/*

 * Copyright (c) 2018, Alliance for Open Media. All rights reserved.

 *

 * This source code is subject to the terms of the BSD 2 Clause License and

 * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License

 * was not distributed with this source code in the LICENSE file, you can

 * obtain it at www.aomedia.org/license/software. If the Alliance for Open

 * Media Patent License 1.0 was not distributed with this source code in the

 * PATENTS file, you can obtain it at www.aomedia.org/license/patent.

 */

#ifndef AOM_AOM_DSP_X86_BLEND_MASK_SSE4_H_

#define AOM_AOM_DSP_X86_BLEND_MASK_SSE4_H_

#include <smmintrin.h>  // SSE4.1

#include <assert.h>

#include "aom/aom_integer.h"

#include "aom_ports/mem.h"

#include "aom_dsp/aom_dsp_common.h"

#include "aom_dsp/blend.h"

#include "aom_dsp/x86/synonyms.h"

#include "config/aom_dsp_rtcd.h"

static inline void blend_a64_d16_mask_w4_sse41(

    uint8_t *dst, const CONV_BUF_TYPE *src0, const CONV_BUF_TYPE *src1,

    const __m128i *m, const __m128i *v_round_offset, const __m128i *v_maxval,

    int shift) {

  const __m128i max_minus_m = _mm_sub_epi16(*v_maxval, *m);

  const __m128i s0 = xx_loadl_64(src0);

  const __m128i s1 = xx_loadl_64(src1);

  const __m128i s0_s1 = _mm_unpacklo_epi16(s0, s1);

  const __m128i m_max_minus_m = _mm_unpacklo_epi16(*m, max_minus_m);

  const __m128i res_a = _mm_madd_epi16(s0_s1, m_max_minus_m);

  const __m128i res_c = _mm_sub_epi32(res_a, *v_round_offset);

  const __m128i res_d = _mm_srai_epi32(res_c, shift);

  const __m128i res_e = _mm_packs_epi32(res_d, res_d);

  const __m128i res = _mm_packus_epi16(res_e, res_e);

  xx_storel_32(dst, res);

}

Unexecuted instantiation: blend_a64_mask_sse4.c:blend_a64_d16_mask_w4_sse41

blend_a64_mask_avx2.c:blend_a64_d16_mask_w4_sse41

Line

Count

Source

30

158k

    int shift) {

31

158k

  const __m128i max_minus_m = _mm_sub_epi16(*v_maxval, *m);

32

158k

  const __m128i s0 = xx_loadl_64(src0);

33

158k

  const __m128i s1 = xx_loadl_64(src1);

34

158k

  const __m128i s0_s1 = _mm_unpacklo_epi16(s0, s1);

35

158k

  const __m128i m_max_minus_m = _mm_unpacklo_epi16(*m, max_minus_m);

36

158k

  const __m128i res_a = _mm_madd_epi16(s0_s1, m_max_minus_m);

37

158k

  const __m128i res_c = _mm_sub_epi32(res_a, *v_round_offset);

38

158k

  const __m128i res_d = _mm_srai_epi32(res_c, shift);

39

158k

  const __m128i res_e = _mm_packs_epi32(res_d, res_d);

40

158k

  const __m128i res = _mm_packus_epi16(res_e, res_e);

41

42

158k

  xx_storel_32(dst, res);

43

158k

}

static inline void blend_a64_d16_mask_w8_sse41(

    uint8_t *dst, const CONV_BUF_TYPE *src0, const CONV_BUF_TYPE *src1,

    const __m128i *m, const __m128i *v_round_offset, const __m128i *v_maxval,

    int shift) {

  const __m128i max_minus_m = _mm_sub_epi16(*v_maxval, *m);

  const __m128i s0 = xx_loadu_128(src0);

  const __m128i s1 = xx_loadu_128(src1);

  __m128i res_lo = _mm_madd_epi16(_mm_unpacklo_epi16(s0, s1),

                                  _mm_unpacklo_epi16(*m, max_minus_m));

  __m128i res_hi = _mm_madd_epi16(_mm_unpackhi_epi16(s0, s1),

                                  _mm_unpackhi_epi16(*m, max_minus_m));

  res_lo = _mm_srai_epi32(_mm_sub_epi32(res_lo, *v_round_offset), shift);

  res_hi = _mm_srai_epi32(_mm_sub_epi32(res_hi, *v_round_offset), shift);

  const __m128i res_e = _mm_packs_epi32(res_lo, res_hi);

  const __m128i res = _mm_packus_epi16(res_e, res_e);

  _mm_storel_epi64((__m128i *)(dst), res);

}

Unexecuted instantiation: blend_a64_mask_sse4.c:blend_a64_d16_mask_w8_sse41

blend_a64_mask_avx2.c:blend_a64_d16_mask_w8_sse41

Line

Count

Source

48

432k

    int shift) {

49

432k

  const __m128i max_minus_m = _mm_sub_epi16(*v_maxval, *m);

50

432k

  const __m128i s0 = xx_loadu_128(src0);

51

432k

  const __m128i s1 = xx_loadu_128(src1);

52

432k

  __m128i res_lo = _mm_madd_epi16(_mm_unpacklo_epi16(s0, s1),

53

432k

                                  _mm_unpacklo_epi16(*m, max_minus_m));

54

432k

  __m128i res_hi = _mm_madd_epi16(_mm_unpackhi_epi16(s0, s1),

55

432k

                                  _mm_unpackhi_epi16(*m, max_minus_m));

56

432k

  res_lo = _mm_srai_epi32(_mm_sub_epi32(res_lo, *v_round_offset), shift);

57

432k

  res_hi = _mm_srai_epi32(_mm_sub_epi32(res_hi, *v_round_offset), shift);

58

432k

  const __m128i res_e = _mm_packs_epi32(res_lo, res_hi);

59

432k

  const __m128i res = _mm_packus_epi16(res_e, res_e);

60

61

432k

  _mm_storel_epi64((__m128i *)(dst), res);

62

432k

}

static inline void aom_lowbd_blend_a64_d16_mask_subw0_subh0_w4_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  for (int i = 0; i < h; ++i) {

    const __m128i m0 = xx_loadl_32(mask);

    const __m128i m = _mm_cvtepu8_epi16(m0);

    blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw0_subh0_w4_sse4_1

Unexecuted instantiation: blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw0_subh0_w4_sse4_1

static inline void aom_lowbd_blend_a64_d16_mask_subw0_subh0_w8_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  for (int i = 0; i < h; ++i) {

    const __m128i m0 = xx_loadl_64(mask);

    const __m128i m = _mm_cvtepu8_epi16(m0);

    blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw0_subh0_w8_sse4_1

blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw0_subh0_w8_sse4_1

Line

Count

Source

87

13.5k

    const __m128i *round_offset, int shift) {

88

13.5k

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

89

174k

  for (int i = 0; i < h; ++i) {

90

161k

    const __m128i m0 = xx_loadl_64(mask);

91

161k

    const __m128i m = _mm_cvtepu8_epi16(m0);

92

161k

    blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

93

161k

                                shift);

94

161k

    mask += mask_stride;

95

161k

    dst += dst_stride;

96

161k

    src0 += src0_stride;

97

161k

    src1 += src1_stride;

98

161k

  }

99

13.5k

}

static inline void aom_lowbd_blend_a64_d16_mask_subw1_subh1_w4_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  const __m128i one_b = _mm_set1_epi8(1);

  const __m128i two_w = _mm_set1_epi16(2);

  for (int i = 0; i < h; ++i) {

    const __m128i m_i0 = xx_loadl_64(mask);

    const __m128i m_i1 = xx_loadl_64(mask + mask_stride);

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

    const __m128i m_acbd = _mm_maddubs_epi16(m_ac, one_b);

    const __m128i m_acbd_2 = _mm_add_epi16(m_acbd, two_w);

    const __m128i m = _mm_srli_epi16(m_acbd_2, 2);

    blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride << 1;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw1_subh1_w4_sse4_1

blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw1_subh1_w4_sse4_1

Line

Count

Source

105

26.6k

    const __m128i *round_offset, int shift) {

106

26.6k

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

107

26.6k

  const __m128i one_b = _mm_set1_epi8(1);

108

26.6k

  const __m128i two_w = _mm_set1_epi16(2);

109

184k

  for (int i = 0; i < h; ++i) {

110

158k

    const __m128i m_i0 = xx_loadl_64(mask);

111

158k

    const __m128i m_i1 = xx_loadl_64(mask + mask_stride);

112

158k

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

113

158k

    const __m128i m_acbd = _mm_maddubs_epi16(m_ac, one_b);

114

158k

    const __m128i m_acbd_2 = _mm_add_epi16(m_acbd, two_w);

115

158k

    const __m128i m = _mm_srli_epi16(m_acbd_2, 2);

116

117

158k

    blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

118

158k

                                shift);

119

158k

    mask += mask_stride << 1;

120

158k

    dst += dst_stride;

121

158k

    src0 += src0_stride;

122

158k

    src1 += src1_stride;

123

158k

  }

124

26.6k

}

static inline void aom_lowbd_blend_a64_d16_mask_subw1_subh1_w8_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  const __m128i one_b = _mm_set1_epi8(1);

  const __m128i two_w = _mm_set1_epi16(2);

  for (int i = 0; i < h; ++i) {

    const __m128i m_i0 = xx_loadu_128(mask);

    const __m128i m_i1 = xx_loadu_128(mask + mask_stride);

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

    const __m128i m_acbd = _mm_maddubs_epi16(m_ac, one_b);

    const __m128i m_acbd_2 = _mm_add_epi16(m_acbd, two_w);

    const __m128i m = _mm_srli_epi16(m_acbd_2, 2);

    blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride << 1;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw1_subh1_w8_sse4_1

blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw1_subh1_w8_sse4_1

Line

Count

Source

130

24.4k

    const __m128i *round_offset, int shift) {

131

24.4k

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

132

24.4k

  const __m128i one_b = _mm_set1_epi8(1);

133

24.4k

  const __m128i two_w = _mm_set1_epi16(2);

134

295k

  for (int i = 0; i < h; ++i) {

135

270k

    const __m128i m_i0 = xx_loadu_128(mask);

136

270k

    const __m128i m_i1 = xx_loadu_128(mask + mask_stride);

137

270k

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

138

270k

    const __m128i m_acbd = _mm_maddubs_epi16(m_ac, one_b);

139

270k

    const __m128i m_acbd_2 = _mm_add_epi16(m_acbd, two_w);

140

270k

    const __m128i m = _mm_srli_epi16(m_acbd_2, 2);

141

142

270k

    blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

143

270k

                                shift);

144

270k

    mask += mask_stride << 1;

145

270k

    dst += dst_stride;

146

270k

    src0 += src0_stride;

147

270k

    src1 += src1_stride;

148

270k

  }

149

24.4k

}

static inline void aom_lowbd_blend_a64_d16_mask_subw1_subh0_w4_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  const __m128i one_b = _mm_set1_epi8(1);

  const __m128i zeros = _mm_setzero_si128();

  for (int i = 0; i < h; ++i) {

    const __m128i m_i0 = xx_loadl_64(mask);

    const __m128i m_ac = _mm_maddubs_epi16(m_i0, one_b);

    const __m128i m = _mm_avg_epu16(m_ac, zeros);

    blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw1_subh0_w4_sse4_1

blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw1_subh0_w4_sse4_1

Line

Count

Source

155

50

    const __m128i *round_offset, int shift) {

156

50

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

157

50

  const __m128i one_b = _mm_set1_epi8(1);

158

50

  const __m128i zeros = _mm_setzero_si128();

159

450

  for (int i = 0; i < h; ++i) {

160

400

    const __m128i m_i0 = xx_loadl_64(mask);

161

400

    const __m128i m_ac = _mm_maddubs_epi16(m_i0, one_b);

162

400

    const __m128i m = _mm_avg_epu16(m_ac, zeros);

163

164

400

    blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

165

400

                                shift);

166

400

    mask += mask_stride;

167

400

    dst += dst_stride;

168

400

    src0 += src0_stride;

169

400

    src1 += src1_stride;

170

400

  }

171

50

}

static inline void aom_lowbd_blend_a64_d16_mask_subw1_subh0_w8_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  const __m128i one_b = _mm_set1_epi8(1);

  const __m128i zeros = _mm_setzero_si128();

  for (int i = 0; i < h; ++i) {

    const __m128i m_i0 = xx_loadu_128(mask);

    const __m128i m_ac = _mm_maddubs_epi16(m_i0, one_b);

    const __m128i m = _mm_avg_epu16(m_ac, zeros);

    blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw1_subh0_w8_sse4_1

blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw1_subh0_w8_sse4_1

Line

Count

Source

177

60

    const __m128i *round_offset, int shift) {

178

60

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

179

60

  const __m128i one_b = _mm_set1_epi8(1);

180

60

  const __m128i zeros = _mm_setzero_si128();

181

732

  for (int i = 0; i < h; ++i) {

182

672

    const __m128i m_i0 = xx_loadu_128(mask);

183

672

    const __m128i m_ac = _mm_maddubs_epi16(m_i0, one_b);

184

672

    const __m128i m = _mm_avg_epu16(m_ac, zeros);

185

186

672

    blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

187

672

                                shift);

188

672

    mask += mask_stride;

189

672

    dst += dst_stride;

190

672

    src0 += src0_stride;

191

672

    src1 += src1_stride;

192

672

  }

193

60

}

static inline void aom_lowbd_blend_a64_d16_mask_subw0_subh1_w4_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  const __m128i zeros = _mm_setzero_si128();

  for (int i = 0; i < h; ++i) {

    const __m128i m_i0 = xx_loadl_64(mask);

    const __m128i m_i1 = xx_loadl_64(mask + mask_stride);

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

    const __m128i m = _mm_cvtepu8_epi16(_mm_avg_epu8(m_ac, zeros));

    blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride << 1;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw0_subh1_w4_sse4_1

Unexecuted instantiation: blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw0_subh1_w4_sse4_1

static inline void aom_lowbd_blend_a64_d16_mask_subw0_subh1_w8_sse4_1(

    uint8_t *dst, uint32_t dst_stride, const CONV_BUF_TYPE *src0,

    uint32_t src0_stride, const CONV_BUF_TYPE *src1, uint32_t src1_stride,

    const uint8_t *mask, uint32_t mask_stride, int h,

    const __m128i *round_offset, int shift) {

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

  const __m128i zeros = _mm_setzero_si128();

  for (int i = 0; i < h; ++i) {

    const __m128i m_i0 = xx_loadl_64(mask);

    const __m128i m_i1 = xx_loadl_64(mask + mask_stride);

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

    const __m128i m = _mm_cvtepu8_epi16(_mm_avg_epu8(m_ac, zeros));

    blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval,

                                shift);

    mask += mask_stride << 1;

    dst += dst_stride;

    src0 += src0_stride;

    src1 += src1_stride;

  }

}

Unexecuted instantiation: blend_a64_mask_sse4.c:aom_lowbd_blend_a64_d16_mask_subw0_subh1_w8_sse4_1

Unexecuted instantiation: blend_a64_mask_avx2.c:aom_lowbd_blend_a64_d16_mask_subw0_subh1_w8_sse4_1

#endif  // AOM_AOM_DSP_X86_BLEND_MASK_SSE4_H_