/*

 * 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

176k

    int shift) {

31

176k

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

32

176k

  const __m128i s0 = xx_loadl_64(src0);

33

176k

  const __m128i s1 = xx_loadl_64(src1);

34

176k

  const __m128i s0_s1 = _mm_unpacklo_epi16(s0, s1);

35

176k

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

36

176k

  const __m128i res_a = _mm_madd_epi16(s0_s1, m_max_minus_m);

37

176k

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

38

176k

  const __m128i res_d = _mm_srai_epi32(res_c, shift);

39

176k

  const __m128i res_e = _mm_packs_epi32(res_d, res_d);

40

176k

  const __m128i res = _mm_packus_epi16(res_e, res_e);

41

42

176k

  xx_storel_32(dst, res);

43

176k

}

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

455k

    int shift) {

49

455k

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

50

455k

  const __m128i s0 = xx_loadu_128(src0);

51

455k

  const __m128i s1 = xx_loadu_128(src1);

52

455k

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

53

455k

                                  _mm_unpacklo_epi16(*m, max_minus_m));

54

455k

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

55

455k

                                  _mm_unpackhi_epi16(*m, max_minus_m));

56

455k

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

57

455k

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

58

455k

  const __m128i res_e = _mm_packs_epi32(res_lo, res_hi);

59

455k

  const __m128i res = _mm_packus_epi16(res_e, res_e);

60

61

455k

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

62

455k

}

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

15.4k

    const __m128i *round_offset, int shift) {

88

15.4k

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

89

194k

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

90

179k

    const __m128i m0 = xx_loadl_64(mask);

91

179k

    const __m128i m = _mm_cvtepu8_epi16(m0);

92

179k

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

93

179k

                                shift);

94

179k

    mask += mask_stride;

95

179k

    dst += dst_stride;

96

179k

    src0 += src0_stride;

97

179k

    src1 += src1_stride;

98

179k

  }

99

15.4k

}

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

30.3k

    const __m128i *round_offset, int shift) {

106

30.3k

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

107

30.3k

  const __m128i one_b = _mm_set1_epi8(1);

108

30.3k

  const __m128i two_w = _mm_set1_epi16(2);

109

206k

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

110

175k

    const __m128i m_i0 = xx_loadl_64(mask);

111

175k

    const __m128i m_i1 = xx_loadl_64(mask + mask_stride);

112

175k

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

113

175k

    const __m128i m_acbd = _mm_maddubs_epi16(m_ac, one_b);

114

175k

    const __m128i m_acbd_2 = _mm_add_epi16(m_acbd, two_w);

115

175k

    const __m128i m = _mm_srli_epi16(m_acbd_2, 2);

116

117

175k

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

118

175k

                                shift);

119

175k

    mask += mask_stride << 1;

120

175k

    dst += dst_stride;

121

175k

    src0 += src0_stride;

122

175k

    src1 += src1_stride;

123

175k

  }

124

30.3k

}

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

25.4k

    const __m128i *round_offset, int shift) {

131

25.4k

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

132

25.4k

  const __m128i one_b = _mm_set1_epi8(1);

133

25.4k

  const __m128i two_w = _mm_set1_epi16(2);

134

301k

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

135

275k

    const __m128i m_i0 = xx_loadu_128(mask);

136

275k

    const __m128i m_i1 = xx_loadu_128(mask + mask_stride);

137

275k

    const __m128i m_ac = _mm_adds_epu8(m_i0, m_i1);

138

275k

    const __m128i m_acbd = _mm_maddubs_epi16(m_ac, one_b);

139

275k

    const __m128i m_acbd_2 = _mm_add_epi16(m_acbd, two_w);

140

275k

    const __m128i m = _mm_srli_epi16(m_acbd_2, 2);

141

142

275k

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

143

275k

                                shift);

144

275k

    mask += mask_stride << 1;

145

275k

    dst += dst_stride;

146

275k

    src0 += src0_stride;

147

275k

    src1 += src1_stride;

148

275k

  }

149

25.4k

}

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

52

    const __m128i *round_offset, int shift) {

156

52

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

157

52

  const __m128i one_b = _mm_set1_epi8(1);

158

52

  const __m128i zeros = _mm_setzero_si128();

159

468

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

160

416

    const __m128i m_i0 = xx_loadl_64(mask);

161

416

    const __m128i m_ac = _mm_maddubs_epi16(m_i0, one_b);

162

416

    const __m128i m = _mm_avg_epu16(m_ac, zeros);

163

164

416

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

165

416

                                shift);

166

416

    mask += mask_stride;

167

416

    dst += dst_stride;

168

416

    src0 += src0_stride;

169

416

    src1 += src1_stride;

170

416

  }

171

52

}

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

62

    const __m128i *round_offset, int shift) {

178

62

  const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA);

179

62

  const __m128i one_b = _mm_set1_epi8(1);

180

62

  const __m128i zeros = _mm_setzero_si128();

181

766

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

182

704

    const __m128i m_i0 = xx_loadu_128(mask);

183

704

    const __m128i m_ac = _mm_maddubs_epi16(m_i0, one_b);

184

704

    const __m128i m = _mm_avg_epu16(m_ac, zeros);

185

186

704

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

187

704

                                shift);

188

704

    mask += mask_stride;

189

704

    dst += dst_stride;

190

704

    src0 += src0_stride;

191

704

    src1 += src1_stride;

192

704

  }

193

62

}

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_