Home
last modified time | relevance | path

Searched refs:vsum2x01 (Results 1 – 24 of 24) sorted by relevance

/external/XNNPACK/src/qs8-igemm/gen/
D3x8c8-minmax-neon-mull-padal.c176 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal() local
184 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
219 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal() local
221 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal()
D3x8c16-minmax-neon-mlal-padal.c200 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal() local
208 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
243 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal() local
245 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal()
D4x8c8-minmax-neon-mull-padal.c209 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal() local
221 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
258 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal() local
260 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
D3x8c8-minmax-neon-mlal-padal.c277 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
285 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mlal_padal()
320 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
322 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mlal_padal()
D4x8c16-minmax-neon-mlal-padal.c241 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal() local
253 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
290 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal() local
292 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
D3x16c8-minmax-neon-mull-padal.c264 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal() local
280 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
345 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal() local
347 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
D4x8c8-minmax-neon-mlal-padal.c336 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
348 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal()
385 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
387 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal()
D3x16c16-minmax-neon-mlal-padal.c312 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal() local
328 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
393 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal() local
395 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
D4x16c8-minmax-neon-mull-padal.c321 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal() local
345 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
414 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal() local
416 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
D4x16c16-minmax-neon-mlal-padal.c385 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
409 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
478 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
480 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
D3x16c8-minmax-neon-mlal-padal.c453 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() local
469 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
534 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() local
536 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
D4x16c8-minmax-neon-mlal-padal.c560 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
584 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
653 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
655 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
/external/XNNPACK/src/qs8-gemm/gen/
D3x8c8-minmax-neon-mull-padal.c158 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal() local
166 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
201 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal() local
203 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mull_padal()
D3x8c16-minmax-neon-mlal-padal.c182 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal() local
190 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
225 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal() local
227 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal()
D4x8c8-minmax-neon-mull-padal.c189 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal() local
201 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
238 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal() local
240 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
D3x8c8-minmax-neon-mlal-padal.c259 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
267 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mlal_padal()
302 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
304 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mlal_padal()
D4x8c16-minmax-neon-mlal-padal.c221 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal() local
233 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
270 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal() local
272 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
D4x8c8-minmax-neon-mlal-padal.c316 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
328 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal()
365 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
367 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal()
D3x16c8-minmax-neon-mull-padal.c246 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal() local
262 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
327 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal() local
329 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
D3x16c16-minmax-neon-mlal-padal.c294 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal() local
310 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
375 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal() local
377 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
D4x16c8-minmax-neon-mull-padal.c301 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal() local
325 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
394 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal() local
396 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
D3x16c8-minmax-neon-mlal-padal.c435 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() local
451 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
516 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() local
518 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
D4x16c16-minmax-neon-mlal-padal.c365 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
389 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
458 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
460 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
D4x16c8-minmax-neon-mlal-padal.c540 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
564 int32x4_t vacc2x0123 = vpaddq_s32(vsum2x01, vsum2x23); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
633 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
635 int32x4_t vacc2x0123 = vcombine_s32(vsum2x01, vsum2x23 ); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()