Hugging Face's logo

Androidonnxfork
/
test

Model card Files Community
5
test
File size: 6,366 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
 
// Auto-generated file. Do not edit!
//   Template: src/f16-vmulcaddc/fma3.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/intrinsics-polyfill.h>
#include <xnnpack/math.h>
#include <xnnpack/vmulcaddc.h>


void xnn_f16_vmulcaddc_minmax_ukernel_c16__fma3_2x(
    size_t rows,
    size_t channels,
    const void* restrict input,
    size_t input_stride,
    const void* restrict weights,
    void* restrict output,
    size_t output_stride,
    const union xnn_f16_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
  assert(rows != 0);
  assert(channels != 0);
  assert(channels % sizeof(uint16_t) == 0);

  const uint16_t* i0 = (const uint16_t*) input;
  uint16_t* o0 = (uint16_t*) output;
  const uint16_t* i1 = (const uint16_t*) ((uintptr_t) i0 + input_stride);
  uint16_t* o1 = (uint16_t*) ((uintptr_t) o0 + output_stride);

  const size_t input_increment = input_stride * 2 - channels;
  const size_t output_increment = output_stride * 2 - channels;

  const __m256 vmin = _mm256_load_ps(params->avx.min);
  const __m256 vmax = _mm256_load_ps(params->avx.max);
  do {
    if XNN_UNPREDICTABLE(rows < 2) {
      i1 = i0;
      o1 = o0;
    }

    const uint16_t* w = (const uint16_t*) weights;
    size_t c = channels;
    for (; c >= 16 * sizeof(uint16_t); c -= 16 * sizeof(uint16_t)) {
      const __m256 vscale01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) w));
      const __m256 vscale89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 8)));

      __m256 vacc0x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
      __m256 vacc0x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i0 + 8)));
      i0 += 16;
      __m256 vacc1x01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
      __m256 vacc1x89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i1 + 8)));
      i1 += 16;

      const __m256 vbias01234567 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 16)));
      const __m256 vbias89ABCDEF = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 24)));
      w += 32;

      vacc0x01234567 = _mm256_fmadd_ps(vacc0x01234567, vscale01234567, vbias01234567);
      vacc0x89ABCDEF = _mm256_fmadd_ps(vacc0x89ABCDEF, vscale89ABCDEF, vbias89ABCDEF);
      vacc1x01234567 = _mm256_fmadd_ps(vacc1x01234567, vscale01234567, vbias01234567);
      vacc1x89ABCDEF = _mm256_fmadd_ps(vacc1x89ABCDEF, vscale89ABCDEF, vbias89ABCDEF);

      vacc0x01234567 = _mm256_max_ps(vacc0x01234567, vmin);
      vacc0x89ABCDEF = _mm256_max_ps(vacc0x89ABCDEF, vmin);
      vacc1x01234567 = _mm256_max_ps(vacc1x01234567, vmin);
      vacc1x89ABCDEF = _mm256_max_ps(vacc1x89ABCDEF, vmin);

      vacc0x01234567 = _mm256_min_ps(vacc0x01234567, vmax);
      vacc0x89ABCDEF = _mm256_min_ps(vacc0x89ABCDEF, vmax);
      vacc1x01234567 = _mm256_min_ps(vacc1x01234567, vmax);
      vacc1x89ABCDEF = _mm256_min_ps(vacc1x89ABCDEF, vmax);

      _mm_storeu_si128((__m128i*) o0, _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_TO_NEAREST_INT));
      _mm_storeu_si128((__m128i*) (o0 + 8), _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_TO_NEAREST_INT));
      o0 += 16;
      _mm_storeu_si128((__m128i*) o1, _mm256_cvtps_ph(vacc1x01234567, _MM_FROUND_TO_NEAREST_INT));
      _mm_storeu_si128((__m128i*) (o1 + 8), _mm256_cvtps_ph(vacc1x89ABCDEF, _MM_FROUND_TO_NEAREST_INT));
      o1 += 16;
    }
    for (; c >= 8 * sizeof(uint16_t); c -= 8 * sizeof(uint16_t)) {
      const __m256 vscale = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) w));

      __m256 vacc0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
      i0 += 8;
      __m256 vacc1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
      i1 += 8;

      const __m256 vbias = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 16)));
      w += 8;

      vacc0 = _mm256_fmadd_ps(vacc0, vscale, vbias);
      vacc1 = _mm256_fmadd_ps(vacc1, vscale, vbias);

      vacc0 = _mm256_max_ps(vacc0, vmin);
      vacc1 = _mm256_max_ps(vacc1, vmin);

      vacc0 = _mm256_min_ps(vacc0, vmax);
      vacc1 = _mm256_min_ps(vacc1, vmax);

      _mm_storeu_si128((__m128i*) o0, _mm256_cvtps_ph(vacc0, _MM_FROUND_TO_NEAREST_INT));
      o0 += 8;
      _mm_storeu_si128((__m128i*) o1, _mm256_cvtps_ph(vacc1, _MM_FROUND_TO_NEAREST_INT));
      o1 += 8;
    }
    if XNN_UNLIKELY(c != 0) {
      const __m256 vscale = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) w));

      __m256 vacc0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
      i0 = (const uint16_t*) ((uintptr_t) i0 + c);
      __m256 vacc1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
      i1 = (const uint16_t*) ((uintptr_t) i1 + c);

      const __m256 vbias = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (w + 16)));

      vacc0 = _mm256_fmadd_ps(vacc0, vscale, vbias);
      vacc1 = _mm256_fmadd_ps(vacc1, vscale, vbias);

      vacc0 = _mm256_max_ps(vacc0, vmin);
      vacc1 = _mm256_max_ps(vacc1, vmin);

      vacc0 = _mm256_min_ps(vacc0, vmax);
      vacc1 = _mm256_min_ps(vacc1, vmax);

      __m128i vh0 = _mm256_cvtps_ph(vacc0, _MM_FROUND_TO_NEAREST_INT);
      __m128i vh1 = _mm256_cvtps_ph(vacc1, _MM_FROUND_TO_NEAREST_INT);

      if (c & (4 * sizeof(uint16_t))) {
        _mm_storel_epi64((__m128i*) o0, vh0);
        _mm_storel_epi64((__m128i*) o1, vh1);

        vh0 = _mm_unpackhi_epi64(vh0, vh0);
        vh1 = _mm_unpackhi_epi64(vh1, vh1);

        o0 += 4;
        o1 += 4;
      }
      if (c & (2 * sizeof(uint16_t))) {
        _mm_storeu_si32(o0, vh0);
        _mm_storeu_si32(o1, vh1);

        vh0 = _mm_srli_epi64(vh0, 32);
        vh1 = _mm_srli_epi64(vh1, 32);

        o0 += 2;
        o1 += 2;
      }
      if (c & (1 * sizeof(uint16_t))) {
        *o0 = (uint16_t) _mm_extract_epi16(vh0, 0);
        *o1 = (uint16_t) _mm_extract_epi16(vh1, 0);

        o0 += 1;
        o1 += 1;
      }
    }
    i0 = (const uint16_t*) ((uintptr_t) i0 + input_increment);
    o0 = (uint16_t*) ((uintptr_t) o0 + output_increment);
    i1 = (const uint16_t*) ((uintptr_t) i1 + input_increment);
    o1 = (uint16_t*) ((uintptr_t) o1 + output_increment);
    rows = doz(rows, 2);
  } while (rows != 0);
}