Hugging Face's logo

Androidonnxfork
/
test

Model card Files Community
5
test
File size: 21,998 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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
 
// Auto-generated file. Do not edit!
//   Template: src/f16-dwconv2d-chw/5x5s2p2-neonfp16arith.c.in
//   Generator: tools/xngen
//
// Copyright 2020 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 <arm_neon.h>

#include <xnnpack/dwconv.h>
#include <xnnpack/intrinsics-polyfill.h>
#include <xnnpack/math.h>


void xnn_f16_dwconv2d_chw_ukernel_5x5s2p2__neonfp16arith_1x8_acc3(
    size_t input_height,
    size_t input_width,
    const void* input,
    const void* weights,
    const void* zero,
    void* output,
    uint32_t padding_top,
    const union xnn_f16_chw_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
  assert(input_height != 0);
  assert(input_width != 0);
  assert(input_width % sizeof(uint16_t) == 0);
  assert(padding_top >= 1);
  assert(padding_top <= 2);

  #if XNN_ARCH_ARM64
    const uint16x8x2_t vminmax = vld2q_dup_u16(&params->neonfp16arith_stride2.min);
    const float16x8_t vmin = vreinterpretq_f16_u16(vminmax.val[0]);
    const float16x8_t vmax = vreinterpretq_f16_u16(vminmax.val[1]);
  #else
    // vld2_dup is to work around aarch32 clang bug with vld1q_dup
    const uint16x4x2_t vminmax = vld2_dup_u16(&params->neonfp16arith_stride2.min);
    const float16x8_t vmin = vreinterpretq_f16_u16(vcombine_u16(vminmax.val[0], vminmax.val[0]));
    const float16x8_t vmax = vreinterpretq_f16_u16(vcombine_u16(vminmax.val[1], vminmax.val[1]));
  #endif
  const uint16x8_t vmask_even = vld1q_u16(params->neonfp16arith_stride2.mask_even);
  const uint16x8_t vmask_odd = vld1q_u16(params->neonfp16arith_stride2.mask_odd);

  const uint16_t* w = (const uint16_t*) weights;
  const float16x8_t vw01234567 = vreinterpretq_f16_u16(vld1q_u16(w));
  const float16x8_t vw89ABCDEF = vreinterpretq_f16_u16(vld1q_u16(w + 8));
  const float16x8_t vwGHIJKLMN = vreinterpretq_f16_u16(vld1q_u16(w + 16));
  const float16x4_t vwOP = vreinterpret_f16_u32(vld1_dup_u32((const void*) (w + 24)));

  const uint32_t padding_top_less_1 = padding_top - 1;
  const size_t input_decrement = round_up_po2(input_width, 16 * sizeof(uint16_t));

  const uint16_t* i0 = zero;
  const uint16_t* i1 = (const uint16_t*) ((uintptr_t) input - ((-padding_top_less_1) & input_width));
  const uint16_t* i2 = (const uint16_t*) ((uintptr_t) i1 + input_width);
  if XNN_UNPREDICTABLE(padding_top_less_1 != 0) {
    i1 = zero;
  }
  const uint16_t* i3 = (const uint16_t*) ((uintptr_t) i2 + input_width);
  const uint16_t* i4 = (const uint16_t*) ((uintptr_t) i3 + input_width);


  uint16_t* o0 = output;

  size_t padded_input_height = input_height + (padding_top_less_1 + 1) + 2 /* padding bottom */;
  size_t output_height = (padded_input_height - 5 /* kernel size */ + 2 /* subsampling */) / 2;
  do {
    if XNN_UNPREDICTABLE(padded_input_height < 6) {
      i3 = zero;
    }
    if XNN_UNPREDICTABLE(padded_input_height < 7) {
      i4 = zero;
    }

    float16x8_t vi0x02468ACE = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi1x02468ACE = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi2x02468ACE = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi3x02468ACE = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi4x02468ACE = vreinterpretq_f16_u16(vmovq_n_u16(0));

    float16x8_t vi0x13579BDF = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi1x13579BDF = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi2x13579BDF = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi3x13579BDF = vreinterpretq_f16_u16(vmovq_n_u16(0));
    float16x8_t vi4x13579BDF = vreinterpretq_f16_u16(vmovq_n_u16(0));

    uint16x8x2_t vi0xGIKMOQSUHJLNPRTV = vld2q_u16(i0); i0 += 16;
    uint16x8x2_t vi1xGIKMOQSUHJLNPRTV = vld2q_u16(i1); i1 += 16;
    uint16x8x2_t vi2xGIKMOQSUHJLNPRTV = vld2q_u16(i2); i2 += 16;
    uint16x8x2_t vi3xGIKMOQSUHJLNPRTV = vld2q_u16(i3); i3 += 16;
    uint16x8x2_t vi4xGIKMOQSUHJLNPRTV = vld2q_u16(i4); i4 += 16;

    size_t w = input_width;
    for (; w > 16 * sizeof(uint16_t); w -= 16 * sizeof(uint16_t)) {
      float16x8_t vo0p0 = vdupq_lane_f16(vget_low_f16(vw01234567), 0);

      // Center column
      float16x8_t vo0p1 = vmulq_lane_f16(vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[0]), vget_low_f16(vw01234567), 3);
      float16x8_t vo0p2 = vmulq_lane_f16(vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[0]), vget_low_f16(vw89ABCDEF), 0);
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[0]), vw89ABCDEF, 5);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[0]), vget_high_f16(vw89ABCDEF), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[0]), vwGHIJKLMN, 2);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[0]), vget_low_f16(vwGHIJKLMN), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[0]), vwGHIJKLMN, 7);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[0]), vget_high_f16(vwGHIJKLMN), 3);
      #endif
      // Right by 2 column
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[1]), vw01234567, 4);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[1]), vget_high_f16(vw01234567), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[1]), vw89ABCDEF, 1);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[1]), vget_low_f16(vw89ABCDEF), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[1]), vw89ABCDEF, 6);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[1]), vget_high_f16(vw89ABCDEF), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[1]), vwGHIJKLMN, 3);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[1]), vget_low_f16(vwGHIJKLMN), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_lane_f16(vo0p2, vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[1]), vwOP, 0);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[1]), vwOP, 0);
      #endif
      // Left by 2 column
      const float16x8_t vi0xEGIKMOQS = vextq_f16(vi0x02468ACE, vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[0]), 7);
      vi0x02468ACE = vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[0]);
      const float16x8_t vi1xEGIKMOQS = vextq_f16(vi1x02468ACE, vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[0]), 7);
      vi1x02468ACE = vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[0]);
      const float16x8_t vi2xEGIKMOQS = vextq_f16(vi2x02468ACE, vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[0]), 7);
      vi2x02468ACE = vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[0]);
      const float16x8_t vi3xEGIKMOQS = vextq_f16(vi3x02468ACE, vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[0]), 7);
      vi3x02468ACE = vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[0]);
      const float16x8_t vi4xEGIKMOQS = vextq_f16(vi4x02468ACE, vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[0]), 7);
      vi4x02468ACE = vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[0]);

      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi0xEGIKMOQS, vw01234567, 1);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi0xEGIKMOQS, vget_low_f16(vw01234567), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi1xEGIKMOQS, vw01234567, 6);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi1xEGIKMOQS, vget_high_f16(vw01234567), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi2xEGIKMOQS, vw89ABCDEF, 3);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi2xEGIKMOQS, vget_low_f16(vw89ABCDEF), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi3xEGIKMOQS, vwGHIJKLMN, 0);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi3xEGIKMOQS, vget_low_f16(vwGHIJKLMN), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi4xEGIKMOQS, vwGHIJKLMN, 5);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi4xEGIKMOQS, vget_high_f16(vwGHIJKLMN), 1);
      #endif
      // Left by 1 column, s1
      const float16x8_t vi0xFHJLNPRT = vextq_f16(vi0x13579BDF, vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[1]), 7);
      vi0x13579BDF = vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[1]);
      const float16x8_t vi1xFHJLNPRT = vextq_f16(vi1x13579BDF, vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[1]), 7);
      vi1x13579BDF = vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[1]);
      const float16x8_t vi2xFHJLNPRT = vextq_f16(vi2x13579BDF, vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[1]), 7);
      vi2x13579BDF = vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[1]);
      const float16x8_t vi3xFHJLNPRT = vextq_f16(vi3x13579BDF, vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[1]), 7);
      vi3x13579BDF = vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[1]);
      const float16x8_t vi4xFHJLNPRT = vextq_f16(vi4x13579BDF, vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[1]), 7);
      vi4x13579BDF = vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[1]);

      const uint16x8x2_t vi0xWYacegikXZbdfhjl = vld2q_u16(i0); i0 += 16;
      const uint16x8x2_t vi1xWYacegikXZbdfhjl = vld2q_u16(i1); i1 += 16;
      const uint16x8x2_t vi2xWYacegikXZbdfhjl = vld2q_u16(i2); i2 += 16;
      const uint16x8x2_t vi3xWYacegikXZbdfhjl = vld2q_u16(i3); i3 += 16;
      const uint16x8x2_t vi4xWYacegikXZbdfhjl = vld2q_u16(i4); i4 += 16;

      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi0xFHJLNPRT, vw01234567, 2);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi0xFHJLNPRT, vget_low_f16(vw01234567), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi1xFHJLNPRT, vw01234567, 7);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi1xFHJLNPRT, vget_high_f16(vw01234567), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi2xFHJLNPRT, vw89ABCDEF, 4);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi2xFHJLNPRT, vget_high_f16(vw89ABCDEF), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi3xFHJLNPRT, vwGHIJKLMN, 1);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi3xFHJLNPRT, vget_low_f16(vwGHIJKLMN), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi4xFHJLNPRT, vwGHIJKLMN, 6);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi4xFHJLNPRT, vget_high_f16(vwGHIJKLMN), 2);
      #endif
      // Right by 1 column
      const float16x8_t vi0xIKMOQSUW = vextq_f16(vreinterpretq_f16_u16(vi0xGIKMOQSUHJLNPRTV.val[0]), vreinterpretq_f16_u16(vi0xWYacegikXZbdfhjl.val[0]), 1);
      vi0xGIKMOQSUHJLNPRTV = vi0xWYacegikXZbdfhjl;
      const float16x8_t vi1xIKMOQSUW = vextq_f16(vreinterpretq_f16_u16(vi1xGIKMOQSUHJLNPRTV.val[0]), vreinterpretq_f16_u16(vi1xWYacegikXZbdfhjl.val[0]), 1);
      vi1xGIKMOQSUHJLNPRTV = vi1xWYacegikXZbdfhjl;
      const float16x8_t vi2xIKMOQSUW = vextq_f16(vreinterpretq_f16_u16(vi2xGIKMOQSUHJLNPRTV.val[0]), vreinterpretq_f16_u16(vi2xWYacegikXZbdfhjl.val[0]), 1);
      vi2xGIKMOQSUHJLNPRTV = vi2xWYacegikXZbdfhjl;
      const float16x8_t vi3xIKMOQSUW = vextq_f16(vreinterpretq_f16_u16(vi3xGIKMOQSUHJLNPRTV.val[0]), vreinterpretq_f16_u16(vi3xWYacegikXZbdfhjl.val[0]), 1);
      vi3xGIKMOQSUHJLNPRTV = vi3xWYacegikXZbdfhjl;
      const float16x8_t vi4xIKMOQSUW = vextq_f16(vreinterpretq_f16_u16(vi4xGIKMOQSUHJLNPRTV.val[0]), vreinterpretq_f16_u16(vi4xWYacegikXZbdfhjl.val[0]), 1);
      vi4xGIKMOQSUHJLNPRTV = vi4xWYacegikXZbdfhjl;

      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi0xIKMOQSUW, vw01234567, 5);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi0xIKMOQSUW, vget_high_f16(vw01234567), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi1xIKMOQSUW, vw89ABCDEF, 2);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi1xIKMOQSUW, vget_low_f16(vw89ABCDEF), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi2xIKMOQSUW, vw89ABCDEF, 7);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi2xIKMOQSUW, vget_high_f16(vw89ABCDEF), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi3xIKMOQSUW, vwGHIJKLMN, 4);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi3xIKMOQSUW, vget_high_f16(vwGHIJKLMN), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_lane_f16(vo0p2, vi4xIKMOQSUW, vwOP, 1);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi4xIKMOQSUW, vwOP, 1);
      #endif
      vo0p0 = vaddq_f16(vo0p0, vo0p1);
      vo0p0 = vaddq_f16(vo0p0, vo0p2);

      float16x8_t vo0 = vmaxq_f16(vo0p0, vmin);

      vo0 = vminq_f16(vo0, vmax);

      vst1q_u16(o0, vreinterpretq_u16_f16(vo0)); o0 += 8;
    }

    // Last block has 1-16 pixels to process.
    assert(w <= 16 * sizeof(uint16_t));
    assert(w >= 1 * sizeof(uint16_t));
    {
      float16x8_t vo0p0 = vdupq_lane_f16(vget_low_f16(vw01234567), 0);

      const float16x8_t vi0xGIKMOQSU = vreinterpretq_f16_u16(vandq_u16(vmask_even, vi0xGIKMOQSUHJLNPRTV.val[0]));
      const float16x8_t vi1xGIKMOQSU = vreinterpretq_f16_u16(vandq_u16(vmask_even, vi1xGIKMOQSUHJLNPRTV.val[0]));
      const float16x8_t vi2xGIKMOQSU = vreinterpretq_f16_u16(vandq_u16(vmask_even, vi2xGIKMOQSUHJLNPRTV.val[0]));
      const float16x8_t vi3xGIKMOQSU = vreinterpretq_f16_u16(vandq_u16(vmask_even, vi3xGIKMOQSUHJLNPRTV.val[0]));
      const float16x8_t vi4xGIKMOQSU = vreinterpretq_f16_u16(vandq_u16(vmask_even, vi4xGIKMOQSUHJLNPRTV.val[0]));

      const float16x8_t vi0xHJLNPRTV = vreinterpretq_f16_u16(vandq_u16(vmask_odd, vi0xGIKMOQSUHJLNPRTV.val[1]));
      const float16x8_t vi1xHJLNPRTV = vreinterpretq_f16_u16(vandq_u16(vmask_odd, vi1xGIKMOQSUHJLNPRTV.val[1]));
      const float16x8_t vi2xHJLNPRTV = vreinterpretq_f16_u16(vandq_u16(vmask_odd, vi2xGIKMOQSUHJLNPRTV.val[1]));
      const float16x8_t vi3xHJLNPRTV = vreinterpretq_f16_u16(vandq_u16(vmask_odd, vi3xGIKMOQSUHJLNPRTV.val[1]));
      const float16x8_t vi4xHJLNPRTV = vreinterpretq_f16_u16(vandq_u16(vmask_odd, vi4xGIKMOQSUHJLNPRTV.val[1]));

      // Center column
      float16x8_t vo0p1 = vmulq_lane_f16(vi0xGIKMOQSU, vget_low_f16(vw01234567), 3);
      float16x8_t vo0p2 = vmulq_lane_f16(vi1xGIKMOQSU, vget_low_f16(vw89ABCDEF), 0);
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi2xGIKMOQSU, vw89ABCDEF, 5);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi2xGIKMOQSU, vget_high_f16(vw89ABCDEF), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi3xGIKMOQSU, vwGHIJKLMN, 2);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi3xGIKMOQSU, vget_low_f16(vwGHIJKLMN), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi4xGIKMOQSU, vwGHIJKLMN, 7);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi4xGIKMOQSU, vget_high_f16(vwGHIJKLMN), 3);
      #endif
      // Right by 1 column
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi0xHJLNPRTV, vw01234567, 4);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi0xHJLNPRTV, vget_high_f16(vw01234567), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi1xHJLNPRTV, vw89ABCDEF, 1);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi1xHJLNPRTV, vget_low_f16(vw89ABCDEF), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi2xHJLNPRTV, vw89ABCDEF, 6);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi2xHJLNPRTV, vget_high_f16(vw89ABCDEF), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi3xHJLNPRTV, vwGHIJKLMN, 3);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi3xHJLNPRTV, vget_low_f16(vwGHIJKLMN), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_lane_f16(vo0p2, vi4xHJLNPRTV, vwOP, 0);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi4xHJLNPRTV, vwOP, 0);
      #endif
      // Left by 2 columns
      const float16x8_t vi0xEGIKMOQS = vextq_f16(vi0x02468ACE, vi0xGIKMOQSU, 7);
      const float16x8_t vi1xEGIKMOQS = vextq_f16(vi1x02468ACE, vi1xGIKMOQSU, 7);
      const float16x8_t vi2xEGIKMOQS = vextq_f16(vi2x02468ACE, vi2xGIKMOQSU, 7);
      const float16x8_t vi3xEGIKMOQS = vextq_f16(vi3x02468ACE, vi3xGIKMOQSU, 7);
      const float16x8_t vi4xEGIKMOQS = vextq_f16(vi4x02468ACE, vi4xGIKMOQSU, 7);

      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi0xEGIKMOQS, vw01234567, 1);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi0xEGIKMOQS, vget_low_f16(vw01234567), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi1xEGIKMOQS, vw01234567, 6);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi1xEGIKMOQS, vget_high_f16(vw01234567), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi2xEGIKMOQS, vw89ABCDEF, 3);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi2xEGIKMOQS, vget_low_f16(vw89ABCDEF), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi3xEGIKMOQS, vwGHIJKLMN, 0);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi3xEGIKMOQS, vget_low_f16(vwGHIJKLMN), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi4xEGIKMOQS, vwGHIJKLMN, 5);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi4xEGIKMOQS, vget_high_f16(vwGHIJKLMN), 1);
      #endif
      // Left by 1 column
      const float16x8_t vi0xFHJLNPRT = vextq_f16(vi0x13579BDF, vi0xHJLNPRTV, 7);
      const float16x8_t vi1xFHJLNPRT = vextq_f16(vi1x13579BDF, vi1xHJLNPRTV, 7);
      const float16x8_t vi2xFHJLNPRT = vextq_f16(vi2x13579BDF, vi2xHJLNPRTV, 7);
      const float16x8_t vi3xFHJLNPRT = vextq_f16(vi3x13579BDF, vi3xHJLNPRTV, 7);
      const float16x8_t vi4xFHJLNPRT = vextq_f16(vi4x13579BDF, vi4xHJLNPRTV, 7);

      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi0xFHJLNPRT, vw01234567, 2);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi0xFHJLNPRT, vget_low_f16(vw01234567), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi1xFHJLNPRT, vw01234567, 7);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi1xFHJLNPRT, vget_high_f16(vw01234567), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi2xFHJLNPRT, vw89ABCDEF, 4);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi2xFHJLNPRT, vget_high_f16(vw89ABCDEF), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi3xFHJLNPRT, vwGHIJKLMN, 1);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi3xFHJLNPRT, vget_low_f16(vwGHIJKLMN), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi4xFHJLNPRT, vwGHIJKLMN, 6);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi4xFHJLNPRT, vget_high_f16(vwGHIJKLMN), 2);
      #endif
      // Right by 2 columns
      const float16x8_t vzero = vreinterpretq_f16_u16(vmovq_n_u16(0));
      const float16x8_t vi0xIKMOQSUW = vextq_f16(vi0xGIKMOQSU, vzero, 1);
      const float16x8_t vi1xIKMOQSUW = vextq_f16(vi1xGIKMOQSU, vzero, 1);
      const float16x8_t vi2xIKMOQSUW = vextq_f16(vi2xGIKMOQSU, vzero, 1);
      const float16x8_t vi3xIKMOQSUW = vextq_f16(vi3xGIKMOQSU, vzero, 1);
      const float16x8_t vi4xIKMOQSUW = vextq_f16(vi4xGIKMOQSU, vzero, 1);

      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi0xIKMOQSUW, vw01234567, 5);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi0xIKMOQSUW, vget_high_f16(vw01234567), 1);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_laneq_f16(vo0p2, vi1xIKMOQSUW, vw89ABCDEF, 2);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi1xIKMOQSUW, vget_low_f16(vw89ABCDEF), 2);
      #endif
      #if XNN_ARCH_ARM64
        vo0p0 = vfmaq_laneq_f16(vo0p0, vi2xIKMOQSUW, vw89ABCDEF, 7);
      #else
        vo0p0 = vmlaq_lane_f16(vo0p0, vi2xIKMOQSUW, vget_high_f16(vw89ABCDEF), 3);
      #endif
      #if XNN_ARCH_ARM64
        vo0p1 = vfmaq_laneq_f16(vo0p1, vi3xIKMOQSUW, vwGHIJKLMN, 4);
      #else
        vo0p1 = vmlaq_lane_f16(vo0p1, vi3xIKMOQSUW, vget_high_f16(vwGHIJKLMN), 0);
      #endif
      #if XNN_ARCH_ARM64
        vo0p2 = vfmaq_lane_f16(vo0p2, vi4xIKMOQSUW, vwOP, 1);
      #else
        vo0p2 = vmlaq_lane_f16(vo0p2, vi4xIKMOQSUW, vwOP, 1);
      #endif
      vo0p0 = vaddq_f16(vo0p0, vo0p1);
      vo0p0 = vaddq_f16(vo0p0, vo0p2);

      float16x8_t vo0 = vmaxq_f16(vo0p0, vmin);

      vo0 = vminq_f16(vo0, vmax);

      const size_t w_tmp = (w + 1 * sizeof(uint16_t)) / (2 * sizeof(uint16_t));

      if XNN_LIKELY(w_tmp == 8) {
        vst1q_u16(o0, vreinterpretq_u16_f16(vo0)); o0 += 8;
      } else {
        float16x4_t vo0_lo = vget_low_f16(vo0);

        if (w_tmp & 4) {
         vst1_u16(o0, vreinterpret_u16_f16(vo0_lo)); o0 += 4;

          vo0_lo = vget_high_f16(vo0);
        }
        if (w_tmp & 2) {
          vst1_lane_u32((void*) o0, vreinterpret_u32_f16(vo0_lo), 0); o0 += 2;

          vo0_lo = vext_f16(vo0_lo, vo0_lo, 2);
        }
        if (w_tmp & 1) {
          vst1_lane_u16(o0, vreinterpret_u16_f16(vo0_lo), 0); o0 += 1;
        }
      }
    }

    i0 = (const uint16_t*) ((uintptr_t) i2 - input_decrement);
    i1 = (const uint16_t*) ((uintptr_t) i3 - input_decrement);
    i2 = (const uint16_t*) ((uintptr_t) i4 - input_decrement);
    i3 = (const uint16_t*) ((uintptr_t) i2 + input_width);
    i4 = (const uint16_t*) ((uintptr_t) i3 + input_width);


    output_height -= 1;
    padded_input_height -= 2;
  } while (output_height != 0);
}