test/Transforms/LoopInterchange/pr43176-move-to-new-latch.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -passes=loop-interchange -cache-line-size=64 -pass-remarks-missed='loop-interchange' -pass-remarks-output=%t -S
; RUN: FileCheck --input-file=%t %s

@b = external dso_local global [5 x i32], align 16

;; Not profitable to interchange, because the access is invariant to j loop.
;;
;;  for(int i=0;i<4;i++) {
;;    for(int j=1;j<4;j++) {
;;      b[i] = ....
;;    }
;; }

; CHECK: --- !Missed
; CHECK-NEXT: Pass:            loop-interchange
; CHECK-NEXT: Name:            InterchangeNotProfitable
; CHECK-NEXT: Function:        test1
; CHECK-NEXT: Args:
; CHECK-NEXT:  - String:          Interchanging loops is not considered to improve cache locality nor vectorization.

define void @test1() {
entry:
  br label %for.body

for.body:                                         ; preds = %for.inc3, %entry
  %inc41 = phi i32 [ %inc4, %for.inc3 ], [ undef, %entry ]
  br label %for.body2

for.body2:                                        ; preds = %for.inc, %for.body
  %lsr.iv = phi i32 [ %lsr.iv.next, %for.inc ], [ 1, %for.body ]
  br label %for.inc

for.inc:                                          ; preds = %for.body2
  %idxprom = sext i32 %inc41 to i64
  %arrayidx = getelementptr inbounds [5 x i32], ptr @b, i64 0, i64 %idxprom
  %0 = load i32, ptr %arrayidx, align 4
  store i32 undef, ptr %arrayidx, align 4
  %cmp = icmp slt i32 %lsr.iv, 4
  %lsr.iv.next = add nuw nsw i32 %lsr.iv, 1
  br i1 %cmp, label %for.body2, label %for.cond1.for.end_crit_edge

for.cond1.for.end_crit_edge:                      ; preds = %for.inc
  br label %for.inc3

for.inc3:                                         ; preds = %for.cond1.for.end_crit_edge
  %inc4 = add nsw i32 %inc41, 1
  br i1 false, label %for.body, label %for.cond.for.end5_crit_edge

for.cond.for.end5_crit_edge:                      ; preds = %for.inc3
  ret void
}


; CHECK: --- !Missed
; CHECK-NEXT: Pass:            loop-interchange
; CHECK-NEXT: Name:            InterchangeNotProfitable
; CHECK-NEXT: Function:        test2
; CHECK-NEXT: Args:
; CHECK-NEXT:  - String:          Interchanging loops is not considered to improve cache locality nor vectorization.

define void @test2() {
entry:
  br label %for.body

for.body:                                         ; preds = %for.inc3, %entry
  %inc41 = phi i32 [ %inc4, %for.inc3 ], [ undef, %entry ]
  br label %for.body2

for.body2:                                        ; preds = %for.inc, %for.body
  %lsr.iv = phi i32 [ %lsr.iv.next, %for.inc ], [ 1, %for.body ]
  br label %for.inc

for.inc:                                          ; preds = %for.body2
  %idxprom = sext i32 %inc41 to i64
  %arrayidx = getelementptr inbounds [5 x i32], ptr @b, i64 0, i64 %idxprom
  %0 = load i32, ptr %arrayidx, align 4
  %cmp = icmp slt i32 %lsr.iv, 4
  %cmp.zext = zext i1 %cmp to i32
  store i32 %cmp.zext, ptr %arrayidx, align 4
  %lsr.iv.next = add nuw nsw i32 %lsr.iv, 1
  br i1 %cmp, label %for.body2, label %for.cond1.for.end_crit_edge

for.cond1.for.end_crit_edge:                      ; preds = %for.inc
  br label %for.inc3

for.inc3:                                         ; preds = %for.cond1.for.end_crit_edge
  %inc4 = add nsw i32 %inc41, 1
  br i1 false, label %for.body, label %for.cond.for.end5_crit_edge

for.cond.for.end5_crit_edge:                      ; preds = %for.inc3
  ret void
}