Skip to content

[VPlan] Preserve nusw in createInBoundsPtrAdd #151549

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
artagnon merged 1 commit into from
Aug 18, 2025
Merged

[VPlan] Preserve nusw in createInBoundsPtrAdd #151549

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
14 changes: 8 additions & 6 deletions llvm/lib/Transforms/Vectorize/LoopVectorizationPlanner.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -256,13 +256,15 @@ class VPBuilder {
new VPInstruction(VPInstruction::PtrAdd, {Ptr, Offset},
GEPNoWrapFlags::none(), DL, Name));
}
VPInstruction *createInBoundsPtrAdd(VPValue *Ptr, VPValue *Offset,
DebugLoc DL = DebugLoc::getUnknown(),
const Twine &Name = "") {
return tryInsertInstruction(
new VPInstruction(VPInstruction::PtrAdd, {Ptr, Offset},
GEPNoWrapFlags::inBounds(), DL, Name));

VPInstruction *createNoWrapPtrAdd(VPValue *Ptr, VPValue *Offset,
GEPNoWrapFlags GEPFlags,
DebugLoc DL = DebugLoc::getUnknown(),
const Twine &Name = "") {
return tryInsertInstruction(new VPInstruction(
VPInstruction::PtrAdd, {Ptr, Offset}, GEPFlags, DL, Name));
}

VPInstruction *createWidePtrAdd(VPValue *Ptr, VPValue *Offset,
DebugLoc DL = DebugLoc::getUnknown(),
const Twine &Name = "") {
Expand Down
4 changes: 1 addition & 3 deletions llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2608,9 +2608,7 @@ void VPlanTransforms::createInterleaveGroups(
VPValue *OffsetVPV =
Plan.getOrAddLiveIn(ConstantInt::get(Plan.getContext(), -Offset));
VPBuilder B(InsertPos);
Addr = NW.isInBounds()
? B.createInBoundsPtrAdd(InsertPos->getAddr(), OffsetVPV)
: B.createPtrAdd(InsertPos->getAddr(), OffsetVPV);
Addr = B.createNoWrapPtrAdd(InsertPos->getAddr(), OffsetVPV, NW);
}
// If the group is reverse, adjust the index to refer to the last vector
// lane instead of the first. We adjust the index from the first vector
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/Transforms/LoopVectorize/interleaved-accesses-different-insert-position.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@ define void @test_ig_insert_pos_at_end_of_vpbb(ptr noalias %dst, ptr noalias %sr
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[TMP3:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr nusw { i16, i16, i16, i16 }, ptr [[SRC]], i64 [[TMP3]], i32 2
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr i8, ptr [[TMP4]], i32 -4
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr nusw i8, ptr [[TMP4]], i32 -4
; CHECK-NEXT: [[WIDE_VEC:%.*]] = load <16 x i16>, ptr [[TMP5]], align 2
; CHECK-NEXT: [[STRIDED_VEC:%.*]] = shufflevector <16 x i16> [[WIDE_VEC]], <16 x i16> poison, <4 x i32> <i32 0, i32 4, i32 8, i32 12>
; CHECK-NEXT: [[STRIDED_VEC1:%.*]] = shufflevector <16 x i16> [[WIDE_VEC]], <16 x i16> poison, <4 x i32> <i32 2, i32 6, i32 10, i32 14>
Expand Down
148 changes: 148 additions & 0 deletions llvm/test/Transforms/LoopVectorize/interleaved-accesses-gep-nowrap-flags.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -185,3 +185,151 @@ loop:
exit:
ret void
}

define void @nusw_preservation_2(ptr %src, ptr noalias %dst) {
; CHECK-LABEL: define void @nusw_preservation_2(
; CHECK-SAME: ptr [[SRC:%.*]], ptr noalias [[DST:%.*]]) {
; CHECK-NEXT: [[ENTRY:.*:]]
; CHECK-NEXT: br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK: [[VECTOR_PH]]:
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[OFFSET_IDX:%.*]] = mul i64 [[INDEX]], 2
; CHECK-NEXT: [[TMP0:%.*]] = or disjoint i64 [[OFFSET_IDX]], 1
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr nusw i8, ptr [[SRC]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr nusw i8, ptr [[TMP1]], i32 -1
; CHECK-NEXT: [[WIDE_VEC:%.*]] = load <8 x i8>, ptr [[TMP2]], align 1
; CHECK-NEXT: [[STRIDED_VEC:%.*]] = shufflevector <8 x i8> [[WIDE_VEC]], <8 x i8> poison, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
; CHECK-NEXT: [[STRIDED_VEC1:%.*]] = shufflevector <8 x i8> [[WIDE_VEC]], <8 x i8> poison, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
; CHECK-NEXT: [[TMP3:%.*]] = add <4 x i8> [[STRIDED_VEC1]], [[STRIDED_VEC]]
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr nusw i8, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT: store <4 x i8> [[TMP3]], ptr [[TMP4]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT: br i1 [[TMP5]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br [[EXIT:label %.*]]
; CHECK: [[SCALAR_PH]]:
;
entry:
br label %loop

loop: ; preds = %loop, %entry
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%iv2 = phi i64 [ 0, %entry ], [ %iv2.next, %loop ]
%or.1 = or disjoint i64 %iv2, 1
%gep.src.or.1 = getelementptr nusw i8, ptr %src, i64 %or.1
%load.src.1 = load i8, ptr %gep.src.or.1, align 1
%gep.src.iv2 = getelementptr nusw i8, ptr %src, i64 %iv2
%load.src.2 = load i8, ptr %gep.src.iv2, align 1
%add = add i8 %load.src.1, %load.src.2
%gep.dst.iv = getelementptr nusw i8, ptr %dst, i64 %iv
store i8 %add, ptr %gep.dst.iv, align 1
%iv2.next = add i64 %iv2, 2
%iv.next = add i64 %iv, 1
%exit.cond = icmp eq i64 %iv.next, 100
br i1 %exit.cond, label %exit, label %loop

exit:
ret void
}

define void @inbounds_preservation_2(ptr %src, ptr noalias %dst) {
; CHECK-LABEL: define void @inbounds_preservation_2(
; CHECK-SAME: ptr [[SRC:%.*]], ptr noalias [[DST:%.*]]) {
; CHECK-NEXT: [[ENTRY:.*:]]
; CHECK-NEXT: br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK: [[VECTOR_PH]]:
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[OFFSET_IDX:%.*]] = mul i64 [[INDEX]], 2
; CHECK-NEXT: [[TMP0:%.*]] = or disjoint i64 [[OFFSET_IDX]], 1
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i8, ptr [[SRC]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i8, ptr [[TMP1]], i32 -1
; CHECK-NEXT: [[WIDE_VEC:%.*]] = load <8 x i8>, ptr [[TMP2]], align 1
; CHECK-NEXT: [[STRIDED_VEC:%.*]] = shufflevector <8 x i8> [[WIDE_VEC]], <8 x i8> poison, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
; CHECK-NEXT: [[STRIDED_VEC1:%.*]] = shufflevector <8 x i8> [[WIDE_VEC]], <8 x i8> poison, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
; CHECK-NEXT: [[TMP3:%.*]] = add <4 x i8> [[STRIDED_VEC1]], [[STRIDED_VEC]]
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds i8, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT: store <4 x i8> [[TMP3]], ptr [[TMP4]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT: br i1 [[TMP5]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br [[EXIT:label %.*]]
; CHECK: [[SCALAR_PH]]:
;
entry:
br label %loop

loop: ; preds = %loop, %entry
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%iv2 = phi i64 [ 0, %entry ], [ %iv2.next, %loop ]
%or.1 = or disjoint i64 %iv2, 1
%gep.src.or.1 = getelementptr inbounds i8, ptr %src, i64 %or.1
%load.src.1 = load i8, ptr %gep.src.or.1, align 1
%gep.src.iv2 = getelementptr inbounds i8, ptr %src, i64 %iv2
%load.src.2 = load i8, ptr %gep.src.iv2, align 1
%add = add i8 %load.src.1, %load.src.2
%gep.dst.iv = getelementptr inbounds i8, ptr %dst, i64 %iv
store i8 %add, ptr %gep.dst.iv, align 1
%iv2.next = add i64 %iv2, 2
%iv.next = add i64 %iv, 1
%exit.cond = icmp eq i64 %iv.next, 100
br i1 %exit.cond, label %exit, label %loop

exit:
ret void
}

define void @nuw_drop_2(ptr %src, ptr noalias %dst) {
; CHECK-LABEL: define void @nuw_drop_2(
; CHECK-SAME: ptr [[SRC:%.*]], ptr noalias [[DST:%.*]]) {
; CHECK-NEXT: [[ENTRY:.*:]]
; CHECK-NEXT: br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK: [[VECTOR_PH]]:
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[OFFSET_IDX:%.*]] = mul i64 [[INDEX]], 2
; CHECK-NEXT: [[TMP0:%.*]] = or disjoint i64 [[OFFSET_IDX]], 1
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr nuw i8, ptr [[SRC]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr i8, ptr [[TMP1]], i32 -1
; CHECK-NEXT: [[WIDE_VEC:%.*]] = load <8 x i8>, ptr [[TMP2]], align 1
; CHECK-NEXT: [[STRIDED_VEC:%.*]] = shufflevector <8 x i8> [[WIDE_VEC]], <8 x i8> poison, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
; CHECK-NEXT: [[STRIDED_VEC1:%.*]] = shufflevector <8 x i8> [[WIDE_VEC]], <8 x i8> poison, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
; CHECK-NEXT: [[TMP3:%.*]] = add <4 x i8> [[STRIDED_VEC1]], [[STRIDED_VEC]]
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr nuw i8, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT: store <4 x i8> [[TMP3]], ptr [[TMP4]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT: br i1 [[TMP5]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br [[EXIT:label %.*]]
; CHECK: [[SCALAR_PH]]:
;
entry:
br label %loop

loop: ; preds = %loop, %entry
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%iv2 = phi i64 [ 0, %entry ], [ %iv2.next, %loop ]
%or.1 = or disjoint i64 %iv2, 1
%gep.src.or.1 = getelementptr nuw i8, ptr %src, i64 %or.1
%load.src.1 = load i8, ptr %gep.src.or.1, align 1
%gep.src.iv2 = getelementptr nuw i8, ptr %src, i64 %iv2
%load.src.2 = load i8, ptr %gep.src.iv2, align 1
%add = add i8 %load.src.1, %load.src.2
%gep.dst.iv = getelementptr nuw i8, ptr %dst, i64 %iv
store i8 %add, ptr %gep.dst.iv, align 1
%iv2.next = add i64 %iv2, 2
%iv.next = add i64 %iv, 1
%exit.cond = icmp eq i64 %iv.next, 100
br i1 %exit.cond, label %exit, label %loop

exit:
ret void
}