| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 4x16c8-minmax-neon-mull-padal.c | 305 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
306 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
307 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
308 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
309 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
310 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
311 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
312 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
313 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
314 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
[all …]
|
| D | 3x16c8-minmax-neon-mull-padal.c | 248 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
249 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
250 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
251 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
252 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
253 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
254 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
255 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
256 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
257 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
[all …]
|
| D | 4x8c8-minmax-neon-mull-padal.c | 201 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
202 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
203 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
204 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
205 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
206 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
207 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
208 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
209 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
210 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
[all …]
|
| D | 2x16c8-minmax-neon-mull-padal.c | 191 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
192 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
193 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
194 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
195 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
196 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
197 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
198 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
199 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
200 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mull_padal()
[all …]
|
| D | 4x16c16-minmax-neon-mlal-padal.c | 369 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
370 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
371 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
372 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
373 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
374 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
375 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
376 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
377 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
378 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
[all …]
|
| D | 3x16c16-minmax-neon-mlal-padal.c | 296 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
297 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
298 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
299 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
300 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
301 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
302 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
303 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
304 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
305 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
[all …]
|
| D | 3x8c8-minmax-neon-mull-padal.c | 168 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
169 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
170 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
171 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
172 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
173 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
174 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
175 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
176 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
177 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
[all …]
|
| D | 4x8c16-minmax-neon-mlal-padal.c | 233 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
234 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
235 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
236 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
237 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
238 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
239 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
240 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
241 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
242 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
[all …]
|
| D | 3x8c16-minmax-neon-mlal-padal.c | 192 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
193 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
194 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
195 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
196 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
197 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
198 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
199 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
200 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
201 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
[all …]
|
| D | 2x16c16-minmax-neon-mlal-padal.c | 223 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
224 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
225 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
226 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
227 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
228 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
229 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
230 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
231 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
232 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_2x16c16__neon_mlal_padal()
[all …]
|
| D | 4x16c8-minmax-neon-mlal-padal.c | 544 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
545 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
546 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
547 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
548 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
549 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
550 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
551 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
552 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
553 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
[all …]
|
| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 4x16c8-minmax-neon-mull-padal.c | 285 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
286 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
287 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
288 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
289 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
290 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
291 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
292 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
293 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
294 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
[all …]
|
| D | 3x16c8-minmax-neon-mull-padal.c | 230 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
231 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
232 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
233 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
234 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
235 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
236 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
237 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
238 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
239 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
[all …]
|
| D | 4x8c8-minmax-neon-mull-padal.c | 181 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
182 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
183 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
184 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
185 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
186 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
187 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
188 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
189 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
190 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
[all …]
|
| D | 2x16c8-minmax-neon-mull-padal.c | 175 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
176 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
177 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
178 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
179 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
180 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
181 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
182 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
183 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
184 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mull_padal()
[all …]
|
| D | 4x16c16-minmax-neon-mlal-padal.c | 349 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
350 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
351 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
352 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
353 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
354 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
355 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
356 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
357 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
358 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
[all …]
|
| D | 3x16c16-minmax-neon-mlal-padal.c | 278 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
279 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
280 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
281 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
282 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
283 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
284 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
285 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
286 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
287 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
[all …]
|
| D | 3x8c8-minmax-neon-mull-padal.c | 150 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
151 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
152 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
153 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
154 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
155 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
156 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
157 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
158 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
159 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
[all …]
|
| D | 2x16c16-minmax-neon-mlal-padal.c | 207 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
208 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
209 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
210 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
211 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
212 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
213 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
214 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
215 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
216 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_2x16c16__neon_mlal_padal()
[all …]
|
| D | 4x8c16-minmax-neon-mlal-padal.c | 213 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
214 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
215 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
216 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
217 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
218 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
219 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
220 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
221 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
222 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
[all …]
|
| D | 2x8c8-minmax-neon-mull-padal.c | 119 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
120 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
121 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
122 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
123 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
124 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
125 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
126 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
127 int32x4_t vacc0x0123 = vpaddq_s32(vsum0x01, vsum0x23); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
128 int32x4_t vacc0x4567 = vpaddq_s32(vsum0x45, vsum0x67); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal()
[all …]
|
| D | 3x8c16-minmax-neon-mlal-padal.c | 174 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
175 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
176 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
177 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
178 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
179 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
180 const int32x4_t vsum1x45 = vpaddq_s32(vacc1x4, vacc1x5); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
181 const int32x4_t vsum1x67 = vpaddq_s32(vacc1x6, vacc1x7); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
182 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
183 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
[all …]
|
| D | 3x16c8-minmax-neon-mlal-padal.c | 419 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
420 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
421 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
422 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
423 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
424 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
425 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
426 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
427 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
428 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
[all …]
|
| D | 4x16c8-minmax-neon-mlal-padal.c | 524 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
525 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
526 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
527 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
528 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
529 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
530 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
531 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
532 const int32x4_t vsum1x01 = vpaddq_s32(vacc1x0, vacc1x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
533 const int32x4_t vsum1x23 = vpaddq_s32(vacc1x2, vacc1x3); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
[all …]
|
| D | 1x16c8-minmax-neon-mull-padal.c | 120 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
121 const int32x4_t vsum0x23 = vpaddq_s32(vacc0x2, vacc0x3); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
122 const int32x4_t vsum0x45 = vpaddq_s32(vacc0x4, vacc0x5); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
123 const int32x4_t vsum0x67 = vpaddq_s32(vacc0x6, vacc0x7); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
124 const int32x4_t vsum0x89 = vpaddq_s32(vacc0x8, vacc0x9); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
125 const int32x4_t vsum0xAB = vpaddq_s32(vacc0x10, vacc0x11); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
126 const int32x4_t vsum0xCD = vpaddq_s32(vacc0x12, vacc0x13); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
127 const int32x4_t vsum0xEF = vpaddq_s32(vacc0x14, vacc0x15); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
128 int32x4_t vacc0x0123 = vpaddq_s32(vsum0x01, vsum0x23); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
129 int32x4_t vacc0x4567 = vpaddq_s32(vsum0x45, vsum0x67); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mull_padal()
[all …]
|