llvm/test/Transforms/LoopVectorize/RISCV/ordered-reduction.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -passes=loop-vectorize -hints-allow-reordering=false -S < %s | FileCheck %s --check-prefix=CHECK-ORDERED
; RUN: opt -passes=loop-vectorize -S < %s | FileCheck %s --check-prefix=CHECK-UNORDERED
; RUN: opt -passes=loop-vectorize -force-ordered-reductions=false -hints-allow-reordering=false -S < %s | FileCheck %s --check-prefix=CHECK-NOT-VECTORIZED
; RUN: opt -passes=loop-vectorize -force-ordered-reductions=true -hints-allow-reordering=false -S < %s | FileCheck %s --check-prefix=CHECK-ORDERED


target triple = "riscv64"

define float @fadd(ptr noalias nocapture readonly %a, i64 %n) #0 {
; CHECK-UNORDERED-LABEL: @fadd(
; CHECK-UNORDERED-NEXT:  entry:
; CHECK-UNORDERED-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N:%.*]], 4
; CHECK-UNORDERED-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK-UNORDERED:       vector.ph:
; CHECK-UNORDERED-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-UNORDERED-NEXT:    [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-UNORDERED-NEXT:    br label [[VECTOR_BODY:%.*]]
; CHECK-UNORDERED:       vector.body:
; CHECK-UNORDERED-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-UNORDERED-NEXT:    [[VEC_PHI:%.*]] = phi <4 x float> [ <float 0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, [[VECTOR_PH]] ], [ [[TMP3:%.*]], [[VECTOR_BODY]] ]
; CHECK-UNORDERED-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-UNORDERED-NEXT:    [[TMP1:%.*]] = getelementptr inbounds float, ptr [[A:%.*]], i64 [[TMP0]]
; CHECK-UNORDERED-NEXT:    [[TMP2:%.*]] = getelementptr inbounds float, ptr [[TMP1]], i32 0
; CHECK-UNORDERED-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP2]], align 4
; CHECK-UNORDERED-NEXT:    [[TMP3]] = fadd <4 x float> [[WIDE_LOAD]], [[VEC_PHI]]
; CHECK-UNORDERED-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-UNORDERED-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-UNORDERED-NEXT:    br i1 [[TMP4]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK-UNORDERED:       middle.block:
; CHECK-UNORDERED-NEXT:    [[TMP5:%.*]] = call float @llvm.vector.reduce.fadd.v4f32(float -0.000000e+00, <4 x float> [[TMP3]])
; CHECK-UNORDERED-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-UNORDERED-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
; CHECK-UNORDERED:       scalar.ph:
; CHECK-UNORDERED-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-UNORDERED-NEXT:    [[BC_MERGE_RDX:%.*]] = phi float [ 0.000000e+00, [[ENTRY]] ], [ [[TMP5]], [[MIDDLE_BLOCK]] ]
; CHECK-UNORDERED-NEXT:    br label [[FOR_BODY:%.*]]
; CHECK-UNORDERED:       for.body:
; CHECK-UNORDERED-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-UNORDERED-NEXT:    [[SUM_07:%.*]] = phi float [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ADD:%.*]], [[FOR_BODY]] ]
; CHECK-UNORDERED-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[IV]]
; CHECK-UNORDERED-NEXT:    [[TMP6:%.*]] = load float, ptr [[ARRAYIDX]], align 4
; CHECK-UNORDERED-NEXT:    [[ADD]] = fadd float [[TMP6]], [[SUM_07]]
; CHECK-UNORDERED-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-UNORDERED-NEXT:    [[EXITCOND_NOT:%.*]] = icmp eq i64 [[IV_NEXT]], [[N]]
; CHECK-UNORDERED-NEXT:    br i1 [[EXITCOND_NOT]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK-UNORDERED:       for.end:
; CHECK-UNORDERED-NEXT:    [[ADD_LCSSA:%.*]] = phi float [ [[ADD]], [[FOR_BODY]] ], [ [[TMP5]], [[MIDDLE_BLOCK]] ]
; CHECK-UNORDERED-NEXT:    ret float [[ADD_LCSSA]]
;
; CHECK-NOT-VECTORIZED-LABEL: @fadd(
; CHECK-NOT-VECTORIZED-NEXT:  entry:
; CHECK-NOT-VECTORIZED-NEXT:    br label [[FOR_BODY:%.*]]
; CHECK-NOT-VECTORIZED:       for.body:
; CHECK-NOT-VECTORIZED-NEXT:    [[IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NOT-VECTORIZED-NEXT:    [[SUM_07:%.*]] = phi float [ 0.000000e+00, [[ENTRY]] ], [ [[ADD:%.*]], [[FOR_BODY]] ]
; CHECK-NOT-VECTORIZED-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A:%.*]], i64 [[IV]]
; CHECK-NOT-VECTORIZED-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
; CHECK-NOT-VECTORIZED-NEXT:    [[ADD]] = fadd float [[TMP0]], [[SUM_07]]
; CHECK-NOT-VECTORIZED-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NOT-VECTORIZED-NEXT:    [[EXITCOND_NOT:%.*]] = icmp eq i64 [[IV_NEXT]], [[N:%.*]]
; CHECK-NOT-VECTORIZED-NEXT:    br i1 [[EXITCOND_NOT]], label [[FOR_END:%.*]], label [[FOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK-NOT-VECTORIZED:       for.end:
; CHECK-NOT-VECTORIZED-NEXT:    [[ADD_LCSSA:%.*]] = phi float [ [[ADD]], [[FOR_BODY]] ]
; CHECK-NOT-VECTORIZED-NEXT:    ret float [[ADD_LCSSA]]
;
; CHECK-ORDERED-LABEL: @fadd(
; CHECK-ORDERED-NEXT:  entry:
; CHECK-ORDERED-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N:%.*]], 4
; CHECK-ORDERED-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK-ORDERED:       vector.ph:
; CHECK-ORDERED-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-ORDERED-NEXT:    [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-ORDERED-NEXT:    br label [[VECTOR_BODY:%.*]]
; CHECK-ORDERED:       vector.body:
; CHECK-ORDERED-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-ORDERED-NEXT:    [[VEC_PHI:%.*]] = phi float [ 0.000000e+00, [[VECTOR_PH]] ], [ [[TMP3:%.*]], [[VECTOR_BODY]] ]
; CHECK-ORDERED-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-ORDERED-NEXT:    [[TMP1:%.*]] = getelementptr inbounds float, ptr [[A:%.*]], i64 [[TMP0]]
; CHECK-ORDERED-NEXT:    [[TMP2:%.*]] = getelementptr inbounds float, ptr [[TMP1]], i32 0
; CHECK-ORDERED-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP2]], align 4
; CHECK-ORDERED-NEXT:    [[TMP3]] = call float @llvm.vector.reduce.fadd.v4f32(float [[VEC_PHI]], <4 x float> [[WIDE_LOAD]])
; CHECK-ORDERED-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-ORDERED-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-ORDERED-NEXT:    br i1 [[TMP4]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK-ORDERED:       middle.block:
; CHECK-ORDERED-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-ORDERED-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
; CHECK-ORDERED:       scalar.ph:
; CHECK-ORDERED-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-ORDERED-NEXT:    [[BC_MERGE_RDX:%.*]] = phi float [ 0.000000e+00, [[ENTRY]] ], [ [[TMP3]], [[MIDDLE_BLOCK]] ]
; CHECK-ORDERED-NEXT:    br label [[FOR_BODY:%.*]]
; CHECK-ORDERED:       for.body:
; CHECK-ORDERED-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-ORDERED-NEXT:    [[SUM_07:%.*]] = phi float [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ADD:%.*]], [[FOR_BODY]] ]
; CHECK-ORDERED-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[IV]]
; CHECK-ORDERED-NEXT:    [[TMP5:%.*]] = load float, ptr [[ARRAYIDX]], align 4
; CHECK-ORDERED-NEXT:    [[ADD]] = fadd float [[TMP5]], [[SUM_07]]
; CHECK-ORDERED-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-ORDERED-NEXT:    [[EXITCOND_NOT:%.*]] = icmp eq i64 [[IV_NEXT]], [[N]]
; CHECK-ORDERED-NEXT:    br i1 [[EXITCOND_NOT]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK-ORDERED:       for.end:
; CHECK-ORDERED-NEXT:    [[ADD_LCSSA:%.*]] = phi float [ [[ADD]], [[FOR_BODY]] ], [ [[TMP3]], [[MIDDLE_BLOCK]] ]
; CHECK-ORDERED-NEXT:    ret float [[ADD_LCSSA]]
;
entry:
  br label %for.body

for.body:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
  %sum.07 = phi float [ 0.000000e+00, %entry ], [ %add, %for.body ]
  %arrayidx = getelementptr inbounds float, ptr %a, i64 %iv
  %0 = load float, ptr %arrayidx, align 4
  %add = fadd float %0, %sum.07
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond.not = icmp eq i64 %iv.next, %n
  br i1 %exitcond.not, label %for.end, label %for.body, !llvm.loop !0

for.end:
  ret float %add
}

attributes #0 = { "target-features"="+f,+v" }
!0 = distinct !{!0, !1, !2}
!1 = !{!"llvm.loop.vectorize.width", i32 4}
!2 = !{!"llvm.loop.interleave.count", i32 1}