llvm/test/Transforms/LoopVectorize/AArch64/partial-reduce-incomplete-chains.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6
; RUN: opt --mattr=+neon,+dotprod -passes=loop-vectorize -force-vector-interleave=1 -enable-epilogue-vectorization=false -S %s | FileCheck %s --check-prefixes=CHECK-NEON

target triple = "arm64-apple-macosx"

define i32 @red_extended_add_incomplete_chain(ptr %start, ptr %end, i32 %offset) {
; CHECK-NEON-LABEL: define i32 @red_extended_add_incomplete_chain(
; CHECK-NEON-SAME: ptr [[START:%.*]], ptr [[END:%.*]], i32 [[OFFSET:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEON-NEXT:  [[ENTRY:.*]]:
; CHECK-NEON-NEXT:    [[START2:%.*]] = ptrtoint ptr [[START]] to i64
; CHECK-NEON-NEXT:    [[END1:%.*]] = ptrtoint ptr [[END]] to i64
; CHECK-NEON-NEXT:    [[TMP0:%.*]] = add i64 [[END1]], 1
; CHECK-NEON-NEXT:    [[TMP1:%.*]] = sub i64 [[TMP0]], [[START2]]
; CHECK-NEON-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP1]], 16
; CHECK-NEON-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK-NEON:       [[VECTOR_PH]]:
; CHECK-NEON-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[TMP1]], 16
; CHECK-NEON-NEXT:    [[N_VEC:%.*]] = sub i64 [[TMP1]], [[N_MOD_VF]]
; CHECK-NEON-NEXT:    [[TMP2:%.*]] = getelementptr i8, ptr [[START]], i64 [[N_VEC]]
; CHECK-NEON-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <16 x i32> poison, i32 [[OFFSET]], i64 0
; CHECK-NEON-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <16 x i32> [[BROADCAST_SPLATINSERT]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK-NEON-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK-NEON:       [[VECTOR_BODY]]:
; CHECK-NEON-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEON-NEXT:    [[VEC_PHI:%.*]] = phi <16 x i32> [ zeroinitializer, %[[VECTOR_PH]] ], [ [[TMP5:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEON-NEXT:    [[NEXT_GEP:%.*]] = getelementptr i8, ptr [[START]], i64 [[INDEX]]
; CHECK-NEON-NEXT:    [[WIDE_LOAD:%.*]] = load <16 x i8>, ptr [[NEXT_GEP]], align 1
; CHECK-NEON-NEXT:    [[TMP3:%.*]] = zext <16 x i8> [[WIDE_LOAD]] to <16 x i32>
; CHECK-NEON-NEXT:    [[TMP4:%.*]] = add <16 x i32> [[VEC_PHI]], [[TMP3]]
; CHECK-NEON-NEXT:    [[TMP5]] = add <16 x i32> [[TMP4]], [[BROADCAST_SPLAT]]
; CHECK-NEON-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEON-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEON-NEXT:    br i1 [[TMP6]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK-NEON:       [[MIDDLE_BLOCK]]:
; CHECK-NEON-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> [[TMP5]])
; CHECK-NEON-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP1]], [[N_VEC]]
; CHECK-NEON-NEXT:    br i1 [[CMP_N]], label %[[EXIT:.*]], label %[[SCALAR_PH]]
; CHECK-NEON:       [[SCALAR_PH]]:
; CHECK-NEON-NEXT:    [[BC_RESUME_VAL:%.*]] = phi ptr [ [[TMP2]], %[[MIDDLE_BLOCK]] ], [ [[START]], %[[ENTRY]] ]
; CHECK-NEON-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ [[TMP7]], %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ]
; CHECK-NEON-NEXT:    br label %[[LOOP:.*]]
; CHECK-NEON:       [[LOOP]]:
; CHECK-NEON-NEXT:    [[PTR_IV:%.*]] = phi ptr [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[GEP_IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEON-NEXT:    [[RED:%.*]] = phi i32 [ [[BC_MERGE_RDX]], %[[SCALAR_PH]] ], [ [[RED_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEON-NEXT:    [[L:%.*]] = load i8, ptr [[PTR_IV]], align 1
; CHECK-NEON-NEXT:    [[L_EXT:%.*]] = zext i8 [[L]] to i32
; CHECK-NEON-NEXT:    [[ADD:%.*]] = add i32 [[RED]], [[L_EXT]]
; CHECK-NEON-NEXT:    [[RED_NEXT]] = add i32 [[ADD]], [[OFFSET]]
; CHECK-NEON-NEXT:    [[GEP_IV_NEXT]] = getelementptr i8, ptr [[PTR_IV]], i64 1
; CHECK-NEON-NEXT:    [[EC:%.*]] = icmp eq ptr [[PTR_IV]], [[END]]
; CHECK-NEON-NEXT:    br i1 [[EC]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK-NEON:       [[EXIT]]:
; CHECK-NEON-NEXT:    [[RED_NEXT_LCSSA:%.*]] = phi i32 [ [[RED_NEXT]], %[[LOOP]] ], [ [[TMP7]], %[[MIDDLE_BLOCK]] ]
; CHECK-NEON-NEXT:    ret i32 [[RED_NEXT_LCSSA]]
;
entry:
  br label %loop

loop:
  %ptr.iv = phi ptr [ %start, %entry ], [ %gep.iv.next, %loop ]
  %red = phi i32 [ 0, %entry ], [ %red.next, %loop ]
  %l = load i8, ptr %ptr.iv, align 1
  %l.ext = zext i8 %l to i32
  %add = add i32 %red, %l.ext
  %red.next = add i32 %add, %offset
  %gep.iv.next = getelementptr i8, ptr %ptr.iv, i64 1
  %ec = icmp eq ptr %ptr.iv, %end
  br i1 %ec, label %exit, label %loop

exit:
  ret i32 %red.next
}