llvm/test/Transforms/LoopVectorize/X86/conversion-cost.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6
; RUN: opt < %s -passes=loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx -S | FileCheck %s

target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.8.0"

define void @conversion_cost1(i32 %n, ptr nocapture %A, ptr nocapture %B) {
; CHECK-LABEL: define void @conversion_cost1(
; CHECK-SAME: i32 [[N:%.*]], ptr captures(none) [[A:%.*]], ptr captures(none) [[B:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEXT:    [[TMP1:%.*]] = icmp sgt i32 [[N]], 3
; CHECK-NEXT:    br i1 [[TMP1]], label %[[ITER_CHECK:.*]], [[DOT_CRIT_EDGE:label %.*]]
; CHECK:       [[ITER_CHECK]]:
; CHECK-NEXT:    [[TMP2:%.*]] = add i32 [[N]], -3
; CHECK-NEXT:    [[TMP3:%.*]] = zext i32 [[TMP2]] to i64
; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP3]], 8
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[VEC_EPILOG_SCALAR_PH:.*]], label %[[VECTOR_MAIN_LOOP_ITER_CHECK:.*]]
; CHECK:       [[VECTOR_MAIN_LOOP_ITER_CHECK]]:
; CHECK-NEXT:    [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[TMP3]], 64
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK1]], label %[[VEC_EPILOG_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = and i64 [[TMP3]], 63
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[TMP3]], [[N_MOD_VF]]
; CHECK-NEXT:    [[TMP4:%.*]] = add i64 3, [[N_VEC]]
; 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 <16 x i8> [ <i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18>, %[[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[STEP_ADD:%.*]] = add <16 x i8> [[VEC_IND]], splat (i8 16)
; CHECK-NEXT:    [[STEP_ADD_2:%.*]] = add <16 x i8> [[STEP_ADD]], splat (i8 16)
; CHECK-NEXT:    [[STEP_ADD_3:%.*]] = add <16 x i8> [[STEP_ADD_2]], splat (i8 16)
; CHECK-NEXT:    [[TMP5:%.*]] = add i64 3, [[INDEX]]
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP5]]
; CHECK-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i8, ptr [[TMP6]], i64 16
; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[TMP6]], i64 32
; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP6]], i64 48
; CHECK-NEXT:    store <16 x i8> [[VEC_IND]], ptr [[TMP6]], align 1
; CHECK-NEXT:    store <16 x i8> [[STEP_ADD]], ptr [[TMP15]], align 1
; CHECK-NEXT:    store <16 x i8> [[STEP_ADD_2]], ptr [[TMP16]], align 1
; CHECK-NEXT:    store <16 x i8> [[STEP_ADD_3]], ptr [[TMP17]], align 1
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 64
; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <16 x i8> [[STEP_ADD_3]], splat (i8 16)
; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP7]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP3]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[CMP_N]], [[DOT_CRIT_EDGE_LOOPEXIT:label %.*]], label %[[VEC_EPILOG_ITER_CHECK:.*]]
; CHECK:       [[VEC_EPILOG_ITER_CHECK]]:
; CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_MOD_VF]], 8
; CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK]], label %[[VEC_EPILOG_SCALAR_PH]], label %[[VEC_EPILOG_PH]], !prof [[PROF3:![0-9]+]]
; CHECK:       [[VEC_EPILOG_PH]]:
; CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[VEC_EPILOG_ITER_CHECK]] ], [ 0, %[[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[TMP4]], %[[VEC_EPILOG_ITER_CHECK]] ], [ 3, %[[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT:    [[N_MOD_VF2:%.*]] = and i64 [[TMP3]], 7
; CHECK-NEXT:    [[N_VEC3:%.*]] = sub i64 [[TMP3]], [[N_MOD_VF2]]
; CHECK-NEXT:    [[TMP8:%.*]] = add i64 3, [[N_VEC3]]
; CHECK-NEXT:    [[TMP9:%.*]] = trunc i64 [[BC_RESUME_VAL]] to i8
; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <8 x i8> poison, i8 [[TMP9]], i64 0
; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <8 x i8> [[BROADCAST_SPLATINSERT]], <8 x i8> poison, <8 x i32> zeroinitializer
; CHECK-NEXT:    [[INDUCTION:%.*]] = add <8 x i8> [[BROADCAST_SPLAT]], <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>
; CHECK-NEXT:    br label %[[VEC_EPILOG_VECTOR_BODY:.*]]
; CHECK:       [[VEC_EPILOG_VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX4:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], %[[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT6:%.*]], %[[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT:    [[VEC_IND5:%.*]] = phi <8 x i8> [ [[INDUCTION]], %[[VEC_EPILOG_PH]] ], [ [[VEC_IND_NEXT7:%.*]], %[[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP10:%.*]] = add i64 3, [[INDEX4]]
; CHECK-NEXT:    [[TMP11:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP10]]
; CHECK-NEXT:    store <8 x i8> [[VEC_IND5]], ptr [[TMP11]], align 1
; CHECK-NEXT:    [[INDEX_NEXT6]] = add nuw i64 [[INDEX4]], 8
; CHECK-NEXT:    [[VEC_IND_NEXT7]] = add <8 x i8> [[VEC_IND5]], splat (i8 8)
; CHECK-NEXT:    [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT6]], [[N_VEC3]]
; CHECK-NEXT:    br i1 [[TMP12]], label %[[VEC_EPILOG_MIDDLE_BLOCK:.*]], label %[[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK:       [[VEC_EPILOG_MIDDLE_BLOCK]]:
; CHECK-NEXT:    [[CMP_N8:%.*]] = icmp eq i64 [[TMP3]], [[N_VEC3]]
; CHECK-NEXT:    br i1 [[CMP_N8]], [[DOT_CRIT_EDGE_LOOPEXIT]], label %[[VEC_EPILOG_SCALAR_PH]]
; CHECK:       [[VEC_EPILOG_SCALAR_PH]]:
; CHECK-NEXT:    [[BC_RESUME_VAL8:%.*]] = phi i64 [ [[TMP8]], %[[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[TMP4]], %[[VEC_EPILOG_ITER_CHECK]] ], [ 3, %[[ITER_CHECK]] ]
; CHECK-NEXT:    br label %[[DOTLR_PH:.*]]
; CHECK:       [[_LR_PH:.*:]]
; CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], %[[DOTLR_PH]] ], [ [[BC_RESUME_VAL8]], %[[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT:    [[TMP13:%.*]] = trunc i64 [[INDVARS_IV]] to i8
; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[INDVARS_IV]]
; CHECK-NEXT:    store i8 [[TMP13]], ptr [[TMP14]], align 1
; CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add i64 [[INDVARS_IV]], 1
; CHECK-NEXT:    [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[N]]
; CHECK-NEXT:    br i1 [[EXITCOND]], [[DOT_CRIT_EDGE_LOOPEXIT]], label %[[DOTLR_PH]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK:       [[__CRIT_EDGE_LOOPEXIT:.*:]]
; CHECK-NEXT:    br [[DOT_CRIT_EDGE]]
; CHECK:       [[__CRIT_EDGE:.*:]]
; CHECK-NEXT:    ret void
;
  %1 = icmp sgt i32 %n, 3
  br i1 %1, label %.lr.ph, label %._crit_edge

.lr.ph:
  %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 3, %0 ]
  %2 = trunc i64 %indvars.iv to i8
  %3 = getelementptr inbounds i8, ptr %A, i64 %indvars.iv
  store i8 %2, ptr %3, align 1
  %indvars.iv.next = add i64 %indvars.iv, 1
  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
  %exitcond = icmp eq i32 %lftr.wideiv, %n
  br i1 %exitcond, label %._crit_edge, label %.lr.ph

._crit_edge:
  ret void
}

define void @conversion_cost2(i32 %n, ptr nocapture %A, ptr nocapture %B) {
; CHECK-LABEL: define void @conversion_cost2(
; CHECK-SAME: i32 [[N:%.*]], ptr captures(none) [[A:%.*]], ptr captures(none) [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[TMP1:%.*]] = icmp sgt i32 [[N]], 9
; CHECK-NEXT:    br i1 [[TMP1]], label %[[DOTLR_PH_PREHEADER:.*]], [[DOT_CRIT_EDGE:label %.*]]
; CHECK:       [[_LR_PH_PREHEADER:.*:]]
; CHECK-NEXT:    [[TMP2:%.*]] = add i32 [[N]], -9
; CHECK-NEXT:    [[TMP3:%.*]] = zext i32 [[TMP2]] to i64
; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP3]], 8
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = and i64 [[TMP3]], 7
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[TMP3]], [[N_MOD_VF]]
; CHECK-NEXT:    [[TMP4:%.*]] = add i64 9, [[N_VEC]]
; 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 <8 x i64> [ <i64 9, i64 10, i64 11, i64 12, i64 13, i64 14, i64 15, i64 16>, %[[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP5:%.*]] = add i64 9, [[INDEX]]
; CHECK-NEXT:    [[TMP6:%.*]] = add nsw <8 x i64> [[VEC_IND]], splat (i64 3)
; CHECK-NEXT:    [[TMP7:%.*]] = sitofp <8 x i64> [[TMP6]] to <8 x float>
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP5]]
; CHECK-NEXT:    store <8 x float> [[TMP7]], ptr [[TMP8]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <8 x i64> [[VEC_IND]], splat (i64 8)
; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP9]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP3]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[CMP_N]], [[DOT_CRIT_EDGE_LOOPEXIT:label %.*]], label %[[SCALAR_PH]]
; CHECK:       [[SCALAR_PH]]:
; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[TMP4]], %[[MIDDLE_BLOCK]] ], [ 9, %[[DOTLR_PH_PREHEADER]] ]
; CHECK-NEXT:    br label %[[DOTLR_PH:.*]]
; CHECK:       [[_LR_PH:.*:]]
; CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], %[[DOTLR_PH]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
; CHECK-NEXT:    [[ADD:%.*]] = add nsw i64 [[INDVARS_IV]], 3
; CHECK-NEXT:    [[TOFP:%.*]] = sitofp i64 [[ADD]] to float
; CHECK-NEXT:    [[GEP:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[INDVARS_IV]]
; CHECK-NEXT:    store float [[TOFP]], ptr [[GEP]], align 4
; CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add i64 [[INDVARS_IV]], 1
; CHECK-NEXT:    [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[N]]
; CHECK-NEXT:    br i1 [[EXITCOND]], [[DOT_CRIT_EDGE_LOOPEXIT]], label %[[DOTLR_PH]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK:       [[__CRIT_EDGE_LOOPEXIT:.*:]]
; CHECK-NEXT:    br [[DOT_CRIT_EDGE]]
; CHECK:       [[__CRIT_EDGE:.*:]]
; CHECK-NEXT:    ret void
;
  %1 = icmp sgt i32 %n, 9
  br i1 %1, label %.lr.ph, label %._crit_edge

.lr.ph:
  %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 9, %0 ]
  %add = add nsw i64 %indvars.iv, 3
  %tofp = sitofp i64 %add to float
  %gep = getelementptr inbounds float, ptr %B, i64 %indvars.iv
  store float %tofp, ptr %gep, align 4
  %indvars.iv.next = add i64 %indvars.iv, 1
  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
  %exitcond = icmp eq i32 %lftr.wideiv, %n
  br i1 %exitcond, label %._crit_edge, label %.lr.ph

._crit_edge:
  ret void
}