File size: 11,447 Bytes
8b7c501
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
// Auto-generated file. Do not edit!
//   Template: src/f16-gavgpool/unipass-f16c.c.in
//   Generator: tools/xngen
//
// Copyright 2022 Google LLC
//
// This source code is licensed under the BSD-style license found in the
// LICENSE file in the root directory of this source tree.

#include <assert.h>

#include <immintrin.h>

#include <xnnpack/gavgpool.h>
#include <xnnpack/intrinsics-polyfill.h>


void xnn_f16_gavgpool_minmax_ukernel_7x__f16c_c32(
    size_t rows,
    size_t channels,
    const void* input,
    size_t input_stride,
    const void* zero,
    void* output,
    const union xnn_f16_scaleminmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
  assert(rows != 0);
  assert(rows <= 7);
  assert(channels != 0);

  const uint16_t* i0 = input;
  const uint16_t* i1 = (const uint16_t*) ((uintptr_t) i0 + input_stride);
  if XNN_UNPREDICTABLE(rows < 2) {
    i1 = (const uint16_t*) zero;
  }
  const uint16_t* i2 = (const uint16_t*) ((uintptr_t) i1 + input_stride);
  if XNN_UNPREDICTABLE(rows <= 2) {
    i2 = (const uint16_t*) zero;
  }
  const uint16_t* i3 = (const uint16_t*) ((uintptr_t) i2 + input_stride);
  if XNN_UNPREDICTABLE(rows < 4) {
    i3 = (const uint16_t*) zero;
  }
  const uint16_t* i4 = (const uint16_t*) ((uintptr_t) i3 + input_stride);
  if XNN_UNPREDICTABLE(rows <= 4) {
    i4 = (const uint16_t*) zero;
  }
  const uint16_t* i5 = (const uint16_t*) ((uintptr_t) i4 + input_stride);
  if XNN_UNPREDICTABLE(rows < 6) {
    i5 = (const uint16_t*) zero;
  }
  const uint16_t* i6 = (const uint16_t*) ((uintptr_t) i5 + input_stride);
  if XNN_UNPREDICTABLE(rows <= 6) {
    i6 = (const uint16_t*) zero;
  }
  uint16_t* o = (uint16_t*) output;

  const __m256 vscale = _mm256_load_ps(params->avx.scale);
  const __m256 vmin = _mm256_load_ps(params->avx.min);
  const __m256 vmax = _mm256_load_ps(params->avx.max);
  for (; channels >= 32; channels -= 32) {
    const __m256 vi0x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
    const __m256 vi0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i0 + 8)));
    const __m256 vi0xGHIJKLMN = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i0 + 16)));
    const __m256 vi0xOPQRSTUV = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i0 + 24)));
    i0 += 32;
    const __m256 vi1x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
    const __m256 vi1x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i1 + 8)));
    const __m256 vi1xGHIJKLMN = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i1 + 16)));
    const __m256 vi1xOPQRSTUV = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i1 + 24)));
    i1 += 32;

    const __m256 vi2x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
    __m128i vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(vi0x01234567, vi1x01234567), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi2x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i2 + 8)));
    __m128i vacc89ABCDEF = _mm256_cvtps_ph(_mm256_add_ps(vi0x89ABCDEF, vi1x89ABCDEF), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi2xGHIJKLMN = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i2 + 16)));
    __m128i vaccGHIJKLMN = _mm256_cvtps_ph(_mm256_add_ps(vi0xGHIJKLMN, vi1xGHIJKLMN), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi2xOPQRSTUV = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i2 + 24)));
    __m128i vaccOPQRSTUV = _mm256_cvtps_ph(_mm256_add_ps(vi0xOPQRSTUV, vi1xOPQRSTUV), _MM_FROUND_TO_NEAREST_INT);
    i2 += 32;

    const __m256 vi3x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
    vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi2x01234567), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi3x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i3 + 8)));
    vacc89ABCDEF = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc89ABCDEF), vi2x89ABCDEF), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi3xGHIJKLMN = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i3 + 16)));
    vaccGHIJKLMN = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccGHIJKLMN), vi2xGHIJKLMN), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi3xOPQRSTUV = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i3 + 24)));
    i3 += 32;
    vaccOPQRSTUV = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccOPQRSTUV), vi2xOPQRSTUV), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi4x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
    vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi3x01234567), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi4x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i4 + 8)));
    vacc89ABCDEF = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc89ABCDEF), vi3x89ABCDEF), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi4xGHIJKLMN = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i4 + 16)));
    vaccGHIJKLMN = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccGHIJKLMN), vi3xGHIJKLMN), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi4xOPQRSTUV = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i4 + 24)));
    i4 += 32;
    vaccOPQRSTUV = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccOPQRSTUV), vi3xOPQRSTUV), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi5x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
    vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi4x01234567), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi5x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i5 + 8)));
    vacc89ABCDEF = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc89ABCDEF), vi4x89ABCDEF), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi5xGHIJKLMN = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i5 + 16)));
    vaccGHIJKLMN = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccGHIJKLMN), vi4xGHIJKLMN), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi5xOPQRSTUV = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i5 + 24)));
    i5 += 32;
    vaccOPQRSTUV = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccOPQRSTUV), vi4xOPQRSTUV), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi6x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
    vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi5x01234567), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi6x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i6 + 8)));
    vacc89ABCDEF = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc89ABCDEF), vi5x89ABCDEF), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi6xGHIJKLMN = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i6 + 16)));
    vaccGHIJKLMN = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccGHIJKLMN), vi5xGHIJKLMN), _MM_FROUND_TO_NEAREST_INT);
    const __m256 vi6xOPQRSTUV = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i6 + 24)));
    i6 += 32;
    vaccOPQRSTUV = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccOPQRSTUV), vi5xOPQRSTUV), _MM_FROUND_TO_NEAREST_INT);
    vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi6x01234567), _MM_FROUND_TO_NEAREST_INT);
    vacc89ABCDEF = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc89ABCDEF), vi6x89ABCDEF), _MM_FROUND_TO_NEAREST_INT);
    vaccGHIJKLMN = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccGHIJKLMN), vi6xGHIJKLMN), _MM_FROUND_TO_NEAREST_INT);
    vaccOPQRSTUV = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vaccOPQRSTUV), vi6xOPQRSTUV), _MM_FROUND_TO_NEAREST_INT);

    vacc01234567 = _mm256_cvtps_ph(_mm256_mul_ps(_mm256_cvtph_ps(vacc01234567), vscale), _MM_FROUND_TO_NEAREST_INT);
    vacc89ABCDEF = _mm256_cvtps_ph(_mm256_mul_ps(_mm256_cvtph_ps(vacc89ABCDEF), vscale), _MM_FROUND_TO_NEAREST_INT);
    vaccGHIJKLMN = _mm256_cvtps_ph(_mm256_mul_ps(_mm256_cvtph_ps(vaccGHIJKLMN), vscale), _MM_FROUND_TO_NEAREST_INT);
    vaccOPQRSTUV = _mm256_cvtps_ph(_mm256_mul_ps(_mm256_cvtph_ps(vaccOPQRSTUV), vscale), _MM_FROUND_TO_NEAREST_INT);

    __m256 vout01234567 = _mm256_max_ps(_mm256_cvtph_ps(vacc01234567), vmin);
    __m256 vout89ABCDEF = _mm256_max_ps(_mm256_cvtph_ps(vacc89ABCDEF), vmin);
    __m256 voutGHIJKLMN = _mm256_max_ps(_mm256_cvtph_ps(vaccGHIJKLMN), vmin);
    __m256 voutOPQRSTUV = _mm256_max_ps(_mm256_cvtph_ps(vaccOPQRSTUV), vmin);

    vout01234567 = _mm256_min_ps(vout01234567, vmax);
    vout89ABCDEF = _mm256_min_ps(vout89ABCDEF, vmax);
    voutGHIJKLMN = _mm256_min_ps(voutGHIJKLMN, vmax);
    voutOPQRSTUV = _mm256_min_ps(voutOPQRSTUV, vmax);

    _mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vout01234567, _MM_FROUND_TO_NEAREST_INT));
    _mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vout89ABCDEF, _MM_FROUND_TO_NEAREST_INT));
    _mm_storeu_si128((__m128i*) (o + 16), _mm256_cvtps_ph(voutGHIJKLMN, _MM_FROUND_TO_NEAREST_INT));
    _mm_storeu_si128((__m128i*) (o + 24), _mm256_cvtps_ph(voutOPQRSTUV, _MM_FROUND_TO_NEAREST_INT));
    o += 32;
  }
  if XNN_UNLIKELY(channels != 0) {
    do {
      const __m256 vi0x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
      i0 += 8;
      const __m256 vi1x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
      i1 += 8;

      const __m256 vi2x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
      __m128i vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(vi0x01234567, vi1x01234567), _MM_FROUND_TO_NEAREST_INT);
      i2 += 8;

      const __m256 vi3x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
      i3 += 8;
      vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi2x01234567), _MM_FROUND_TO_NEAREST_INT);
      const __m256 vi4x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
      i4 += 8;
      vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi3x01234567), _MM_FROUND_TO_NEAREST_INT);
      const __m256 vi5x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
      i5 += 8;
      vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi4x01234567), _MM_FROUND_TO_NEAREST_INT);
      const __m256 vi6x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
      i6 += 8;
      vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi5x01234567), _MM_FROUND_TO_NEAREST_INT);
      vacc01234567 = _mm256_cvtps_ph(_mm256_add_ps(_mm256_cvtph_ps(vacc01234567), vi6x01234567), _MM_FROUND_TO_NEAREST_INT);

      vacc01234567 = _mm256_cvtps_ph(_mm256_mul_ps(_mm256_cvtph_ps(vacc01234567), vscale), _MM_FROUND_TO_NEAREST_INT);
      __m256 vout01234567 = _mm256_max_ps(_mm256_cvtph_ps(vacc01234567), vmin);
      vout01234567 = _mm256_min_ps(vout01234567, vmax);

      if XNN_LIKELY(channels >= 8) {
        _mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vout01234567, _MM_FROUND_TO_NEAREST_INT));
        o += 8;
        channels -= 8;
      } else {
        __m128i vh01234567 = _mm256_cvtps_ph(vout01234567, _MM_FROUND_TO_NEAREST_INT);
        if (channels & 4) {
          _mm_storel_epi64((__m128i*) o, vh01234567);
          o += 4;
          vh01234567 = _mm_unpackhi_epi64(vh01234567, vh01234567);
        }
        if (channels & 2) {
          _mm_storeu_si32(o, vh01234567);
          o += 2;
          vh01234567 = _mm_srli_epi64(vh01234567, 32);
        }
        if (channels & 1) {
          *o = (uint16_t) _mm_extract_epi16(vh01234567, 0);
        }
        channels = 0;
      }
    } while (channels != 0);
  }
}