[LV] Add tests for epilogue vectorization with widened inductions.

Includes a test for the miscompile in #57712.

1 files changed, 274 insertions, 0 deletions

diff --git a/llvm/test/Transforms/LoopVectorize/optimal-epilog-vectorization-widen-inductions.ll b/llvm/test/Transforms/LoopVectorize/optimal-epilog-vectorization-widen-inductions.ll
new file mode 100644
index 000000000000..22009293b7f5
--- /dev/null
+++ b/llvm/test/Transforms/LoopVectorize/optimal-epilog-vectorization-widen-inductions.ll
@@ -0,0 +1,274 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -passes='loop-vectorize' -force-vector-width=2 -enable-epilogue-vectorization -epilogue-vectorization-force-VF=2 -S %s | FileCheck %s
+
+target triple = "arm64-apple-macosx"
+
+; Test for #57712.
+define void @test_widen_ptr_induction(ptr %ptr.start.1) {
+; CHECK-LABEL: @test_widen_ptr_induction(
+; CHECK-NEXT:  iter.check:
+; CHECK-NEXT:    br i1 false, label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
+; CHECK:       vector.main.loop.iter.check:
+; CHECK-NEXT:    br i1 false, label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
+; CHECK:       vector.ph:
+; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
+; CHECK:       vector.body:
+; CHECK-NEXT:    [[POINTER_PHI:%.*]] = phi ptr [ [[PTR_START_1:%.*]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <2 x i64> <i64 0, i64 1>
+; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <2 x i64> <i64 2, i64 3>
+; CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <2 x ptr> [[TMP0]], zeroinitializer
+; CHECK-NEXT:    [[TMP3:%.*]] = icmp ne <2 x ptr> [[TMP1]], zeroinitializer
+; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <2 x i1> [[TMP2]], i32 0
+; CHECK-NEXT:    tail call void @llvm.assume(i1 [[TMP4]])
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x i1> [[TMP2]], i32 1
+; CHECK-NEXT:    tail call void @llvm.assume(i1 [[TMP5]])
+; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x i1> [[TMP3]], i32 0
+; CHECK-NEXT:    tail call void @llvm.assume(i1 [[TMP6]])
+; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <2 x i1> [[TMP3]], i32 1
+; CHECK-NEXT:    tail call void @llvm.assume(i1 [[TMP7]])
+; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <2 x ptr> [[TMP0]], i32 0
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr i8, ptr [[TMP8]], i32 0
+; CHECK-NEXT:    store <2 x i8> zeroinitializer, ptr [[TMP9]], align 1
+; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr i8, ptr [[TMP8]], i32 2
+; CHECK-NEXT:    store <2 x i8> zeroinitializer, ptr [[TMP10]], align 1
+; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
+; CHECK-NEXT:    [[PTR_IND]] = getelementptr i8, ptr [[POINTER_PHI]], i64 4
+; CHECK-NEXT:    [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], 10000
+; CHECK-NEXT:    br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK:       middle.block:
+; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10001, 10000
+; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
+; CHECK:       vec.epilog.iter.check:
+; CHECK-NEXT:    [[IND_END1:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 10000
+; CHECK-NEXT:    br i1 true, label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
+; CHECK:       vec.epilog.ph:
+; CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ 10000, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
+; CHECK-NEXT:    [[IND_END:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 10000
+; CHECK-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
+; CHECK:       vec.epilog.vector.body:
+; CHECK-NEXT:    [[POINTER_PHI5:%.*]] = phi ptr [ [[PTR_START_1]], [[VEC_EPILOG_PH]] ], [ [[PTR_IND6:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; CHECK-NEXT:    [[INDEX4:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT7:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP12:%.*]] = getelementptr i8, ptr [[POINTER_PHI5]], <2 x i64> <i64 0, i64 1>
+; CHECK-NEXT:    [[TMP13:%.*]] = icmp ne <2 x ptr> [[TMP12]], zeroinitializer
+; CHECK-NEXT:    [[TMP14:%.*]] = extractelement <2 x i1> [[TMP13]], i32 0
+; CHECK-NEXT:    tail call void @llvm.assume(i1 [[TMP14]])
+; CHECK-NEXT:    [[TMP15:%.*]] = extractelement <2 x i1> [[TMP13]], i32 1
+; CHECK-NEXT:    tail call void @llvm.assume(i1 [[TMP15]])
+; CHECK-NEXT:    [[TMP16:%.*]] = extractelement <2 x ptr> [[TMP12]], i32 0
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr i8, ptr [[TMP16]], i32 0
+; CHECK-NEXT:    store <2 x i8> zeroinitializer, ptr [[TMP17]], align 1
+; CHECK-NEXT:    [[INDEX_NEXT7]] = add nuw i64 [[INDEX4]], 2
+; CHECK-NEXT:    [[PTR_IND6]] = getelementptr i8, ptr [[POINTER_PHI5]], i64 2
+; CHECK-NEXT:    [[TMP18:%.*]] = icmp eq i64 [[INDEX_NEXT7]], 10000
+; CHECK-NEXT:    br i1 [[TMP18]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP2:![0-9]+]]
+; CHECK:       vec.epilog.middle.block:
+; CHECK-NEXT:    [[CMP_N3:%.*]] = icmp eq i64 10001, 10000
+; CHECK-NEXT:    br i1 [[CMP_N3]], label [[EXIT]], label [[VEC_EPILOG_SCALAR_PH]]
+; CHECK:       vec.epilog.scalar.ph:
+; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 10000, [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ 10000, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK:%.*]] ]
+; CHECK-NEXT:    [[BC_RESUME_VAL2:%.*]] = phi ptr [ [[IND_END]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END1]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[PTR_START_1]], [[ITER_CHECK]] ]
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT:    [[PTR_IV:%.*]] = phi ptr [ [[BC_RESUME_VAL2]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[PTR_IV_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT:    [[CMP_I_I_I_I:%.*]] = icmp ne ptr [[PTR_IV]], null
+; CHECK-NEXT:    tail call void @llvm.assume(i1 [[CMP_I_I_I_I]])
+; CHECK-NEXT:    store i8 0, ptr [[PTR_IV]], align 1
+; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
+; CHECK-NEXT:    [[PTR_IV_NEXT]] = getelementptr inbounds i8, ptr [[PTR_IV]], i64 1
+; CHECK-NEXT:    [[EC:%.*]] = icmp eq i64 [[IV]], 10000
+; CHECK-NEXT:    br i1 [[EC]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP4:![0-9]+]]
+; CHECK:       exit:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %loop
+
+loop:
+  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
+  %ptr.iv = phi ptr [ %ptr.start.1, %entry ], [ %ptr.iv.next, %loop ]
+  %cmp.i.i.i.i = icmp ne ptr %ptr.iv, null
+  tail call void @llvm.assume(i1 %cmp.i.i.i.i)
+  store i8 0, ptr %ptr.iv, align 1
+  %iv.next = add nuw nsw i64 %iv, 1
+  %ptr.iv.next = getelementptr inbounds i8, ptr %ptr.iv, i64 1
+  %ec = icmp eq i64 %iv, 10000
+  br i1 %ec, label %exit, label %loop
+
+exit:
+  ret void
+}
+
+declare void @llvm.assume(i1 noundef)
+
+
+define void @test_widen_induction(i32 %init, ptr %A, i32 %N, i32 %step) {
+; CHECK-LABEL: @test_widen_induction(
+; 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:    [[VEC_IND:%.*]] = phi <2 x i64> [ <i64 0, i64 1>, [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[STEP_ADD:%.*]] = add <2 x i64> [[VEC_IND]], <i64 2, i64 2>
+; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 2
+; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i64, ptr [[A:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i64, ptr [[TMP2]], i32 0
+; CHECK-NEXT:    store <2 x i64> [[VEC_IND]], ptr [[TMP4]], align 4
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i64, ptr [[TMP2]], i32 2
+; CHECK-NEXT:    store <2 x i64> [[STEP_ADD]], ptr [[TMP5]], align 4
+; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
+; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i64> [[STEP_ADD]], <i64 2, i64 2>
+; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1000
+; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
+; CHECK:       middle.block:
+; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1000, 1000
+; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
+; CHECK:       scalar.ph:
+; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 1000, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    [[IV_1:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_1_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT:    [[GEP_A:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[IV_1]]
+; CHECK-NEXT:    store i64 [[IV_1]], ptr [[GEP_A]], align 4
+; CHECK-NEXT:    [[IV_1_NEXT]] = add nuw nsw i64 [[IV_1]], 1
+; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_1_NEXT]], 1000
+; CHECK-NEXT:    br i1 [[EXITCOND]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP6:![0-9]+]]
+; CHECK:       exit:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %loop
+
+loop:
+  %iv.1 = phi i64 [ 0, %entry ], [ %iv.1.next, %loop ]
+  %gep.A = getelementptr inbounds i64, ptr %A, i64 %iv.1
+  store i64 %iv.1, ptr %gep.A, align 4
+  %iv.1.next = add nuw nsw i64 %iv.1, 1
+  %exitcond = icmp eq i64 %iv.1.next, 1000
+  br i1 %exitcond, label %exit, label %loop
+
+exit:
+  ret void
+}
+
+define void @test_widen_extended_induction(ptr %dst) {
+; CHECK-LABEL: @test_widen_extended_induction(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_SCEVCHECK:%.*]]
+; CHECK:       vector.scevcheck:
+; CHECK-NEXT:    br i1 true, label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
+; CHECK:       vector.ph:
+; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
+; CHECK:       vector.body:
+; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_IND:%.*]] = phi <2 x i8> [ <i8 0, i8 1>, [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[STEP_ADD:%.*]] = add <2 x i8> [[VEC_IND]], <i8 2, i8 2>
+; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = trunc i32 [[INDEX]] to i8
+; CHECK-NEXT:    [[TMP0:%.*]] = add i8 [[OFFSET_IDX]], 0
+; CHECK-NEXT:    [[TMP1:%.*]] = add i8 [[OFFSET_IDX]], 2
+; CHECK-NEXT:    [[TMP2:%.*]] = zext i8 [[TMP0]] to i64
+; CHECK-NEXT:    [[TMP3:%.*]] = zext i8 [[TMP1]] to i64
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [6 x i8], ptr [[DST:%.*]], i64 0, i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [6 x i8], ptr [[DST]], i64 0, i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i8, ptr [[TMP4]], i32 0
+; CHECK-NEXT:    store <2 x i8> [[VEC_IND]], ptr [[TMP6]], align 1
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i8, ptr [[TMP4]], i32 2
+; CHECK-NEXT:    store <2 x i8> [[STEP_ADD]], ptr [[TMP7]], align 1
+; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
+; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i8> [[STEP_ADD]], <i8 2, i8 2>
+; CHECK-NEXT:    [[TMP8:%.*]] = icmp eq i32 [[INDEX_NEXT]], 10000
+; CHECK-NEXT:    br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
+; CHECK:       middle.block:
+; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i32 10000, 10000
+; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
+; CHECK:       scalar.ph:
+; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i8 [ 16, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ], [ 0, [[VECTOR_SCEVCHECK]] ]
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    [[IV:%.*]] = phi i8 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT:    [[IV_EXT:%.*]] = zext i8 [[IV]] to i64
+; CHECK-NEXT:    [[ARRAYIDX1449:%.*]] = getelementptr inbounds [6 x i8], ptr [[DST]], i64 0, i64 [[IV_EXT]]
+; CHECK-NEXT:    store i8 [[IV]], ptr [[ARRAYIDX1449]], align 1
+; CHECK-NEXT:    [[IV_NEXT]] = add i8 [[IV]], 1
+; CHECK-NEXT:    [[IV_NEXT_EXT:%.*]] = zext i8 [[IV_NEXT]] to i32
+; CHECK-NEXT:    [[EC:%.*]] = icmp eq i32 [[IV_NEXT_EXT]], 10000
+; CHECK-NEXT:    br i1 [[EC]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP8:![0-9]+]]
+; CHECK:       exit:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %loop
+
+loop:
+  %iv = phi i8 [ 0, %entry ], [ %iv.next, %loop ]
+  %iv.ext = zext i8 %iv to i64
+  %arrayidx1449 = getelementptr inbounds [6 x i8], ptr %dst, i64 0, i64 %iv.ext
+  store i8 %iv, ptr %arrayidx1449, align 1
+  %iv.next = add i8 %iv, 1
+  %iv.next.ext = zext i8 %iv.next to i32
+  %ec = icmp eq i32  %iv.next.ext, 10000
+  br i1 %ec, label %exit, label %loop
+
+exit:
+  ret void
+}
+
+define void @test_widen_truncated_induction(ptr %A) {
+; CHECK-LABEL: @test_widen_truncated_induction(
+; 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:    [[VEC_IND:%.*]] = phi <2 x i8> [ <i8 0, i8 1>, [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[STEP_ADD:%.*]] = add <2 x i8> [[VEC_IND]], <i8 2, i8 2>
+; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 2
+; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i8, ptr [[A:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i8, ptr [[TMP2]], i32 0
+; CHECK-NEXT:    store <2 x i8> [[VEC_IND]], ptr [[TMP4]], align 1
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i8, ptr [[TMP2]], i32 2
+; CHECK-NEXT:    store <2 x i8> [[STEP_ADD]], ptr [[TMP5]], align 1
+; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
+; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i8> [[STEP_ADD]], <i8 2, i8 2>
+; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], 10000
+; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP9:![0-9]+]]
+; CHECK:       middle.block:
+; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10000, 10000
+; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
+; CHECK:       scalar.ph:
+; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 10000, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT:    [[IV_TRUNC:%.*]] = trunc i64 [[IV]] to i8
+; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[IV]]
+; CHECK-NEXT:    store i8 [[IV_TRUNC]], ptr [[ARRAYIDX]], align 1
+; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
+; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp ne i64 [[IV_NEXT]], 10000
+; CHECK-NEXT:    br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT]], !llvm.loop [[LOOP10:![0-9]+]]
+; CHECK:       exit:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %loop
+
+loop:
+  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
+  %iv.trunc = trunc i64 %iv to i8
+  %arrayidx = getelementptr inbounds i8, ptr %A, i64 %iv
+  store i8 %iv.trunc, ptr %arrayidx, align 1
+  %iv.next = add nuw nsw i64 %iv, 1
+  %exitcond = icmp ne i64 %iv.next, 10000
+  br i1 %exitcond, label %loop, label %exit
+
+exit:
+  ret void
+}