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
|
/*
* Copyright (c) 2022-2023 Arm Limited.
*
* SPDX-License-Identifier: MIT
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in all
* copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifdef __ARM_FEATURE_SVE
#ifdef ARM_COMPUTE_ENABLE_SME2
#include "arm_gemm.hpp"
#include <cstdint>
#include "../../asmlib.hpp"
#include "../../utils.hpp"
namespace arm_gemm {
void sme2_interleaved_nomerge_u8q_mopa_1VLx4VL(const uint8_t *const A, const uint8_t *const B, uint8_t *const C, int ldc, const int M, const int N, const int K, const int32_t *const bias, const Requantize32 &rq, const int n_0, bool accumulate, int32_t *const accumulator_buffer)
{
struct KernelArgs
{
KernelArgs(
const uint8_t *const A,
const uint8_t *const B,
uint8_t *const C, const int ldc,
const int M, const int N, const int K,
const int32_t *const bias,
const Requantize32 &rq,
const int n_0,
bool accumulate,
int32_t *const accumulator_buffer
) : A(A),
B(B), kstride_bytes(roundup(K, 4) * sizeof(uint8_t)),
C(C), ldcb(ldc * sizeof(uint8_t)),
M(M), N(N), K(K),
n_loops(((K / 4) - 1) / 2), n_tail_iters(((K / 4) - 1) % 2),
bias(bias), n_0(n_0),
accumulator_buffer(accumulator_buffer),
flags(0x0)
{
if (accumulate)
{
flags |= 1 << 0; // FILL_ACCUMULATORS_FROM_BUFFER
}
if (C == nullptr)
{
flags |= 1 << 1; // STORE_ACCUMULATORS_TO_BUFFER
}
if (rq.per_channel_requant)
{
flags |= 1 << 2; // PER_CHANNEL_QUANTISATION
}
}
const uint8_t *const A;
const uint8_t *const B;
const long kstride_bytes;
uint8_t *const C;
const long ldcb;
const long M, N, K, n_loops, n_tail_iters;
int32_t min = std::numeric_limits<uint8_t>::min();
int32_t max = std::numeric_limits<uint8_t>::max();
const int32_t *const bias;
const int n_0;
int32_t *const accumulator_buffer;
uint64_t flags;
};
// Construct arguments for this kernel
KernelArgs args(A, B, C, ldc, M, N, K, bias, rq, n_0, accumulate, accumulator_buffer);
__asm__ __volatile__(
"ldr x14, [%x[args], %[offsetof_flags]]\n"
".inst 0xd503477f // SMSTART ZA\n"
"ptrue p1.b\n"
".inst 0x25207811 // ptrue pn9.b\n"
"ldr x13, [%x[args], %[offsetof_accumulator_buffer]]\n"
"ldr x11, [%x[args], %[offsetof_accumulator_buffer]]\n"
"tbz x14, #0, 2f\n"
"mov x12, #0x0\n"
"cntw x20\n"
"1:" // Initial accumulator load from buffer: Loop
".inst 0xa040c5ac // ld1w { z12.s-z15.s }, pn9.b/Z, [x13]\n"
".inst 0xc0840580 // mova za0h.s[x12], { z12.s-z15.s }\n"
".inst 0xa041c5bc // ld1w { z28.s-z31.s }, pn9.b/Z, [x13, #0x4, MUL VL]\n"
".inst 0xc0840781 // mova za1h.s[x12], { z28.s-z31.s }\n"
".inst 0xa042c5bc // ld1w { z28.s-z31.s }, pn9.b/Z, [x13, #0x8, MUL VL]\n"
".inst 0xc0840782 // mova za2h.s[x12], { z28.s-z31.s }\n"
".inst 0xa043c5a4 // ld1w { z4.s-z7.s }, pn9.b/Z, [x13, #0xc, MUL VL]\n"
".inst 0xc0840483 // mova za3h.s[x12], { z4.s-z7.s }\n"
"add x12, x12, #0x4\n"
"cmp x12, x20\n"
"addvl x13, x13, #16\n"
"blt 1b\n"
"2:" // Initial accumulator load from buffer: End
"ldr w10, [%x[args], %[offsetof_M]]\n"
"mov x9, #0x0\n"
"mov x28, #0x0\n"
"ldr w27, [%x[args], %[offsetof_N]]\n"
"ldr x26, [%x[args], %[offsetof_A]]\n"
"3:" // M and N loop
"mov x25, x26\n"
".inst 0x25bb6790 // whilelt pn8.s, x28, x27, VLx4\n"
"tbnz x14, #0, 4f\n"
"ldr x20, [%x[args], %[offsetof_bias]]\n"
".inst 0xc00800ff // zero { zad0, zad1, zad2, zad3, zad4, zad5, zad6, zad7 }\n"
"cbz x20, 5f\n"
".inst 0xa01cc299 // ldnt1w { z24.s-z27.s }, p8/Z, [x20, x28, LSL #2]\n"
".inst 0xc0902700 // addha za0.s, p1/M, p1/M, z24.s\n"
".inst 0xc0902721 // addha za1.s, p1/M, p1/M, z25.s\n"
".inst 0xc0902742 // addha za2.s, p1/M, p1/M, z26.s\n"
".inst 0xc0902763 // addha za3.s, p1/M, p1/M, z27.s\n"
"4:" // Prepare accumulators: Test for last block
"mov x20, x28\n"
"mov x21, x9\n"
"incw x20, ALL, MUL #4\n"
"incw x21\n"
"cmp x20, x27\n"
"csel x21, x9, x21, LT\n"
"mov x20, x14\n"
"bfm x14, XZR, #0x0, #0x0 // bfc x14, #0x0, #0x1\n"
"cmp x21, x10\n"
"csel x14, x20, x14, LT\n"
"5:" // Prepare accumulators: End
"ldr x20, [%x[args], %[offsetof_K]]\n"
"add x20, x20, #0x3\n"
"lsr x20, x20, #0x2\n"
"ldr x23, [%x[args], %[offsetof_B]]\n"
"lsr x22, x20, #0x2\n"
"and x21, x20, #0x3\n"
"ldr x20, [%x[args], %[offsetof_kstride_bytes]]\n"
"madd x23, x28, x20, x23\n" // bptr = B + n * kstride_bytes
"cbz x22, 8f\n"
"subs x22, x22, #0x1\n"
"ld1b { z10.b }, p1/Z, [x25]\n"
".inst 0xa04086fd // ldnt1b { z28.b-z31.b }, pn9.b/Z, [x23]\n"
"ld1b { z16.b }, p1/Z, [x25, #1, MUL VL]\n"
".inst 0xa04186ed // ldnt1b { z12.b-z15.b }, pn9.b/Z, [x23, #0x4, MUL VL]\n"
"ld1b { z21.b }, p1/Z, [x25, #2, MUL VL]\n"
".inst 0xa04286f9 // ldnt1b { z24.b-z27.b }, pn9.b/Z, [x23, #0x8, MUL VL]\n"
"ld1b { z19.b }, p1/Z, [x25, #3, MUL VL]\n"
"addvl x25, x25, #4\n"
".inst 0xa04386e1 // ldnt1b { z0.b-z3.b }, pn9.b/Z, [x23, #0xc, MUL VL]\n"
"addvl x23, x23, #16\n"
"ble 7f\n"
"6:" // K loop
".inst 0xa1bc2540 // umopa za0.s, p1/M, p1/M, z10.b, z28.b\n"
"subs x22, x22, #0x1\n"
".inst 0xa1bd2541 // umopa za1.s, p1/M, p1/M, z10.b, z29.b\n"
".inst 0xa1be2542 // umopa za2.s, p1/M, p1/M, z10.b, z30.b\n"
".inst 0xa1bf2543 // umopa za3.s, p1/M, p1/M, z10.b, z31.b\n"
"ld1b { z10.b }, p1/Z, [x25]\n"
".inst 0xa1ac2600 // umopa za0.s, p1/M, p1/M, z16.b, z12.b\n"
".inst 0xa04086fd // ldnt1b { z28.b-z31.b }, pn9.b/Z, [x23]\n"
".inst 0xa1ad2601 // umopa za1.s, p1/M, p1/M, z16.b, z13.b\n"
".inst 0xa1ae2602 // umopa za2.s, p1/M, p1/M, z16.b, z14.b\n"
".inst 0xa1af2603 // umopa za3.s, p1/M, p1/M, z16.b, z15.b\n"
"ld1b { z16.b }, p1/Z, [x25, #1, MUL VL]\n"
".inst 0xa1b826a0 // umopa za0.s, p1/M, p1/M, z21.b, z24.b\n"
".inst 0xa04186ed // ldnt1b { z12.b-z15.b }, pn9.b/Z, [x23, #0x4, MUL VL]\n"
".inst 0xa1b926a1 // umopa za1.s, p1/M, p1/M, z21.b, z25.b\n"
".inst 0xa1ba26a2 // umopa za2.s, p1/M, p1/M, z21.b, z26.b\n"
".inst 0xa1bb26a3 // umopa za3.s, p1/M, p1/M, z21.b, z27.b\n"
"ld1b { z21.b }, p1/Z, [x25, #2, MUL VL]\n"
".inst 0xa04286f9 // ldnt1b { z24.b-z27.b }, pn9.b/Z, [x23, #0x8, MUL VL]\n"
".inst 0xa1a02660 // umopa za0.s, p1/M, p1/M, z19.b, z0.b\n"
".inst 0xa1a12661 // umopa za1.s, p1/M, p1/M, z19.b, z1.b\n"
".inst 0xa1a22662 // umopa za2.s, p1/M, p1/M, z19.b, z2.b\n"
".inst 0xa1a32663 // umopa za3.s, p1/M, p1/M, z19.b, z3.b\n"
"ld1b { z19.b }, p1/Z, [x25, #3, MUL VL]\n"
"addvl x25, x25, #4\n"
".inst 0xa04386e1 // ldnt1b { z0.b-z3.b }, pn9.b/Z, [x23, #0xc, MUL VL]\n"
"addvl x23, x23, #16\n"
"bgt 6b\n"
"7:" // K loop tail
".inst 0xa1bc2540 // umopa za0.s, p1/M, p1/M, z10.b, z28.b\n"
".inst 0xa1bd2541 // umopa za1.s, p1/M, p1/M, z10.b, z29.b\n"
".inst 0xa1be2542 // umopa za2.s, p1/M, p1/M, z10.b, z30.b\n"
".inst 0xa1bf2543 // umopa za3.s, p1/M, p1/M, z10.b, z31.b\n"
".inst 0xa1ac2600 // umopa za0.s, p1/M, p1/M, z16.b, z12.b\n"
".inst 0xa1ad2601 // umopa za1.s, p1/M, p1/M, z16.b, z13.b\n"
".inst 0xa1ae2602 // umopa za2.s, p1/M, p1/M, z16.b, z14.b\n"
".inst 0xa1af2603 // umopa za3.s, p1/M, p1/M, z16.b, z15.b\n"
".inst 0xa1b826a0 // umopa za0.s, p1/M, p1/M, z21.b, z24.b\n"
".inst 0xa1b926a1 // umopa za1.s, p1/M, p1/M, z21.b, z25.b\n"
".inst 0xa1ba26a2 // umopa za2.s, p1/M, p1/M, z21.b, z26.b\n"
".inst 0xa1bb26a3 // umopa za3.s, p1/M, p1/M, z21.b, z27.b\n"
".inst 0xa1a02660 // umopa za0.s, p1/M, p1/M, z19.b, z0.b\n"
".inst 0xa1a12661 // umopa za1.s, p1/M, p1/M, z19.b, z1.b\n"
".inst 0xa1a22662 // umopa za2.s, p1/M, p1/M, z19.b, z2.b\n"
".inst 0xa1a32663 // umopa za3.s, p1/M, p1/M, z19.b, z3.b\n"
"8:" // K oddments
"cbz x21, 10f\n"
"9:" // K oddments: Loop
"ld1b { z10.b }, p1/Z, [x25]\n"
"subs x21, x21, #0x1\n"
"addvl x25, x25, #1\n"
".inst 0xa04086fc // ld1b { z28.b-z31.b }, pn9.b/Z, [x23]\n"
"addvl x23, x23, #4\n"
".inst 0xa1bc2540 // umopa za0.s, p1/M, p1/M, z10.b, z28.b\n"
".inst 0xa1bd2541 // umopa za1.s, p1/M, p1/M, z10.b, z29.b\n"
".inst 0xa1be2542 // umopa za2.s, p1/M, p1/M, z10.b, z30.b\n"
".inst 0xa1bf2543 // umopa za3.s, p1/M, p1/M, z10.b, z31.b\n"
"bgt 9b\n"
"10:" // K oddments: End
"ld1w { z14.s }, p1/Z, [x25]\n"
"addvl x25, x25, #1\n"
".inst 0xc09125c0 // addva za0.s, p1/M, p1/M, z14.s\n"
".inst 0xc09125c1 // addva za1.s, p1/M, p1/M, z14.s\n"
".inst 0xc09125c2 // addva za2.s, p1/M, p1/M, z14.s\n"
".inst 0xc09125c3 // addva za3.s, p1/M, p1/M, z14.s\n"
"tbz x14, #1, 14f\n"
"tbz x14, #0, 12f\n"
"mov x12, #0x0\n"
"cntw x20\n"
"11:" // Store to partial result buffer: Store and refill: Loop
".inst 0xa040c5b8 // ld1w { z24.s-z27.s }, pn9.b/Z, [x13]\n"
".inst 0xc086041c // mova { z28.s-z31.s }, za0h.s[x12]\n"
".inst 0xc0840700 // mova za0h.s[x12], { z24.s-z27.s }\n"
".inst 0xc0860428 // mova { z8.s-z11.s }, za1h.s[x12]\n"
".inst 0xa041c5b0 // ld1w { z16.s-z19.s }, pn9.b/Z, [x13, #0x4, MUL VL]\n"
".inst 0xc0840601 // mova za1h.s[x12], { z16.s-z19.s }\n"
".inst 0xc0860458 // mova { z24.s-z27.s }, za2h.s[x12]\n"
".inst 0xc086046c // mova { z12.s-z15.s }, za3h.s[x12]\n"
".inst 0xa042c5a4 // ld1w { z4.s-z7.s }, pn9.b/Z, [x13, #0x8, MUL VL]\n"
".inst 0xc0840482 // mova za2h.s[x12], { z4.s-z7.s }\n"
".inst 0xa043c5a4 // ld1w { z4.s-z7.s }, pn9.b/Z, [x13, #0xc, MUL VL]\n"
".inst 0xc0840483 // mova za3h.s[x12], { z4.s-z7.s }\n"
"add x12, x12, #0x4\n"
"cmp x12, x20\n"
".inst 0xa060c57c // st1w { z28.s-z31.s }, pn9.b, [x11]\n"
"addvl x13, x13, #16\n"
".inst 0xa061c568 // st1w { z8.s-z11.s }, pn9.b, [x11, #0x4, MUL VL]\n"
".inst 0xa062c578 // st1w { z24.s-z27.s }, pn9.b, [x11, #0x8, MUL VL]\n"
".inst 0xa063c56c // st1w { z12.s-z15.s }, pn9.b, [x11, #0xc, MUL VL]\n"
"addvl x11, x11, #16\n"
"blt 11b\n"
"b 21f\n"
"12:" // Store to partial result buffer: Store only
"mov x12, #0x0\n"
"cntw x20\n"
"13:" // Store to partial result buffer: Store only: Loop
".inst 0xc086041c // mova { z28.s-z31.s }, za0h.s[x12]\n"
".inst 0xc0860420 // mova { z0.s-z3.s }, za1h.s[x12]\n"
".inst 0xa060c57c // st1w { z28.s-z31.s }, pn9.b, [x11]\n"
".inst 0xc0860448 // mova { z8.s-z11.s }, za2h.s[x12]\n"
".inst 0xc0860470 // mova { z16.s-z19.s }, za3h.s[x12]\n"
".inst 0xa061c560 // st1w { z0.s-z3.s }, pn9.b, [x11, #0x4, MUL VL]\n"
"add x12, x12, #0x4\n"
"cmp x12, x20\n"
".inst 0xa062c568 // st1w { z8.s-z11.s }, pn9.b, [x11, #0x8, MUL VL]\n"
".inst 0xa063c570 // st1w { z16.s-z19.s }, pn9.b, [x11, #0xc, MUL VL]\n"
"addvl x11, x11, #16\n"
"blt 13b\n"
"b 21f\n"
"14:" // Store to output array
"ldr x24, [%x[args], %[offsetof_C]]\n"
"add x24, x24, x28\n" // C += n
"sub x23, x10, x9\n"
"ld1rw { z12.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_mul]]\n"
"ldr x22, [%x[args], %[offsetof_ldcb]]\n"
"madd x24, x9, x22, x24\n" // C += m * ldc
"ld1rw { z13.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_mul]]\n"
"ld1rw { z14.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_mul]]\n"
"ld1rw { z15.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_mul]]\n"
"ld1rw { z4.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_right_shift]]\n"
"ld1rw { z5.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_right_shift]]\n"
"ld1rw { z6.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_right_shift]]\n"
"ld1rw { z7.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_per_layer_right_shift]]\n"
"ld1rw { z1.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_c_offset]]\n"
"ld1rw { z21.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_minval]]\n"
"ld1rw { z20.s }, p1/Z, [%x[rq], %[offsetof_Requantize32_maxval]]\n"
"tbz x14, #2, 15f\n"
"ldr w21, [%x[args], %[offsetof_n_0]]\n"
"add x21, x21, x28\n"
"ldr x20, [%x[rq], %[offsetof_Requantize32_per_channel_muls]]\n"
"add x20, x20, x21, LSL #2\n"
".inst 0xa040c28c // ld1w { z12.s-z15.s }, p8/Z, [x20]\n"
"ldr x20, [%x[rq], %[offsetof_Requantize32_per_channel_right_shifts]]\n"
"add x20, x20, x21, LSL #2\n"
".inst 0xa040c284 // ld1w { z4.s-z7.s }, p8/Z, [x20]\n"
"15:" // Store to output array: Load per-channel parameters: End
"cntw x20\n"
"whilelt p0.b, x28, x27\n"
"cmp x23, x20\n"
"csel x20, x23, x20, LT\n"
"lsr x21, x20, #0x1\n"
"mov x12, #0x0\n"
"and x20, x20, #0x1\n"
"cbz x21, 17f\n"
"16:" // Store to output array: Accumulator row 0 loop
".inst 0xc086001a // mova { z26.s-z27.s }, za0h.s[x12, 0:1]\n"
".inst 0xc086005c // mova { z28.s-z29.s }, za1h.s[x12, 0:1]\n"
".inst 0xc1aca41a // sqdmulh { z26.s-z27.s }, { z26.s-z27.s }, z12.s\n"
".inst 0xc0860096 // mova { z22.s-z23.s }, za2h.s[x12, 0:1]\n"
".inst 0xc08600d0 // mova { z16.s-z17.s }, za3h.s[x12, 0:1]\n"
".inst 0xc1ada41c // sqdmulh { z28.s-z29.s }, { z28.s-z29.s }, z13.s\n"
".inst 0xc1aea416 // sqdmulh { z22.s-z23.s }, { z22.s-z23.s }, z14.s\n"
"add x12, x12, #0x2\n"
"cmp x12, x21, LSL #1\n"
".inst 0xc1afa410 // sqdmulh { z16.s-z17.s }, { z16.s-z17.s }, z15.s\n"
".inst 0xc1a4a23a // srshl { z26.s-z27.s }, { z26.s-z27.s }, z4.s\n"
".inst 0xc1a5a23c // srshl { z28.s-z29.s }, { z28.s-z29.s }, z5.s\n"
".inst 0xc1a6a236 // srshl { z22.s-z23.s }, { z22.s-z23.s }, z6.s\n"
".inst 0xc1a7a230 // srshl { z16.s-z17.s }, { z16.s-z17.s }, z7.s\n"
".inst 0xc1a1a31a // add { z26.s-z27.s }, { z26.s-z27.s }, z1.s\n"
".inst 0xc1a1a31c // add { z28.s-z29.s }, { z28.s-z29.s }, z1.s\n"
".inst 0xc1a1a316 // add { z22.s-z23.s }, { z22.s-z23.s }, z1.s\n"
".inst 0xc1a1a310 // add { z16.s-z17.s }, { z16.s-z17.s }, z1.s\n"
".inst 0xc1b4c6ba // sclamp { z26.s-z27.s }, z21.s, z20.s\n"
".inst 0xc1b4c6bc // sclamp { z28.s-z29.s }, z21.s, z20.s\n"
"uzp1 z19.b, z26.b, z28.b\n"
".inst 0xc1b4c6b6 // sclamp { z22.s-z23.s }, z21.s, z20.s\n"
".inst 0xc1b4c6b0 // sclamp { z16.s-z17.s }, z21.s, z20.s\n"
"uzp1 z16.b, z22.b, z16.b\n"
"uzp1 z18.b, z27.b, z29.b\n"
"uzp1 z17.b, z23.b, z17.b\n"
"uzp1 z16.b, z19.b, z16.b\n"
"st1b { z16.b }, p0, [x24]\n"
"add x24, x24, x22\n"
"uzp1 z16.b, z18.b, z17.b\n"
"st1b { z16.b }, p0, [x24]\n"
"add x24, x24, x22\n"
"blt 16b\n"
"17:" // Store to output array: Accumulator row 0 oddments
"cbz x20, 18f\n"
".inst 0xc0860002 // mova { z2.s-z3.s }, za0h.s[x12, 0:1]\n"
".inst 0xc0860058 // mova { z24.s-z25.s }, za1h.s[x12, 0:1]\n"
".inst 0xc1aca402 // sqdmulh { z2.s-z3.s }, { z2.s-z3.s }, z12.s\n"
".inst 0xc0860090 // mova { z16.s-z17.s }, za2h.s[x12, 0:1]\n"
".inst 0xc08600ca // mova { z10.s-z11.s }, za3h.s[x12, 0:1]\n"
".inst 0xc1ada418 // sqdmulh { z24.s-z25.s }, { z24.s-z25.s }, z13.s\n"
".inst 0xc1aea410 // sqdmulh { z16.s-z17.s }, { z16.s-z17.s }, z14.s\n"
".inst 0xc1afa40a // sqdmulh { z10.s-z11.s }, { z10.s-z11.s }, z15.s\n"
".inst 0xc1a4a222 // srshl { z2.s-z3.s }, { z2.s-z3.s }, z4.s\n"
".inst 0xc1a5a238 // srshl { z24.s-z25.s }, { z24.s-z25.s }, z5.s\n"
".inst 0xc1a6a230 // srshl { z16.s-z17.s }, { z16.s-z17.s }, z6.s\n"
".inst 0xc1a7a22a // srshl { z10.s-z11.s }, { z10.s-z11.s }, z7.s\n"
".inst 0xc1a1a302 // add { z2.s-z3.s }, { z2.s-z3.s }, z1.s\n"
".inst 0xc1a1a318 // add { z24.s-z25.s }, { z24.s-z25.s }, z1.s\n"
".inst 0xc1a1a310 // add { z16.s-z17.s }, { z16.s-z17.s }, z1.s\n"
".inst 0xc1a1a30a // add { z10.s-z11.s }, { z10.s-z11.s }, z1.s\n"
".inst 0xc1b4c6a2 // sclamp { z2.s-z3.s }, z21.s, z20.s\n"
".inst 0xc1b4c6b8 // sclamp { z24.s-z25.s }, z21.s, z20.s\n"
"uzp1 z23.b, z2.b, z24.b\n"
".inst 0xc1b4c6b0 // sclamp { z16.s-z17.s }, z21.s, z20.s\n"
".inst 0xc1b4c6aa // sclamp { z10.s-z11.s }, z21.s, z20.s\n"
"uzp1 z16.b, z16.b, z10.b\n"
"uzp1 z16.b, z23.b, z16.b\n"
"st1b { z16.b }, p0, [x24]\n"
"18:" // Store to output array: Accumulator row 0 oddments: End
"19:" // Store to output array: End
"tbz x14, #0, 21f\n"
"mov x12, #0x0\n"
"cntw x20\n"
"20:" // Store to output array: Refill accumulators: Loop
".inst 0xa040c5b0 // ld1w { z16.s-z19.s }, pn9.b/Z, [x13]\n"
".inst 0xc0840600 // mova za0h.s[x12], { z16.s-z19.s }\n"
".inst 0xa041c5ac // ld1w { z12.s-z15.s }, pn9.b/Z, [x13, #0x4, MUL VL]\n"
".inst 0xc0840581 // mova za1h.s[x12], { z12.s-z15.s }\n"
".inst 0xa042c5b0 // ld1w { z16.s-z19.s }, pn9.b/Z, [x13, #0x8, MUL VL]\n"
".inst 0xc0840602 // mova za2h.s[x12], { z16.s-z19.s }\n"
".inst 0xa043c5ac // ld1w { z12.s-z15.s }, pn9.b/Z, [x13, #0xc, MUL VL]\n"
".inst 0xc0840583 // mova za3h.s[x12], { z12.s-z15.s }\n"
"add x12, x12, #0x4\n"
"cmp x12, x20\n"
"addvl x13, x13, #16\n"
"blt 20b\n"
"21:" // End block
"incw x28, ALL, MUL #4\n"
"cmp x28, x27\n"
"blt 3b\n"
"incw x9\n"
"cmp x9, x10\n"
"mov x28, #0x0\n"
"mov x26, x25\n"
"blt 3b\n"
".inst 0xd503467f // SMSTOP\n"
:
: [args] "r" (&args), [offsetof_A] "I" (offsetof(KernelArgs, A)), [offsetof_B] "I" (offsetof(KernelArgs, B)), [offsetof_C] "I" (offsetof(KernelArgs, C)), [offsetof_K] "I" (offsetof(KernelArgs, K)), [offsetof_M] "I" (offsetof(KernelArgs, M)), [offsetof_N] "I" (offsetof(KernelArgs, N)), [offsetof_Requantize32_c_offset] "I" (offsetof(Requantize32, c_offset)), [offsetof_Requantize32_maxval] "I" (offsetof(Requantize32, maxval)), [offsetof_Requantize32_minval] "I" (offsetof(Requantize32, minval)), [offsetof_Requantize32_per_channel_muls] "I" (offsetof(Requantize32, per_channel_muls)), [offsetof_Requantize32_per_channel_right_shifts] "I" (offsetof(Requantize32, per_channel_right_shifts)), [offsetof_Requantize32_per_layer_mul] "I" (offsetof(Requantize32, per_layer_mul)), [offsetof_Requantize32_per_layer_right_shift] "I" (offsetof(Requantize32, per_layer_right_shift)), [offsetof_accumulator_buffer] "I" (offsetof(KernelArgs, accumulator_buffer)), [offsetof_bias] "I" (offsetof(KernelArgs, bias)), [offsetof_flags] "I" (offsetof(KernelArgs, flags)), [offsetof_kstride_bytes] "I" (offsetof(KernelArgs, kstride_bytes)), [offsetof_ldcb] "I" (offsetof(KernelArgs, ldcb)), [offsetof_n_0] "I" (offsetof(KernelArgs, n_0)), [rq] "r" (&rq)
: "cc", "memory", "p0", "p1", "p2", "p3", "p4", "p5", "p6", "p7", "p8", "p9", "p10", "p11", "p12", "p13", "p14", "p15", "x9", "x10", "x11", "x12", "x13", "x14", "x20", "x21", "x22", "x23", "x24", "x25", "x26", "x27", "x28", "z0", "z1", "z2", "z3", "z4", "z5", "z6", "z7", "z8", "z9", "z10", "z11", "z12", "z13", "z14", "z15", "z16", "z17", "z18", "z19", "z20", "z21", "z22", "z23", "z24", "z25", "z26", "z27", "z28", "z29", "z30", "z31"
);
}
} // namespace arm_gemm
#endif // ARM_COMPUTE_ENABLE_SME2
#endif // __ARM_FEATURE_SVE
|
