// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5

// RUN: %clang_cc1        -triple aarch64-none-linux-gnu -target-feature +neon -target-feature +bf16 -target-feature +fp8 -target-feature +fp8dot2 -target-feature +fp8dot4 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,sroa | FileCheck %s
// RUN: %clang_cc1 -x c++ -triple aarch64-none-linux-gnu -target-feature +neon -target-feature +bf16 -target-feature +fp8 -target-feature +fp8dot2 -target-feature +fp8dot4 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,sroa | FileCheck %s -check-prefix CHECK-CXX

// RUN: %clang_cc1        -triple aarch64-none-linux-gnu -target-feature +neon -target-feature +bf16 -target-feature +fp8 -target-feature +fp8dot2 -target-feature +fp8dot4 -O3 -Werror -Wall -S -o /dev/null %s

// REQUIRES: aarch64-registered-target

#include <arm_neon.h>

// CHECK-LABEL: define dso_local <4 x half> @test_vdot_f16(
// CHECK-SAME: <4 x half> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x half> [[VD]] to <4 x i16>
// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT2_I:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
// CHECK-NEXT:    [[FDOT21_I:%.*]] = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.v4f16.v8i8(<4 x half> [[FDOT2_I]], <8 x i8> [[VN]], <8 x i8> [[VM]])
// CHECK-NEXT:    ret <4 x half> [[FDOT21_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <4 x half> @_Z13test_vdot_f1613__Float16x4_t13__Mfloat8x8_tS0_m(
// CHECK-CXX-SAME: <4 x half> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = bitcast <4 x half> [[VD]] to <4 x i16>
// CHECK-CXX-NEXT:    [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT2_I:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
// CHECK-CXX-NEXT:    [[FDOT21_I:%.*]] = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.v4f16.v8i8(<4 x half> [[FDOT2_I]], <8 x i8> [[VN]], <8 x i8> [[VM]])
// CHECK-CXX-NEXT:    ret <4 x half> [[FDOT21_I]]
//
float16x4_t test_vdot_f16(float16x4_t vd, mfloat8x8_t vn, mfloat8x8_t vm, fpm_t fpmr) {
  return vdot_f16_mf8_fpm(vd, vn, vm, fpmr);
}

// CHECK-LABEL: define dso_local <8 x half> @test_vdotq_f16(
// CHECK-SAME: <8 x half> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <8 x half> [[VD]] to <8 x i16>
// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT2_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
// CHECK-NEXT:    [[FDOT21_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.v8f16.v16i8(<8 x half> [[FDOT2_I]], <16 x i8> [[VN]], <16 x i8> [[VM]])
// CHECK-NEXT:    ret <8 x half> [[FDOT21_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z14test_vdotq_f1613__Float16x8_t14__Mfloat8x16_tS0_m(
// CHECK-CXX-SAME: <8 x half> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = bitcast <8 x half> [[VD]] to <8 x i16>
// CHECK-CXX-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT2_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
// CHECK-CXX-NEXT:    [[FDOT21_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.v8f16.v16i8(<8 x half> [[FDOT2_I]], <16 x i8> [[VN]], <16 x i8> [[VM]])
// CHECK-CXX-NEXT:    ret <8 x half> [[FDOT21_I]]
//
float16x8_t test_vdotq_f16(float16x8_t vd, mfloat8x16_t vn, mfloat8x16_t vm, fpm_t fpmr) {
  return vdotq_f16_mf8_fpm(vd, vn, vm, fpmr);
}

// CHECK-LABEL: define dso_local <4 x half> @test_vdot_lane_f16(
// CHECK-SAME: <4 x half> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x half> [[VD]] to <4 x i16>
// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-NEXT:    [[TMP2:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
// CHECK-NEXT:    [[FDOT2_LANE1:%.*]] = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v4f16.v8i8(<4 x half> [[FDOT2_LANE]], <8 x i8> [[VN]], <16 x i8> [[TMP2]], i32 3)
// CHECK-NEXT:    ret <4 x half> [[FDOT2_LANE1]]
//
// CHECK-CXX-LABEL: define dso_local noundef <4 x half> @_Z18test_vdot_lane_f1613__Float16x4_t13__Mfloat8x8_tS0_m(
// CHECK-CXX-SAME: <4 x half> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = bitcast <4 x half> [[VD]] to <4 x i16>
// CHECK-CXX-NEXT:    [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-CXX-NEXT:    [[TMP2:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
// CHECK-CXX-NEXT:    [[FDOT2_LANE1:%.*]] = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v4f16.v8i8(<4 x half> [[FDOT2_LANE]], <8 x i8> [[VN]], <16 x i8> [[TMP2]], i32 3)
// CHECK-CXX-NEXT:    ret <4 x half> [[FDOT2_LANE1]]
//
float16x4_t test_vdot_lane_f16(float16x4_t vd, mfloat8x8_t vn, mfloat8x8_t vm, fpm_t fpmr) {
  return vdot_lane_f16_mf8_fpm(vd, vn, vm, 3, fpmr);
}

// CHECK-LABEL: define dso_local <4 x half> @test_vdot_laneq_f16(
// CHECK-SAME: <4 x half> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x half> [[VD]] to <4 x i16>
// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
// CHECK-NEXT:    [[FDOT2_LANE1:%.*]] = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v4f16.v8i8(<4 x half> [[FDOT2_LANE]], <8 x i8> [[VN]], <16 x i8> [[VM]], i32 7)
// CHECK-NEXT:    ret <4 x half> [[FDOT2_LANE1]]
//
// CHECK-CXX-LABEL: define dso_local noundef <4 x half> @_Z19test_vdot_laneq_f1613__Float16x4_t13__Mfloat8x8_t14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <4 x half> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = bitcast <4 x half> [[VD]] to <4 x i16>
// CHECK-CXX-NEXT:    [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
// CHECK-CXX-NEXT:    [[FDOT2_LANE1:%.*]] = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v4f16.v8i8(<4 x half> [[FDOT2_LANE]], <8 x i8> [[VN]], <16 x i8> [[VM]], i32 7)
// CHECK-CXX-NEXT:    ret <4 x half> [[FDOT2_LANE1]]
//
float16x4_t test_vdot_laneq_f16(float16x4_t vd, mfloat8x8_t vn, mfloat8x16_t vm, fpm_t fpmr) {
  return vdot_laneq_f16_mf8_fpm(vd, vn, vm, 7, fpmr);
}

// CHECK-LABEL: define dso_local <8 x half> @test_vdotq_lane_f16(
// CHECK-SAME: <8 x half> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <8 x half> [[VD]] to <8 x i16>
// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-NEXT:    [[TMP2:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
// CHECK-NEXT:    [[FDOT2_LANE1:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v8f16.v16i8(<8 x half> [[FDOT2_LANE]], <16 x i8> [[VN]], <16 x i8> [[TMP2]], i32 3)
// CHECK-NEXT:    ret <8 x half> [[FDOT2_LANE1]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z19test_vdotq_lane_f1613__Float16x8_t14__Mfloat8x16_t13__Mfloat8x8_tm(
// CHECK-CXX-SAME: <8 x half> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = bitcast <8 x half> [[VD]] to <8 x i16>
// CHECK-CXX-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-CXX-NEXT:    [[TMP2:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
// CHECK-CXX-NEXT:    [[FDOT2_LANE1:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v8f16.v16i8(<8 x half> [[FDOT2_LANE]], <16 x i8> [[VN]], <16 x i8> [[TMP2]], i32 3)
// CHECK-CXX-NEXT:    ret <8 x half> [[FDOT2_LANE1]]
//
float16x8_t test_vdotq_lane_f16(float16x8_t vd, mfloat8x16_t vn, mfloat8x8_t vm, fpm_t fpmr) {
  return vdotq_lane_f16_mf8_fpm(vd, vn, vm, 3, fpmr);
}

// CHECK-LABEL: define dso_local <8 x half> @test_vdotq_laneq_f16(
// CHECK-SAME: <8 x half> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <8 x half> [[VD]] to <8 x i16>
// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
// CHECK-NEXT:    [[FDOT2_LANE1:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v8f16.v16i8(<8 x half> [[FDOT2_LANE]], <16 x i8> [[VN]], <16 x i8> [[VM]], i32 7)
// CHECK-NEXT:    ret <8 x half> [[FDOT2_LANE1]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z20test_vdotq_laneq_f1613__Float16x8_t14__Mfloat8x16_tS0_m(
// CHECK-CXX-SAME: <8 x half> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = bitcast <8 x half> [[VD]] to <8 x i16>
// CHECK-CXX-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT2_LANE:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
// CHECK-CXX-NEXT:    [[FDOT2_LANE1:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v8f16.v16i8(<8 x half> [[FDOT2_LANE]], <16 x i8> [[VN]], <16 x i8> [[VM]], i32 7)
// CHECK-CXX-NEXT:    ret <8 x half> [[FDOT2_LANE1]]
//
float16x8_t test_vdotq_laneq_f16(float16x8_t vd, mfloat8x16_t vn, mfloat8x16_t vm, fpm_t fpmr) {
  return vdotq_laneq_f16_mf8_fpm(vd, vn, vm, 7, fpmr);
}

// CHECK-LABEL: define dso_local <2 x float> @test_vdot_f32(
// CHECK-SAME: <2 x float> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT4_I:%.*]] = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.v2f32.v8i8(<2 x float> [[VD]], <8 x i8> [[VN]], <8 x i8> [[VM]])
// CHECK-NEXT:    ret <2 x float> [[FDOT4_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <2 x float> @_Z13test_vdot_f3213__Float32x2_t13__Mfloat8x8_tS0_m(
// CHECK-CXX-SAME: <2 x float> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT4_I:%.*]] = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.v2f32.v8i8(<2 x float> [[VD]], <8 x i8> [[VN]], <8 x i8> [[VM]])
// CHECK-CXX-NEXT:    ret <2 x float> [[FDOT4_I]]
//
float32x2_t test_vdot_f32(float32x2_t vd, mfloat8x8_t vn, mfloat8x8_t vm, fpm_t fpmr) {
  return vdot_f32_mf8_fpm(vd, vn, vm, fpmr);
}

// CHECK-LABEL: define dso_local <4 x float> @test_vdotq_f32(
// CHECK-SAME: <4 x float> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT4_I:%.*]] = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.v4f32.v16i8(<4 x float> [[VD]], <16 x i8> [[VN]], <16 x i8> [[VM]])
// CHECK-NEXT:    ret <4 x float> [[FDOT4_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <4 x float> @_Z14test_vdotq_f3213__Float32x4_t14__Mfloat8x16_tS0_m(
// CHECK-CXX-SAME: <4 x float> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT4_I:%.*]] = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.v4f32.v16i8(<4 x float> [[VD]], <16 x i8> [[VN]], <16 x i8> [[VM]])
// CHECK-CXX-NEXT:    ret <4 x float> [[FDOT4_I]]
//
float32x4_t test_vdotq_f32(float32x4_t vd, mfloat8x16_t vn, mfloat8x16_t vm, fpm_t fpmr) {
  return vdotq_f32_mf8_fpm(vd, vn, vm, fpmr);
}

// CHECK-LABEL: define dso_local <2 x float> @test_vdot_lane_f32(
// CHECK-SAME: <2 x float> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT4_LANE:%.*]] = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v2f32.v8i8(<2 x float> [[VD]], <8 x i8> [[VN]], <16 x i8> [[TMP0]], i32 1)
// CHECK-NEXT:    ret <2 x float> [[FDOT4_LANE]]
//
// CHECK-CXX-LABEL: define dso_local noundef <2 x float> @_Z18test_vdot_lane_f3213__Float32x2_t13__Mfloat8x8_tS0_m(
// CHECK-CXX-SAME: <2 x float> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT4_LANE:%.*]] = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v2f32.v8i8(<2 x float> [[VD]], <8 x i8> [[VN]], <16 x i8> [[TMP0]], i32 1)
// CHECK-CXX-NEXT:    ret <2 x float> [[FDOT4_LANE]]
//
float32x2_t test_vdot_lane_f32(float32x2_t vd, mfloat8x8_t vn, mfloat8x8_t vm, fpm_t fpmr) {
  return vdot_lane_f32_mf8_fpm(vd, vn, vm, 1, fpmr);
}

// CHECK-LABEL: define dso_local <2 x float> @test_vdot_laneq_f32(
// CHECK-SAME: <2 x float> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT4_LANE:%.*]] = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v2f32.v8i8(<2 x float> [[VD]], <8 x i8> [[VN]], <16 x i8> [[VM]], i32 3)
// CHECK-NEXT:    ret <2 x float> [[FDOT4_LANE]]
//
// CHECK-CXX-LABEL: define dso_local noundef <2 x float> @_Z19test_vdot_laneq_f3213__Float32x2_t13__Mfloat8x8_t14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <2 x float> noundef [[VD:%.*]], <8 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT4_LANE:%.*]] = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v2f32.v8i8(<2 x float> [[VD]], <8 x i8> [[VN]], <16 x i8> [[VM]], i32 3)
// CHECK-CXX-NEXT:    ret <2 x float> [[FDOT4_LANE]]
//
float32x2_t test_vdot_laneq_f32(float32x2_t vd, mfloat8x8_t vn, mfloat8x16_t vm, fpm_t fpmr) {
  return vdot_laneq_f32_mf8_fpm(vd, vn, vm, 3, fpmr);
}

// CHECK-LABEL: define dso_local <4 x float> @test_vdotq_lane_f32(
// CHECK-SAME: <4 x float> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[TMP0:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT4_LANE:%.*]] = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v4f32.v16i8(<4 x float> [[VD]], <16 x i8> [[VN]], <16 x i8> [[TMP0]], i32 1)
// CHECK-NEXT:    ret <4 x float> [[FDOT4_LANE]]
//
// CHECK-CXX-LABEL: define dso_local noundef <4 x float> @_Z19test_vdotq_lane_f3213__Float32x4_t14__Mfloat8x16_t13__Mfloat8x8_tm(
// CHECK-CXX-SAME: <4 x float> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <8 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    [[TMP0:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VM]], i64 0)
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT4_LANE:%.*]] = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v4f32.v16i8(<4 x float> [[VD]], <16 x i8> [[VN]], <16 x i8> [[TMP0]], i32 1)
// CHECK-CXX-NEXT:    ret <4 x float> [[FDOT4_LANE]]
//
float32x4_t test_vdotq_lane_f32(float32x4_t vd, mfloat8x16_t vn, mfloat8x8_t vm, fpm_t fpmr) {
  return vdotq_lane_f32_mf8_fpm(vd, vn, vm, 1, fpmr);
}

// CHECK-LABEL: define dso_local <4 x float> @test_vdotq_laneq_f32(
// CHECK-SAME: <4 x float> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-NEXT:    [[FDOT4_LANE:%.*]] = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v4f32.v16i8(<4 x float> [[VD]], <16 x i8> [[VN]], <16 x i8> [[VM]], i32 3)
// CHECK-NEXT:    ret <4 x float> [[FDOT4_LANE]]
//
// CHECK-CXX-LABEL: define dso_local noundef <4 x float> @_Z20test_vdotq_laneq_f3213__Float32x4_t14__Mfloat8x16_tS0_m(
// CHECK-CXX-SAME: <4 x float> noundef [[VD:%.*]], <16 x i8> [[VN:%.*]], <16 x i8> [[VM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT:  [[ENTRY:.*:]]
// CHECK-CXX-NEXT:    call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
// CHECK-CXX-NEXT:    [[FDOT4_LANE:%.*]] = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v4f32.v16i8(<4 x float> [[VD]], <16 x i8> [[VN]], <16 x i8> [[VM]], i32 3)
// CHECK-CXX-NEXT:    ret <4 x float> [[FDOT4_LANE]]
//
float32x4_t test_vdotq_laneq_f32(float32x4_t vd, mfloat8x16_t vn, mfloat8x16_t vm, fpm_t fpmr) {
  return vdotq_laneq_f32_mf8_fpm(vd, vn, vm, 3, fpmr);
}