; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=arm-eabi -mattr=+armv8.2-a,+fullfp16,+neon -float-abi=hard < %s | FileCheck %s --check-prefix=CHECKHARD
; RUN: llc -mtriple=arm-eabi -mattr=+armv8.2-a,+fullfp16,+neon -float-abi=soft < %s | FileCheck %s --check-prefix=CHECKSOFT

define float @test_vget_lane_f16_1(<4 x half> %a) nounwind {
; CHECKHARD-LABEL: test_vget_lane_f16_1:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vcvtt.f32.f16 s0, s0
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: test_vget_lane_f16_1:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d0, r0, r1
; CHECKSOFT-NEXT:    vcvtt.f32.f16 s0, s0
; CHECKSOFT-NEXT:    vmov r0, s0
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <4 x half> %a, i32 1
  %conv = fpext half %elt to float
  ret float %conv
}

define float @test_vget_lane_f16_2(<4 x half> %a) nounwind {
; CHECKHARD-LABEL: test_vget_lane_f16_2:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vcvtb.f32.f16 s0, s1
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: test_vget_lane_f16_2:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d0, r0, r1
; CHECKSOFT-NEXT:    vcvtb.f32.f16 s0, s1
; CHECKSOFT-NEXT:    vmov r0, s0
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <4 x half> %a, i32 2
  %conv = fpext half %elt to float
  ret float %conv
}

define float @test_vget_laneq_f16_6(<8 x half> %a) nounwind {
; CHECKHARD-LABEL: test_vget_laneq_f16_6:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vcvtb.f32.f16 s0, s3
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: test_vget_laneq_f16_6:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d1, r2, r3
; CHECKSOFT-NEXT:    vcvtb.f32.f16 s0, s3
; CHECKSOFT-NEXT:    vmov r0, s0
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <8 x half> %a, i32 6
  %conv = fpext half %elt to float
  ret float %conv
}

define float @test_vget_laneq_f16_7(<8 x half> %a) nounwind {
; CHECKHARD-LABEL: test_vget_laneq_f16_7:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vcvtt.f32.f16 s0, s3
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: test_vget_laneq_f16_7:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d1, r2, r3
; CHECKSOFT-NEXT:    vcvtt.f32.f16 s0, s3
; CHECKSOFT-NEXT:    vmov r0, s0
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <8 x half> %a, i32 7
  %conv = fpext half %elt to float
  ret float %conv
}

define <4 x half> @insert_v4f16(half %a) {
; CHECKHARD-LABEL: insert_v4f16:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    @ kill: def $s0 killed $s0 def $d0
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: insert_v4f16:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov.f16 s0, r0
; CHECKSOFT-NEXT:    vmov r0, r1, d0
; CHECKSOFT-NEXT:    bx lr
entry:
  %res = insertelement <4 x half> undef, half %a, i32 0
  ret <4 x half> %res
}

define <8 x half> @insert_v8f16(half %a) {
; CHECKHARD-LABEL: insert_v8f16:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    @ kill: def $s0 killed $s0 def $q0
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: insert_v8f16:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov.f16 s0, r0
; CHECKSOFT-NEXT:    vmov r2, r3, d1
; CHECKSOFT-NEXT:    vmov r0, r1, d0
; CHECKSOFT-NEXT:    bx lr
entry:
  %res = insertelement <8 x half> undef, half %a, i32 0
  ret <8 x half> %res
}

define <4 x half> @test_vset_lane_f16(<4 x half> %a, float %fb) nounwind {
; CHECKHARD-LABEL: test_vset_lane_f16:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vcvtt.f16.f32 s1, s2
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: test_vset_lane_f16:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d0, r0, r1
; CHECKSOFT-NEXT:    vmov s2, r2
; CHECKSOFT-NEXT:    vcvtt.f16.f32 s1, s2
; CHECKSOFT-NEXT:    vmov r0, r1, d0
; CHECKSOFT-NEXT:    bx lr
entry:
  %b = fptrunc float %fb to half
  %x = insertelement <4 x half> %a, half %b, i32 3
  ret <4 x half> %x
}

define <8 x half> @test_vset_laneq_f16_1(<8 x half> %a, float %fb) nounwind {
; CHECKHARD-LABEL: test_vset_laneq_f16_1:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vcvtt.f16.f32 s0, s4
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: test_vset_laneq_f16_1:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d1, r2, r3
; CHECKSOFT-NEXT:    vldr s4, [sp]
; CHECKSOFT-NEXT:    vmov d0, r0, r1
; CHECKSOFT-NEXT:    vcvtt.f16.f32 s0, s4
; CHECKSOFT-NEXT:    vmov r2, r3, d1
; CHECKSOFT-NEXT:    vmov r0, r1, d0
; CHECKSOFT-NEXT:    bx lr
entry:
  %b = fptrunc float %fb to half
  %x = insertelement <8 x half> %a, half %b, i32 1
  ret <8 x half> %x
}

define <8 x half> @test_vset_laneq_f16_7(<8 x half> %a, float %fb) nounwind {
; CHECKHARD-LABEL: test_vset_laneq_f16_7:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vcvtt.f16.f32 s3, s4
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: test_vset_laneq_f16_7:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d1, r2, r3
; CHECKSOFT-NEXT:    vldr s4, [sp]
; CHECKSOFT-NEXT:    vmov d0, r0, r1
; CHECKSOFT-NEXT:    vcvtt.f16.f32 s3, s4
; CHECKSOFT-NEXT:    vmov r0, r1, d0
; CHECKSOFT-NEXT:    vmov r2, r3, d1
; CHECKSOFT-NEXT:    bx lr
entry:
  %b = fptrunc float %fb to half
  %x = insertelement <8 x half> %a, half %b, i32 7
  ret <8 x half> %x
}

define arm_aapcs_vfpcc <8 x half> @shuffle3step_f16(<32 x half> %src) {
; CHECKHARD-LABEL: shuffle3step_f16:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vmov r1, s0
; CHECKHARD-NEXT:    vmovx.f16 s12, s1
; CHECKHARD-NEXT:    vmov r0, s12
; CHECKHARD-NEXT:    vrev32.16 d16, d3
; CHECKHARD-NEXT:    vext.16 d17, d4, d5, #2
; CHECKHARD-NEXT:    vmovx.f16 s12, s4
; CHECKHARD-NEXT:    vext.16 d16, d16, d3, #1
; CHECKHARD-NEXT:    vext.16 d16, d17, d16, #2
; CHECKHARD-NEXT:    vext.16 d16, d16, d17, #1
; CHECKHARD-NEXT:    vext.16 d17, d16, d16, #1
; CHECKHARD-NEXT:    vmov.16 d16[0], r1
; CHECKHARD-NEXT:    vmov.16 d16[1], r0
; CHECKHARD-NEXT:    vmov r0, s3
; CHECKHARD-NEXT:    vmov.16 d16[2], r0
; CHECKHARD-NEXT:    vmov r0, s12
; CHECKHARD-NEXT:    vmovx.f16 s12, s0
; CHECKHARD-NEXT:    vmov r1, s12
; CHECKHARD-NEXT:    vmovx.f16 s12, s3
; CHECKHARD-NEXT:    vmov.16 d16[3], r0
; CHECKHARD-NEXT:    vmov r0, s2
; CHECKHARD-NEXT:    vmov.16 d18[0], r1
; CHECKHARD-NEXT:    vmov.16 d18[1], r0
; CHECKHARD-NEXT:    vmov r0, s12
; CHECKHARD-NEXT:    vdup.16 q3, d3[1]
; CHECKHARD-NEXT:    vmov r1, s12
; CHECKHARD-NEXT:    vmovx.f16 s12, s9
; CHECKHARD-NEXT:    vmov.16 d18[2], r0
; CHECKHARD-NEXT:    vmov r0, s5
; CHECKHARD-NEXT:    vmov.16 d18[3], r0
; CHECKHARD-NEXT:    vmov r0, s8
; CHECKHARD-NEXT:    vmov.16 d19[0], r1
; CHECKHARD-NEXT:    vmov.16 d19[1], r0
; CHECKHARD-NEXT:    vmov r0, s12
; CHECKHARD-NEXT:    vmov.16 d19[2], r0
; CHECKHARD-NEXT:    vmov r0, s11
; CHECKHARD-NEXT:    vmov.16 d19[3], r0
; CHECKHARD-NEXT:    vadd.f16 q8, q8, q9
; CHECKHARD-NEXT:    vext.16 d18, d0, d1, #2
; CHECKHARD-NEXT:    vmovx.f16 s0, s8
; CHECKHARD-NEXT:    vmov r0, s0
; CHECKHARD-NEXT:    vdup.16 q0, d3[2]
; CHECKHARD-NEXT:    vext.16 d19, d18, d2, #3
; CHECKHARD-NEXT:    vmov r1, s0
; CHECKHARD-NEXT:    vext.16 d18, d2, d18, #1
; CHECKHARD-NEXT:    vmovx.f16 s0, s11
; CHECKHARD-NEXT:    vext.16 d18, d18, d19, #2
; CHECKHARD-NEXT:    vext.16 d18, d18, d18, #1
; CHECKHARD-NEXT:    vmov.16 d19[0], r1
; CHECKHARD-NEXT:    vmov.16 d19[1], r0
; CHECKHARD-NEXT:    vmov r0, s10
; CHECKHARD-NEXT:    vmov.16 d19[2], r0
; CHECKHARD-NEXT:    vmov r0, s0
; CHECKHARD-NEXT:    vmov.16 d19[3], r0
; CHECKHARD-NEXT:    vadd.f16 q0, q8, q9
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: shuffle3step_f16:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov r1, s0
; CHECKSOFT-NEXT:    vmovx.f16 s12, s1
; CHECKSOFT-NEXT:    vmov r0, s12
; CHECKSOFT-NEXT:    vrev32.16 d16, d3
; CHECKSOFT-NEXT:    vext.16 d17, d4, d5, #2
; CHECKSOFT-NEXT:    vmovx.f16 s12, s4
; CHECKSOFT-NEXT:    vext.16 d16, d16, d3, #1
; CHECKSOFT-NEXT:    vext.16 d16, d17, d16, #2
; CHECKSOFT-NEXT:    vext.16 d16, d16, d17, #1
; CHECKSOFT-NEXT:    vext.16 d17, d16, d16, #1
; CHECKSOFT-NEXT:    vmov.16 d16[0], r1
; CHECKSOFT-NEXT:    vmov.16 d16[1], r0
; CHECKSOFT-NEXT:    vmov r0, s3
; CHECKSOFT-NEXT:    vmov.16 d16[2], r0
; CHECKSOFT-NEXT:    vmov r0, s12
; CHECKSOFT-NEXT:    vmovx.f16 s12, s0
; CHECKSOFT-NEXT:    vmov r1, s12
; CHECKSOFT-NEXT:    vmovx.f16 s12, s3
; CHECKSOFT-NEXT:    vmov.16 d16[3], r0
; CHECKSOFT-NEXT:    vmov r0, s2
; CHECKSOFT-NEXT:    vmov.16 d18[0], r1
; CHECKSOFT-NEXT:    vmov.16 d18[1], r0
; CHECKSOFT-NEXT:    vmov r0, s12
; CHECKSOFT-NEXT:    vdup.16 q3, d3[1]
; CHECKSOFT-NEXT:    vmov r1, s12
; CHECKSOFT-NEXT:    vmovx.f16 s12, s9
; CHECKSOFT-NEXT:    vmov.16 d18[2], r0
; CHECKSOFT-NEXT:    vmov r0, s5
; CHECKSOFT-NEXT:    vmov.16 d18[3], r0
; CHECKSOFT-NEXT:    vmov r0, s8
; CHECKSOFT-NEXT:    vmov.16 d19[0], r1
; CHECKSOFT-NEXT:    vmov.16 d19[1], r0
; CHECKSOFT-NEXT:    vmov r0, s12
; CHECKSOFT-NEXT:    vmov.16 d19[2], r0
; CHECKSOFT-NEXT:    vmov r0, s11
; CHECKSOFT-NEXT:    vmov.16 d19[3], r0
; CHECKSOFT-NEXT:    vadd.f16 q8, q8, q9
; CHECKSOFT-NEXT:    vext.16 d18, d0, d1, #2
; CHECKSOFT-NEXT:    vmovx.f16 s0, s8
; CHECKSOFT-NEXT:    vmov r0, s0
; CHECKSOFT-NEXT:    vdup.16 q0, d3[2]
; CHECKSOFT-NEXT:    vext.16 d19, d18, d2, #3
; CHECKSOFT-NEXT:    vmov r1, s0
; CHECKSOFT-NEXT:    vext.16 d18, d2, d18, #1
; CHECKSOFT-NEXT:    vmovx.f16 s0, s11
; CHECKSOFT-NEXT:    vext.16 d18, d18, d19, #2
; CHECKSOFT-NEXT:    vext.16 d18, d18, d18, #1
; CHECKSOFT-NEXT:    vmov.16 d19[0], r1
; CHECKSOFT-NEXT:    vmov.16 d19[1], r0
; CHECKSOFT-NEXT:    vmov r0, s10
; CHECKSOFT-NEXT:    vmov.16 d19[2], r0
; CHECKSOFT-NEXT:    vmov r0, s0
; CHECKSOFT-NEXT:    vmov.16 d19[3], r0
; CHECKSOFT-NEXT:    vadd.f16 q0, q8, q9
; CHECKSOFT-NEXT:    bx lr
entry:
  %s1 = shufflevector <32 x half> %src, <32 x half> undef, <8 x i32> <i32 0, i32 3, i32 6, i32 9, i32 12, i32 15, i32 18, i32 21>
  %s2 = shufflevector <32 x half> %src, <32 x half> undef, <8 x i32> <i32 1, i32 4, i32 7, i32 10, i32 13, i32 16, i32 19, i32 22>
  %s3 = shufflevector <32 x half> %src, <32 x half> undef, <8 x i32> <i32 2, i32 5, i32 8, i32 11, i32 14, i32 17, i32 20, i32 23>
  %a = fadd <8 x half> %s1, %s2
  %r = fadd <8 x half> %a, %s3
  ret <8 x half> %r
}

define i16 @extract_v4i16(<4 x half> %a) {
; CHECKHARD-LABEL: extract_v4i16:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vmov.u16 r0, d0[0]
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: extract_v4i16:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d16, r0, r1
; CHECKSOFT-NEXT:    vmov.u16 r0, d16[0]
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <4 x half> %a, i32 0
  %t = bitcast half %elt to i16
  ret i16 %t
}

define i16 @extract_v8i16(<8 x half> %a) {
; CHECKHARD-LABEL: extract_v8i16:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vmov.u16 r0, d0[0]
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: extract_v8i16:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d16, r0, r1
; CHECKSOFT-NEXT:    vmov.u16 r0, d16[0]
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <8 x half> %a, i32 0
  %t = bitcast half %elt to i16
  ret i16 %t
}

define i32 @extract_v4s32(<4 x half> %a) {
; CHECKHARD-LABEL: extract_v4s32:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vmov.u16 r0, d0[0]
; CHECKHARD-NEXT:    sxth r0, r0
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: extract_v4s32:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d16, r0, r1
; CHECKSOFT-NEXT:    vmov.u16 r0, d16[0]
; CHECKSOFT-NEXT:    sxth r0, r0
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <4 x half> %a, i32 0
  %t = bitcast half %elt to i16
  %s = sext i16 %t to i32
  ret i32 %s
}

define i32 @extract_v8s32(<8 x half> %a) {
; CHECKHARD-LABEL: extract_v8s32:
; CHECKHARD:       @ %bb.0: @ %entry
; CHECKHARD-NEXT:    vmov.u16 r0, d0[0]
; CHECKHARD-NEXT:    sxth r0, r0
; CHECKHARD-NEXT:    bx lr
;
; CHECKSOFT-LABEL: extract_v8s32:
; CHECKSOFT:       @ %bb.0: @ %entry
; CHECKSOFT-NEXT:    vmov d16, r0, r1
; CHECKSOFT-NEXT:    vmov.u16 r0, d16[0]
; CHECKSOFT-NEXT:    sxth r0, r0
; CHECKSOFT-NEXT:    bx lr
entry:
  %elt = extractelement <8 x half> %a, i32 0
  %t = bitcast half %elt to i16
  %s = sext i16 %t to i32
  ret i32 %s
}