llvm/test/Transforms/LoopVectorize/AArch64/wide-vscale-range-max-runtime-vf.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6
; RUN: opt -passes=loop-vectorize -mtriple=aarch64-unknown-linux-gnu -mattr=+sve -S %s | FileCheck %s

; The maximum runtime VF is 4 * 2^30 = 2^32. The trip count (8) is not
; guaranteed to be a multiple of the VF for all values of vscale.
define void @wide_vscale_range_needs_tail_folding(ptr noalias %A) #0 {
; CHECK-LABEL: define void @wide_vscale_range_needs_tail_folding(
; CHECK-SAME: ptr noalias [[A:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT:    [[TMP1:%.*]] = shl nuw i64 [[TMP0]], 2
; CHECK-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 4 x i1> @llvm.get.active.lane.mask.nxv4i1.i64(i64 0, i64 8)
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 4 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[VECTOR_PH]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = call <vscale x 4 x i32> @llvm.masked.load.nxv4i32.p0(ptr align 4 [[TMP2]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i32> poison)
; CHECK-NEXT:    [[TMP3:%.*]] = add <vscale x 4 x i32> [[WIDE_LOAD]], splat (i32 1)
; CHECK-NEXT:    call void @llvm.masked.store.nxv4i32.p0(<vscale x 4 x i32> [[TMP3]], ptr align 4 [[TMP2]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP1]]
; CHECK-NEXT:    [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 4 x i1> @llvm.get.active.lane.mask.nxv4i1.i64(i64 [[INDEX_NEXT]], i64 8)
; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <vscale x 4 x i1> [[ACTIVE_LANE_MASK_NEXT]], i64 0
; CHECK-NEXT:    [[TMP5:%.*]] = xor i1 [[TMP4]], true
; CHECK-NEXT:    br i1 [[TMP5]], label %[[SCALAR_PH:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       [[SCALAR_PH]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr inbounds i32, ptr %A, i64 %iv
  %l = load i32, ptr %gep, align 4
  %add = add i32 %l, 1
  store i32 %add, ptr %gep, align 4
  %iv.next = add i64 %iv, 1
  %ec = icmp eq i64 %iv.next, 8
  br i1 %ec, label %exit, label %loop

exit:
  ret void
}

; Same maximum runtime VF of 2^32, but here it divides the trip count of 2^32,
; so no tail remains for any VF that may be chosen and the tail does not need
; folding.
define void @wide_vscale_range_no_tail_for_any_vf(ptr noalias %A) #1 {
; CHECK-LABEL: define void @wide_vscale_range_no_tail_for_any_vf(
; CHECK-SAME: ptr noalias [[A:%.*]]) #[[ATTR1:[0-9]+]] {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT:    [[TMP1:%.*]] = shl nuw i64 [[TMP0]], 2
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, ptr [[TMP2]], align 4
; CHECK-NEXT:    [[TMP3:%.*]] = add <vscale x 4 x i32> [[WIDE_LOAD]], splat (i32 1)
; CHECK-NEXT:    store <vscale x 4 x i32> [[TMP3]], ptr [[TMP2]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP1]]
; CHECK-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4294967296
; CHECK-NEXT:    br i1 [[TMP4]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr inbounds i32, ptr %A, i64 %iv
  %l = load i32, ptr %gep, align 4
  %add = add i32 %l, 1
  store i32 %add, ptr %gep, align 4
  %iv.next = add i64 %iv, 1
  %ec = icmp eq i64 %iv.next, 4294967296
  br i1 %ec, label %exit, label %loop

exit:
  ret void
}

attributes #0 = { vscale_range(1,1073741824) }
attributes #1 = { optsize vscale_range(1,1073741824) }