; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s -passes=loop-vectorize,dce,instcombine -S | FileCheck -check-prefix=GFX9 %s
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mcpu=fiji < %s -passes=loop-vectorize,dce,instcombine -S | FileCheck -check-prefix=VI %s
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii < %s -passes=loop-vectorize,dce,instcombine -S | FileCheck -check-prefix=CI %s

define half @vectorize_v2f16_loop(ptr addrspace(1) noalias %s) {
; GFX9-LABEL: @vectorize_v2f16_loop(
; GFX9-NEXT:  entry:
; GFX9-NEXT:    br label [[VECTOR_PH:%.*]]
; GFX9:       vector.ph:
; GFX9-NEXT:    br label [[VECTOR_BODY:%.*]]
; GFX9:       vector.body:
; GFX9-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; GFX9-NEXT:    [[VEC_PHI:%.*]] = phi <2 x half> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP2:%.*]], [[VECTOR_BODY]] ]
; GFX9-NEXT:    [[VEC_PHI1:%.*]] = phi <2 x half> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP3:%.*]], [[VECTOR_BODY]] ]
; GFX9-NEXT:    [[TMP0:%.*]] = getelementptr inbounds [2 x i8], ptr addrspace(1) [[S:%.*]], i64 [[INDEX]]
; GFX9-NEXT:    [[TMP1:%.*]] = getelementptr inbounds nuw i8, ptr addrspace(1) [[TMP0]], i64 4
; GFX9-NEXT:    [[WIDE_LOAD:%.*]] = load <2 x half>, ptr addrspace(1) [[TMP0]], align 2
; GFX9-NEXT:    [[WIDE_LOAD2:%.*]] = load <2 x half>, ptr addrspace(1) [[TMP1]], align 2
; GFX9-NEXT:    [[TMP2]] = fadd fast <2 x half> [[VEC_PHI]], [[WIDE_LOAD]]
; GFX9-NEXT:    [[TMP3]] = fadd fast <2 x half> [[VEC_PHI1]], [[WIDE_LOAD2]]
; GFX9-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; GFX9-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], 256
; GFX9-NEXT:    br i1 [[TMP4]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; GFX9:       middle.block:
; GFX9-NEXT:    br label [[FOR_BODY:%.*]]
; GFX9:       for.end:
; GFX9-NEXT:    [[BIN_RDX:%.*]] = fadd fast <2 x half> [[TMP3]], [[TMP2]]
; GFX9-NEXT:    [[ADD_LCSSA:%.*]] = call fast half @llvm.vector.reduce.fadd.v2f16(half 0xH0000, <2 x half> [[BIN_RDX]])
; GFX9-NEXT:    ret half [[ADD_LCSSA]]
;
; VI-LABEL: @vectorize_v2f16_loop(
; VI-NEXT:  entry:
; VI-NEXT:    br label [[VECTOR_PH:%.*]]
; VI:       vector.ph:
; VI-NEXT:    br label [[VECTOR_BODY:%.*]]
; VI:       vector.body:
; VI-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; VI-NEXT:    [[VEC_PHI:%.*]] = phi <2 x half> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP2:%.*]], [[VECTOR_BODY]] ]
; VI-NEXT:    [[VEC_PHI1:%.*]] = phi <2 x half> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP3:%.*]], [[VECTOR_BODY]] ]
; VI-NEXT:    [[TMP0:%.*]] = getelementptr inbounds [2 x i8], ptr addrspace(1) [[S:%.*]], i64 [[INDEX]]
; VI-NEXT:    [[TMP1:%.*]] = getelementptr inbounds nuw i8, ptr addrspace(1) [[TMP0]], i64 4
; VI-NEXT:    [[WIDE_LOAD:%.*]] = load <2 x half>, ptr addrspace(1) [[TMP0]], align 2
; VI-NEXT:    [[WIDE_LOAD2:%.*]] = load <2 x half>, ptr addrspace(1) [[TMP1]], align 2
; VI-NEXT:    [[TMP2]] = fadd fast <2 x half> [[VEC_PHI]], [[WIDE_LOAD]]
; VI-NEXT:    [[TMP3]] = fadd fast <2 x half> [[VEC_PHI1]], [[WIDE_LOAD2]]
; VI-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; VI-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], 256
; VI-NEXT:    br i1 [[TMP4]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; VI:       middle.block:
; VI-NEXT:    br label [[FOR_BODY:%.*]]
; VI:       for.end:
; VI-NEXT:    [[BIN_RDX:%.*]] = fadd fast <2 x half> [[TMP3]], [[TMP2]]
; VI-NEXT:    [[ADD_LCSSA:%.*]] = call fast half @llvm.vector.reduce.fadd.v2f16(half 0xH0000, <2 x half> [[BIN_RDX]])
; VI-NEXT:    ret half [[ADD_LCSSA]]
;
; CI-LABEL: @vectorize_v2f16_loop(
; CI-NEXT:  entry:
; CI-NEXT:    br label [[FOR_BODY:%.*]]
; CI:       for.body:
; CI-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CI-NEXT:    [[Q_04:%.*]] = phi half [ 0xH0000, [[ENTRY]] ], [ [[ADD:%.*]], [[FOR_BODY]] ]
; CI-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [2 x i8], ptr addrspace(1) [[S:%.*]], i64 [[INDVARS_IV]]
; CI-NEXT:    [[TMP0:%.*]] = load half, ptr addrspace(1) [[ARRAYIDX]], align 2
; CI-NEXT:    [[ADD]] = fadd fast half [[Q_04]], [[TMP0]]
; CI-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CI-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 256
; CI-NEXT:    br i1 [[EXITCOND]], label [[FOR_END:%.*]], label [[FOR_BODY]]
; CI:       for.end:
; CI-NEXT:    ret half [[ADD]]
;
entry:
  br label %for.body

for.body:
  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
  %q.04 = phi half [ 0.0, %entry ], [ %add, %for.body ]
  %arrayidx = getelementptr inbounds half, ptr addrspace(1) %s, i64 %indvars.iv
  %0 = load half, ptr addrspace(1) %arrayidx, align 2
  %add = fadd fast half %q.04, %0
  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
  %exitcond = icmp eq i64 %indvars.iv.next, 256
  br i1 %exitcond, label %for.end, label %for.body

for.end:
  %add.lcssa = phi half [ %add, %for.body ]
  ret half %add.lcssa
}