Skip to content

[OpenMPIRBuilder] Fix tripcount not a multiple of tile size #154999

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Aug 26, 2025
Merged

[OpenMPIRBuilder] Fix tripcount not a multiple of tile size #154999

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6dc21da
[OpenMPIRBuilder] Fix tripcount not a multiple of tile size
Meinersbur Aug 22, 2025
7e6de45
Reduce change noise
Meinersbur Aug 22, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion clang/test/OpenMP/irbuilder_unroll_partial_factor_for.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,7 @@ void unroll_partial_heuristic_for(int n, float *a, float *b, float *c, float *d)
// CHECK-NEXT: br i1 [[OMP_FLOOR0_CMP]], label [[OMP_FLOOR0_BODY:%.*]], label [[OMP_FLOOR0_EXIT:%.*]]
// CHECK: omp_floor0.body:
// CHECK-NEXT: [[TMP13:%.*]] = add i32 [[OMP_FLOOR0_IV]], [[TMP9]]
// CHECK-NEXT: [[TMP14:%.*]] = icmp eq i32 [[TMP13]], [[OMP_FLOOR0_TRIPCOUNT]]
// CHECK-NEXT: [[TMP14:%.*]] = icmp eq i32 [[TMP13]], [[TMP4]]
// CHECK-NEXT: [[TMP15:%.*]] = select i1 [[TMP14]], i32 [[TMP5]], i32 13
// CHECK-NEXT: br label [[OMP_TILE0_PREHEADER:%.*]]
// CHECK: omp_tile0.preheader:
Expand Down
2 changes: 1 addition & 1 deletion clang/test/OpenMP/irbuilder_unroll_partial_heuristic_constant_for.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -85,7 +85,7 @@ void unroll_partial_heuristic_constant_for(float *a, float *b, float *c, float *
// CHECK-NEXT: br i1 [[OMP_FLOOR0_CMP]], label [[OMP_FLOOR0_BODY:%.*]], label [[OMP_FLOOR0_EXIT:%.*]]
// CHECK: omp_floor0.body:
// CHECK-NEXT: [[TMP12:%.*]] = add i32 [[OMP_FLOOR0_IV]], [[TMP8]]
// CHECK-NEXT: [[TMP13:%.*]] = icmp eq i32 [[TMP12]], [[OMP_FLOOR0_TRIPCOUNT]]
// CHECK-NEXT: [[TMP13:%.*]] = icmp eq i32 [[TMP12]], [[TMP3]]
// CHECK-NEXT: [[TMP14:%.*]] = select i1 [[TMP13]], i32 [[TMP4]], i32 4
// CHECK-NEXT: br label [[OMP_TILE0_PREHEADER:%.*]]
// CHECK: omp_tile0.preheader:
Expand Down
2 changes: 1 addition & 1 deletion clang/test/OpenMP/irbuilder_unroll_partial_heuristic_runtime_for.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,7 +87,7 @@ void unroll_partial_heuristic_runtime_for(int n, float *a, float *b, float *c, f
// CHECK-NEXT: br i1 [[OMP_FLOOR0_CMP]], label [[OMP_FLOOR0_BODY:%.*]], label [[OMP_FLOOR0_EXIT:%.*]]
// CHECK: omp_floor0.body:
// CHECK-NEXT: [[TMP13:%.*]] = add i32 [[OMP_FLOOR0_IV]], [[TMP9]]
// CHECK-NEXT: [[TMP14:%.*]] = icmp eq i32 [[TMP13]], [[OMP_FLOOR0_TRIPCOUNT]]
// CHECK-NEXT: [[TMP14:%.*]] = icmp eq i32 [[TMP13]], [[TMP4]]
// CHECK-NEXT: [[TMP15:%.*]] = select i1 [[TMP14]], i32 [[TMP5]], i32 4
// CHECK-NEXT: br label [[OMP_TILE0_PREHEADER:%.*]]
// CHECK: omp_tile0.preheader:
Expand Down
2 changes: 1 addition & 1 deletion clang/test/OpenMP/irbuilder_unroll_unroll_partial_factor.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,7 @@ void unroll_partial_factor_for(float *a, float *b, float *c, float *d) {
// CHECK-NEXT: br i1 [[OMP_FLOOR0_CMP]], label [[OMP_FLOOR0_BODY:%.*]], label [[OMP_FLOOR0_EXIT:%.*]]
// CHECK: omp_floor0.body:
// CHECK-NEXT: [[TMP12:%.*]] = add i32 [[OMP_FLOOR0_IV]], [[TMP8]]
// CHECK-NEXT: [[TMP13:%.*]] = icmp eq i32 [[TMP12]], [[OMP_FLOOR0_TRIPCOUNT]]
// CHECK-NEXT: [[TMP13:%.*]] = icmp eq i32 [[TMP12]], [[TMP3]]
// CHECK-NEXT: [[TMP14:%.*]] = select i1 [[TMP13]], i32 [[TMP4]], i32 2
// CHECK-NEXT: br label [[OMP_TILE0_PREHEADER:%.*]]
// CHECK: omp_tile0.preheader:
Expand Down
2 changes: 1 addition & 1 deletion clang/test/OpenMP/irbuilder_unroll_unroll_partial_heuristic.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@ void unroll_unroll_partial_heuristic(float *a, float *b, float *c, float *d) {
// CHECK-NEXT: [[OMP_FLOOR0_CMP:%.*]] = icmp ult i32 [[OMP_FLOOR0_IV]], [[OMP_FLOOR0_TRIPCOUNT]]
// CHECK-NEXT: br i1 [[OMP_FLOOR0_CMP]], label [[OMP_FLOOR0_BODY:%.*]], label [[OMP_FLOOR0_EXIT:%.*]]
// CHECK: omp_floor0.body:
// CHECK-NEXT: [[TMP7:%.*]] = icmp eq i32 [[OMP_FLOOR0_IV]], [[OMP_FLOOR0_TRIPCOUNT]]
// CHECK-NEXT: [[TMP7:%.*]] = icmp eq i32 [[OMP_FLOOR0_IV]], [[TMP3]]
// CHECK-NEXT: [[TMP8:%.*]] = select i1 [[TMP7]], i32 [[TMP4]], i32 8
// CHECK-NEXT: br label [[OMP_TILE0_PREHEADER:%.*]]
// CHECK: omp_tile0.preheader:
Expand Down
11 changes: 6 additions & 5 deletions llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5582,13 +5582,13 @@ OpenMPIRBuilder::tileLoops(DebugLoc DL, ArrayRef<CanonicalLoopInfo *> Loops,
// Compute the trip counts of the floor loops.
Builder.SetCurrentDebugLocation(DL);
Builder.restoreIP(OutermostLoop->getPreheaderIP());
SmallVector<Value *, 4> FloorCount, FloorRems;
SmallVector<Value *, 4> FloorCompleteCount, FloorCount, FloorRems;
for (int i = 0; i < NumLoops; ++i) {
Value *TileSize = TileSizes[i];
Value *OrigTripCount = OrigTripCounts[i];
Type *IVType = OrigTripCount->getType();

Value *FloorTripCount = Builder.CreateUDiv(OrigTripCount, TileSize);
Value *FloorCompleteTripCount = Builder.CreateUDiv(OrigTripCount, TileSize);
Value *FloorTripRem = Builder.CreateURem(OrigTripCount, TileSize);

// 0 if tripcount divides the tilesize, 1 otherwise.
Expand All @@ -5602,11 +5602,12 @@ OpenMPIRBuilder::tileLoops(DebugLoc DL, ArrayRef<CanonicalLoopInfo *> Loops,
Builder.CreateICmpNE(FloorTripRem, ConstantInt::get(IVType, 0));

FloorTripOverflow = Builder.CreateZExt(FloorTripOverflow, IVType);
FloorTripCount =
Builder.CreateAdd(FloorTripCount, FloorTripOverflow,
Value *FloorTripCount =
Builder.CreateAdd(FloorCompleteTripCount, FloorTripOverflow,
"omp_floor" + Twine(i) + ".tripcount", true);

// Remember some values for later use.
FloorCompleteCount.push_back(FloorCompleteTripCount);
FloorCount.push_back(FloorTripCount);
FloorRems.push_back(FloorTripRem);
}
Expand Down Expand Up @@ -5661,7 +5662,7 @@ OpenMPIRBuilder::tileLoops(DebugLoc DL, ArrayRef<CanonicalLoopInfo *> Loops,
Value *TileSize = TileSizes[i];

Value *FloorIsEpilogue =
Builder.CreateICmpEQ(FloorLoop->getIndVar(), FloorCount[i]);
Builder.CreateICmpEQ(FloorLoop->getIndVar(), FloorCompleteCount[i]);
Value *TileTripCount =
Builder.CreateSelect(FloorIsEpilogue, FloorRems[i], TileSize);

Expand Down