = todo("aligned");
-  };
   auto checkAllocate = [&todo](auto op, LogicalResult &result) {
     if (!op.getAllocateVars().empty() || !op.getAllocatorVars().empty())
       result = todo("allocate");
@@ -271,7 +267,6 @@ static LogicalResult checkImplementationStatus(Operation &op) {
       })
       .Case([&](omp::ParallelOp op) { checkAllocate(op, result); })
       .Case([&](omp::SimdOp op) {
-        checkAligned(op, result);
         checkLinear(op, result);
         checkNontemporal(op, result);
         checkPrivate(op, result);
@@ -2257,6 +2252,24 @@ convertOmpSimd(Operation &opInst, llvm::IRBuilderBase &builder,
 
   llvm::MapVector<llvm::Value *, llvm::Value *> alignedVars;
   llvm::omp::OrderKind order = convertOrderKind(simdOp.getOrder());
+  llvm::BasicBlock *sourceBlock = builder.GetInsertBlock();
+  std::optional<ArrayAttr> alignmentValues = simdOp.getAlignments();
+  mlir::OperandRange operands = simdOp.getAlignedVars();
+  for (size_t i = 0; i < operands.size(); ++i) {
+    llvm::Value *alignment = nullptr;
+    llvm::Value *llvmVal = moduleTranslation.lookupValue(operands[i]);
+    llvm::Type *ty = llvmVal->getType();
+    if (auto intAttr = llvm::dyn_cast<IntegerAttr>((*alignmentValues)[i])) {
+      alignment = builder.getInt64(intAttr.getInt());
+      assert(ty->isPointerTy() && "Invalid type for aligned variable");
+      assert(alignment && "Invalid alignment value");
+      auto curInsert = builder.saveIP();
+      builder.SetInsertPoint(sourceBlock->getTerminator());
+      llvmVal = builder.CreateLoad(ty, llvmVal);
+      builder.restoreIP(curInsert);
+      alignedVars[llvmVal] = alignment;
+    }
+  }
   ompBuilder->applySimd(loopInfo, alignedVars,
                         simdOp.getIfExpr()
                             ? moduleTranslation.lookupValue(simdOp.getIfExpr())
diff --git a/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir b/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir
new file mode 100644
index 00000000000000..234604e4b664ad
--- /dev/null
+++ b/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir
@@ -0,0 +1,60 @@
+// RUN: mlir-translate -mlir-to-llvmir -split-input-file %s | FileCheck %s
+
+//CHECK-LABEL: define void @_QPsimd_aligned_pointer() {
+//CHECK:   %[[A_PTR:.*]] = alloca { ptr, i64, i32, i8, i8, i8, i8 }, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_PTR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_pointer() {
+  %1 = llvm.mlir.constant(1 : i64) : i64
+  %2 = llvm.alloca %1 x !llvm.struct<(ptr, i64, i32, i8, i8, i8, i8)> {bindc_name = "x"} : (i64) -> !llvm.ptr
+  %3 = llvm.alloca %1 x i32 {bindc_name = "i", pinned} : (i64) -> !llvm.ptr
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  %5 = llvm.mlir.constant(10 : i32) : i32
+  %6 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%2 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%4) to (%5) inclusive step (%6) {
+      llvm.store %arg0, %3 : i32, !llvm.ptr
+      omp.yield
+    }
+  }
+  llvm.return
+}
+
+//CHECK-LABEL: define void @_QPsimd_aligned_cptr() {
+//CHECK:   %[[A_CPTR:.*]] = alloca %_QM__fortran_builtinsT__builtin_c_ptr, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_CPTR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_cptr() {
+  %0 = llvm.mlir.constant(1 : i64) : i64
+  %1 = llvm.alloca %0 x !llvm.struct<"_QM__fortran_builtinsT__builtin_c_ptr", (i64)> {bindc_name = "a"} : (i64) -> !llvm.ptr
+  %2 = llvm.mlir.constant(1 : i64) : i64
+  %3 = llvm.alloca %2 x i32 {bindc_name = "i", pinned} : (i64) -> !llvm.ptr
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  %5 = llvm.mlir.constant(10 : i32) : i32
+  %6 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%1 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%4) to (%5) inclusive step (%6) {
+      llvm.store %arg0, %3 : i32, !llvm.ptr
+      omp.yield
+    }
+  }
+  llvm.return
+}
+
+//CHECK-LABEL: define void @_QPsimd_aligned_allocatable() {
+//CHECK:   %[[A_ADDR:.*]] = alloca { ptr, i64, i32, i8, i8, i8, i8, [1 x [3 x i64]] }, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_ADDR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_allocatable() {
+  %0 = llvm.mlir.constant(1 : i64) : i64
+  %1 = llvm.alloca %0 x !llvm.struct<(ptr, i64, i32, i8, i8, i8, i8, array<1 x array<3 x i64>>)> {bindc_name = "a"} : (i64) -> !llvm.ptr
+  %2 = llvm.mlir.constant(1 : i32) : i32
+  %3 = llvm.mlir.constant(10 : i32) : i32
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%1 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%2) to (%3) inclusive step (%4) {
+      omp.yield
+    }
+  }
+  llvm.return
+}
diff --git a/mlir/test/Target/LLVMIR/openmp-todo.mlir b/mlir/test/Target/LLVMIR/openmp-todo.mlir
index c1e0014d1f571f..010f6048706280 100644
--- a/mlir/test/Target/LLVMIR/openmp-todo.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-todo.mlir
@@ -127,18 +127,6 @@ llvm.func @sections_private(%x : !llvm.ptr) {
   llvm.return
 }
 
-// -----
-
-llvm.func @simd_aligned(%lb : i32, %ub : i32, %step : i32, %x : !llvm.ptr) {
-  // expected-error@below {{not yet implemented: Unhandled clause aligned in omp.simd operation}}
-  // expected-error@below {{LLVM Translation failed for operation: omp.simd}}
-  omp.simd aligned(%x : !llvm.ptr -> 32) {
-    omp.loop_nest (%iv) : i32 = (%lb) to (%ub) step (%step) {
-      omp.yield
-    }
-  }
-  llvm.return
-}
 
 // -----

llvmbot · 2024-12-11T10:26:33Z

@llvm/pr-subscribers-mlir-llvm

Author: Kaviya Rajendiran (kaviya2510)

Changes

Added lowering support for aligned clause to LLVM IR

Full diff: https://github.com/llvm/llvm-project/pull/119536.diff

7 Files Affected:

(modified) clang/test/OpenMP/irbuilder_simd_aligned.cpp (+3-3)
(modified) flang/lib/Lower/OpenMP/ClauseProcessor.cpp (+2)
(modified) llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp (+2-1)
(modified) llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp (+4-4)
(modified) mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp (+18-5)
(added) mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir (+60)
(modified) mlir/test/Target/LLVMIR/openmp-todo.mlir (-12)

diff --git a/clang/test/OpenMP/irbuilder_simd_aligned.cpp b/clang/test/OpenMP/irbuilder_simd_aligned.cpp
index 1c3dc49b717ed4..721fde6d954958 100644
--- a/clang/test/OpenMP/irbuilder_simd_aligned.cpp
+++ b/clang/test/OpenMP/irbuilder_simd_aligned.cpp
@@ -70,8 +70,11 @@ void simple(float *a, float *b, int *c) {
 // CHECK-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP3:![0-9]+]]
 // CHECK:       for.end:
 // CHECK-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[A_ADDR]], align 8
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP4]], i64 128) ]
 // CHECK-NEXT:    [[TMP5:%.*]] = load ptr, ptr [[P]], align 8
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP5]], i64 64) ]
 // CHECK-NEXT:    [[ARRAYDECAY:%.*]] = getelementptr inbounds [32 x i32], ptr [[D]], i64 0, i64 0
+// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[ARRAYDECAY]], i64 16) ]
 // CHECK-NEXT:    store i32 3, ptr [[I1]], align 4
 // CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
 // CHECK-NEXT:    store ptr [[I1]], ptr [[TMP6]], align 8
@@ -82,9 +85,6 @@ void simple(float *a, float *b, int *c) {
 // CHECK-NEXT:    [[DOTCOUNT:%.*]] = load i32, ptr [[DOTCOUNT_ADDR]], align 4
 // CHECK-NEXT:    br label [[OMP_LOOP_PREHEADER:%.*]]
 // CHECK:       omp_loop.preheader:
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP4]], i64 128) ]
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[TMP5]], i64 64) ]
-// CHECK-NEXT:    call void @llvm.assume(i1 true) [ "align"(ptr [[ARRAYDECAY]], i64 16) ]
 // CHECK-NEXT:    br label [[OMP_LOOP_HEADER:%.*]]
 // CHECK:       omp_loop.header:
 // CHECK-NEXT:    [[OMP_LOOP_IV:%.*]] = phi i32 [ 0, [[OMP_LOOP_PREHEADER]] ], [ [[OMP_LOOP_NEXT:%.*]], [[OMP_LOOP_INC:%.*]] ]
diff --git a/flang/lib/Lower/OpenMP/ClauseProcessor.cpp b/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
index 48c559a78b9bc4..1b226482a8c13c 100644
--- a/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
+++ b/flang/lib/Lower/OpenMP/ClauseProcessor.cpp
@@ -606,6 +606,8 @@ addAlignedClause(lower::AbstractConverter &converter,
   // Do not generate alignment assumption if alignment is less than or equal to
   // 0.
   if (alignment > 0) {
+    // alignment value must be power of 2
+    assert((alignment & (alignment - 1)) == 0 && "alignment is not power of 2");
     auto &objects = std::get<omp::ObjectList>(clause.t);
     if (!objects.empty())
       genObjectList(objects, converter, alignedVars);
diff --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
index 21004e6a15d495..2fb53609e84397 100644
--- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
+++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
@@ -5302,10 +5302,11 @@ void OpenMPIRBuilder::applySimd(CanonicalLoopInfo *CanonicalLoop,
   Loop *L = LI.getLoopFor(CanonicalLoop->getHeader());
   if (AlignedVars.size()) {
     InsertPointTy IP = Builder.saveIP();
-    Builder.SetInsertPoint(CanonicalLoop->getPreheader()->getTerminator());
     for (auto &AlignedItem : AlignedVars) {
       Value *AlignedPtr = AlignedItem.first;
       Value *Alignment = AlignedItem.second;
+      Instruction *loadInst = dyn_cast<Instruction>(AlignedPtr);
+      Builder.SetInsertPoint(loadInst->getNextNode());
       Builder.CreateAlignmentAssumption(F->getDataLayout(),
                                         AlignedPtr, Alignment);
     }
diff --git a/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp b/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
index 630cd03c688012..4ec1b281d4d2d3 100644
--- a/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
+++ b/llvm/unittests/Frontend/OpenMPIRBuilderTest.cpp
@@ -1993,6 +1993,7 @@ TEST_F(OpenMPIRBuilderTest, ApplySimdCustomAligned) {
   OpenMPIRBuilder OMPBuilder(*M);
   IRBuilder<> Builder(BB);
   const int AlignmentValue = 32;
+  llvm::BasicBlock *sourceBlock = Builder.GetInsertBlock();
   AllocaInst *Alloc1 =
       Builder.CreateAlloca(Builder.getPtrTy(), Builder.getInt64(1));
   LoadInst *Load1 = Builder.CreateLoad(Alloc1->getAllocatedType(), Alloc1);
@@ -2031,13 +2032,12 @@ TEST_F(OpenMPIRBuilderTest, ApplySimdCustomAligned) {
 
   // Check if number of assumption instructions is equal to number of aligned
   // variables
-  BasicBlock *LoopPreheader = CLI->getPreheader();
-  size_t NumAssummptionCallsInPreheader = count_if(
-      *LoopPreheader, [](Instruction &I) { return isa<AssumeInst>(I); });
+  size_t NumAssummptionCallsInPreheader =
+      count_if(*sourceBlock, [](Instruction &I) { return isa<AssumeInst>(I); });
   EXPECT_EQ(NumAssummptionCallsInPreheader, AlignedVars.size());
 
   // Check if variables are correctly aligned
-  for (Instruction &Instr : *LoopPreheader) {
+  for (Instruction &Instr : *sourceBlock) {
     if (!isa<AssumeInst>(Instr))
       continue;
     AssumeInst *AssumeInstruction = cast<AssumeInst>(&Instr);
diff --git a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
index 49fe509800491a..b4ccb77f45a42c 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp
@@ -150,10 +150,6 @@ static LogicalResult checkImplementationStatus(Operation &op) {
                           << " operation";
   };
 
-  auto checkAligned = [&todo](auto op, LogicalResult &result) {
-    if (!op.getAlignedVars().empty() || op.getAlignments())
-      result = todo("aligned");
-  };
   auto checkAllocate = [&todo](auto op, LogicalResult &result) {
     if (!op.getAllocateVars().empty() || !op.getAllocatorVars().empty())
       result = todo("allocate");
@@ -271,7 +267,6 @@ static LogicalResult checkImplementationStatus(Operation &op) {
       })
       .Case([&](omp::ParallelOp op) { checkAllocate(op, result); })
       .Case([&](omp::SimdOp op) {
-        checkAligned(op, result);
         checkLinear(op, result);
         checkNontemporal(op, result);
         checkPrivate(op, result);
@@ -2257,6 +2252,24 @@ convertOmpSimd(Operation &opInst, llvm::IRBuilderBase &builder,
 
   llvm::MapVector<llvm::Value *, llvm::Value *> alignedVars;
   llvm::omp::OrderKind order = convertOrderKind(simdOp.getOrder());
+  llvm::BasicBlock *sourceBlock = builder.GetInsertBlock();
+  std::optional<ArrayAttr> alignmentValues = simdOp.getAlignments();
+  mlir::OperandRange operands = simdOp.getAlignedVars();
+  for (size_t i = 0; i < operands.size(); ++i) {
+    llvm::Value *alignment = nullptr;
+    llvm::Value *llvmVal = moduleTranslation.lookupValue(operands[i]);
+    llvm::Type *ty = llvmVal->getType();
+    if (auto intAttr = llvm::dyn_cast<IntegerAttr>((*alignmentValues)[i])) {
+      alignment = builder.getInt64(intAttr.getInt());
+      assert(ty->isPointerTy() && "Invalid type for aligned variable");
+      assert(alignment && "Invalid alignment value");
+      auto curInsert = builder.saveIP();
+      builder.SetInsertPoint(sourceBlock->getTerminator());
+      llvmVal = builder.CreateLoad(ty, llvmVal);
+      builder.restoreIP(curInsert);
+      alignedVars[llvmVal] = alignment;
+    }
+  }
   ompBuilder->applySimd(loopInfo, alignedVars,
                         simdOp.getIfExpr()
                             ? moduleTranslation.lookupValue(simdOp.getIfExpr())
diff --git a/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir b/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir
new file mode 100644
index 00000000000000..234604e4b664ad
--- /dev/null
+++ b/mlir/test/Target/LLVMIR/openmp-simd-aligned.mlir
@@ -0,0 +1,60 @@
+// RUN: mlir-translate -mlir-to-llvmir -split-input-file %s | FileCheck %s
+
+//CHECK-LABEL: define void @_QPsimd_aligned_pointer() {
+//CHECK:   %[[A_PTR:.*]] = alloca { ptr, i64, i32, i8, i8, i8, i8 }, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_PTR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_pointer() {
+  %1 = llvm.mlir.constant(1 : i64) : i64
+  %2 = llvm.alloca %1 x !llvm.struct<(ptr, i64, i32, i8, i8, i8, i8)> {bindc_name = "x"} : (i64) -> !llvm.ptr
+  %3 = llvm.alloca %1 x i32 {bindc_name = "i", pinned} : (i64) -> !llvm.ptr
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  %5 = llvm.mlir.constant(10 : i32) : i32
+  %6 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%2 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%4) to (%5) inclusive step (%6) {
+      llvm.store %arg0, %3 : i32, !llvm.ptr
+      omp.yield
+    }
+  }
+  llvm.return
+}
+
+//CHECK-LABEL: define void @_QPsimd_aligned_cptr() {
+//CHECK:   %[[A_CPTR:.*]] = alloca %_QM__fortran_builtinsT__builtin_c_ptr, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_CPTR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_cptr() {
+  %0 = llvm.mlir.constant(1 : i64) : i64
+  %1 = llvm.alloca %0 x !llvm.struct<"_QM__fortran_builtinsT__builtin_c_ptr", (i64)> {bindc_name = "a"} : (i64) -> !llvm.ptr
+  %2 = llvm.mlir.constant(1 : i64) : i64
+  %3 = llvm.alloca %2 x i32 {bindc_name = "i", pinned} : (i64) -> !llvm.ptr
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  %5 = llvm.mlir.constant(10 : i32) : i32
+  %6 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%1 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%4) to (%5) inclusive step (%6) {
+      llvm.store %arg0, %3 : i32, !llvm.ptr
+      omp.yield
+    }
+  }
+  llvm.return
+}
+
+//CHECK-LABEL: define void @_QPsimd_aligned_allocatable() {
+//CHECK:   %[[A_ADDR:.*]] = alloca { ptr, i64, i32, i8, i8, i8, i8, [1 x [3 x i64]] }, i64 1, align 8
+//CHECK:   %[[A_VAL:.*]] = load ptr, ptr %[[A_ADDR]], align 8
+//CHECK:   call void @llvm.assume(i1 true) [ "align"(ptr %[[A_VAL]], i64 256) ]
+llvm.func @_QPsimd_aligned_allocatable() {
+  %0 = llvm.mlir.constant(1 : i64) : i64
+  %1 = llvm.alloca %0 x !llvm.struct<(ptr, i64, i32, i8, i8, i8, i8, array<1 x array<3 x i64>>)> {bindc_name = "a"} : (i64) -> !llvm.ptr
+  %2 = llvm.mlir.constant(1 : i32) : i32
+  %3 = llvm.mlir.constant(10 : i32) : i32
+  %4 = llvm.mlir.constant(1 : i32) : i32
+  omp.simd aligned(%1 : !llvm.ptr -> 256 : i64) {
+    omp.loop_nest (%arg0) : i32 = (%2) to (%3) inclusive step (%4) {
+      omp.yield
+    }
+  }
+  llvm.return
+}
diff --git a/mlir/test/Target/LLVMIR/openmp-todo.mlir b/mlir/test/Target/LLVMIR/openmp-todo.mlir
index c1e0014d1f571f..010f6048706280 100644
--- a/mlir/test/Target/LLVMIR/openmp-todo.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-todo.mlir
@@ -127,18 +127,6 @@ llvm.func @sections_private(%x : !llvm.ptr) {
   llvm.return
 }
 
-// -----
-
-llvm.func @simd_aligned(%lb : i32, %ub : i32, %step : i32, %x : !llvm.ptr) {
-  // expected-error@below {{not yet implemented: Unhandled clause aligned in omp.simd operation}}
-  // expected-error@below {{LLVM Translation failed for operation: omp.simd}}
-  omp.simd aligned(%x : !llvm.ptr -> 32) {
-    omp.loop_nest (%iv) : i32 = (%lb) to (%ub) step (%step) {
-      omp.yield
-    }
-  }
-  llvm.return
-}
 
 // -----

[MLIR][OpenMP] Lowering aligned clause to LLVM IR for SIMD directive

ed29417

kaviya2510 requested review from skatrak, tblah, kiranchandramohan, DominikAdamski, NimishMishra, kiranktp and Thirumalai-Shaktivel December 11, 2024 10:25

llvmbot added clang Clang issues not falling into any other category mlir:llvm mlir flang Flang issues not falling into any other category mlir:openmp flang:fir-hlfir flang:openmp clang:openmp OpenMP related changes to Clang labels Dec 11, 2024

kaviya2510 mentioned this pull request Dec 11, 2024

[OpenMP][MLIR] Lowering aligned clause to LLVM IR #111676

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[MLIR][OpenMP] Lowering aligned clause to LLVM IR for SIMD directive #119536

[MLIR][OpenMP] Lowering aligned clause to LLVM IR for SIMD directive #119536

kaviya2510 commented Dec 11, 2024 •

edited

Loading

github-actions bot commented Dec 11, 2024

llvmbot commented Dec 11, 2024 •

edited

Loading

llvmbot commented Dec 11, 2024

llvmbot commented Dec 11, 2024

[MLIR][OpenMP] Lowering aligned clause to LLVM IR for SIMD directive #119536

Are you sure you want to change the base?

[MLIR][OpenMP] Lowering aligned clause to LLVM IR for SIMD directive #119536

Conversation

kaviya2510 commented Dec 11, 2024 • edited Loading

github-actions bot commented Dec 11, 2024

llvmbot commented Dec 11, 2024 • edited Loading

llvmbot commented Dec 11, 2024

llvmbot commented Dec 11, 2024

kaviya2510 commented Dec 11, 2024 •

edited

Loading

llvmbot commented Dec 11, 2024 •

edited

Loading