; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --filter-out-after "scalar.ph\:" --version 5
; RUN: opt -S -passes=loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -force-target-supports-scalable-vectors=true -scalable-vectorization=on < %s | FileCheck %s --check-prefix=CHECKUF1
; RUN: opt -S -passes=loop-vectorize -force-vector-interleave=2 -force-vector-width=4 -force-target-supports-scalable-vectors=true -scalable-vectorization=on < %s | FileCheck %s --check-prefix=CHECKUF2

; For an interleave factor of 2, vscale is scaled by 8 instead of 4 (and thus shifted left by 3 instead of 2).
; There is also the increment for the next iteration, e.g. instead of indexing IDXB, it indexes at IDXB + vscale * 4.
define void @loop(i64 %N, ptr noalias %a, ptr noalias %b) {
; CHECKUF1-LABEL: define void @loop(
; CHECKUF1-SAME: i64 [[N:%.*]], ptr noalias [[A:%.*]], ptr noalias [[B:%.*]]) {
; CHECKUF1-NEXT:  [[ENTRY:.*:]]
; CHECKUF1-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; CHECKUF1-NEXT:    [[TMP1:%.*]] = shl nuw i64 [[TMP0]], 2
; CHECKUF1-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; CHECKUF1-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECKUF1:       [[VECTOR_PH]]:
; CHECKUF1-NEXT:    [[TMP6:%.*]] = shl nuw i64 [[TMP0]], 2
; CHECKUF1-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP6]]
; CHECKUF1-NEXT:    [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECKUF1-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECKUF1:       [[VECTOR_BODY]]:
; CHECKUF1-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECKUF1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds double, ptr [[B]], i64 [[INDEX]]
; CHECKUF1-NEXT:    [[WIDE_LOAD:%.*]] = load <vscale x 4 x double>, ptr [[TMP7]], align 8
; CHECKUF1-NEXT:    [[TMP8:%.*]] = fadd <vscale x 4 x double> [[WIDE_LOAD]], splat (double 1.000000e+00)
; CHECKUF1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds double, ptr [[A]], i64 [[INDEX]]
; CHECKUF1-NEXT:    store <vscale x 4 x double> [[TMP8]], ptr [[TMP9]], align 8
; CHECKUF1-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP6]]
; CHECKUF1-NEXT:    [[TMP10:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECKUF1-NEXT:    br i1 [[TMP10]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECKUF1:       [[MIDDLE_BLOCK]]:
; CHECKUF1-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECKUF1-NEXT:    br i1 [[CMP_N]], [[EXIT:label %.*]], label %[[SCALAR_PH]]
; CHECKUF1:       [[SCALAR_PH]]:
;
; CHECKUF2-LABEL: define void @loop(
; CHECKUF2-SAME: i64 [[N:%.*]], ptr noalias [[A:%.*]], ptr noalias [[B:%.*]]) {
; CHECKUF2-NEXT:  [[ENTRY:.*:]]
; CHECKUF2-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; CHECKUF2-NEXT:    [[TMP1:%.*]] = shl nuw i64 [[TMP0]], 3
; CHECKUF2-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; CHECKUF2-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECKUF2:       [[VECTOR_PH]]:
; CHECKUF2-NEXT:    [[TMP3:%.*]] = shl nuw i64 [[TMP0]], 2
; CHECKUF2-NEXT:    [[TMP6:%.*]] = shl nuw i64 [[TMP3]], 1
; CHECKUF2-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP6]]
; CHECKUF2-NEXT:    [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECKUF2-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECKUF2:       [[VECTOR_BODY]]:
; CHECKUF2-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECKUF2-NEXT:    [[TMP7:%.*]] = getelementptr inbounds double, ptr [[B]], i64 [[INDEX]]
; CHECKUF2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds double, ptr [[TMP7]], i64 [[TMP3]]
; CHECKUF2-NEXT:    [[WIDE_LOAD:%.*]] = load <vscale x 4 x double>, ptr [[TMP7]], align 8
; CHECKUF2-NEXT:    [[WIDE_LOAD3:%.*]] = load <vscale x 4 x double>, ptr [[TMP9]], align 8
; CHECKUF2-NEXT:    [[TMP10:%.*]] = fadd <vscale x 4 x double> [[WIDE_LOAD]], splat (double 1.000000e+00)
; CHECKUF2-NEXT:    [[TMP11:%.*]] = fadd <vscale x 4 x double> [[WIDE_LOAD3]], splat (double 1.000000e+00)
; CHECKUF2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds double, ptr [[A]], i64 [[INDEX]]
; CHECKUF2-NEXT:    [[TMP14:%.*]] = getelementptr inbounds double, ptr [[TMP12]], i64 [[TMP3]]
; CHECKUF2-NEXT:    store <vscale x 4 x double> [[TMP10]], ptr [[TMP12]], align 8
; CHECKUF2-NEXT:    store <vscale x 4 x double> [[TMP11]], ptr [[TMP14]], align 8
; CHECKUF2-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP6]]
; CHECKUF2-NEXT:    [[TMP15:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECKUF2-NEXT:    br i1 [[TMP15]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECKUF2:       [[MIDDLE_BLOCK]]:
; CHECKUF2-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECKUF2-NEXT:    br i1 [[CMP_N]], [[EXIT:label %.*]], label %[[SCALAR_PH]]
; CHECKUF2:       [[SCALAR_PH]]:
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %arrayidx = getelementptr inbounds double, ptr %b, i64 %iv
  %0 = load double, ptr %arrayidx, align 8
  %add = fadd double %0, 1.000000e+00
  %arrayidx2 = getelementptr inbounds double, ptr %a, i64 %iv
  store double %add, ptr %arrayidx2, align 8
  %iv.next = add nuw nsw i64 %iv, 1
  %ec = icmp eq i64 %iv.next, %N
  br i1 %ec, label %exit, label %loop, !llvm.loop !1

exit:
  ret void
}

!1 = distinct !{!1, !2}
!2 = !{!"llvm.loop.vectorize.scalable.enable", i1 true}