Home
last modified time | relevance | path

Searched refs:vacc1x15 (Results 1 – 18 of 18) sorted by relevance

/external/XNNPACK/src/qs8-igemm/gen/
D2x16c8-minmax-rndnu-neon-mull.c82 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mull() local
182 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mull()
206 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mull()
269 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mull()
D2x16c8-minmax-rndnu-neon-mlal.c82 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local
232 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
321 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
345 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
408 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
D2x16c16-minmax-rndnu-neon-mlal.c82 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c16__neon_mlal() local
214 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c16__neon_mlal()
238 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c16__neon_mlal()
300 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c16__neon_mlal()
D3x16c8-minmax-rndnu-neon-mlal.c86 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() local
305 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
427 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
452 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
527 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
D3x16c8-minmax-rndnu-neon-mull.c86 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
238 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull()
263 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull()
338 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull()
D3x16c16-minmax-rndnu-neon-mlal.c86 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
286 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal()
311 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal()
385 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal()
D4x16c8-minmax-rndnu-neon-mull.c90 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mull() local
294 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mull()
320 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mull()
407 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mull()
D4x16c8-minmax-rndnu-neon-mlal.c90 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local
378 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
533 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
559 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
646 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
D4x16c16-minmax-rndnu-neon-mlal.c90 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c16__neon_mlal() local
358 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c16__neon_mlal()
384 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c16__neon_mlal()
470 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c16__neon_mlal()
/external/XNNPACK/src/qs8-gemm/gen/
D2x16c8-minmax-rndnu-neon-mull.c81 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mull() local
169 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mull()
190 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mull()
253 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mull()
D2x16c8-minmax-rndnu-neon-mlal.c81 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local
219 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
308 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
329 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
392 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
D2x16c16-minmax-rndnu-neon-mlal.c81 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c16__neon_mlal() local
201 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c16__neon_mlal()
222 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c16__neon_mlal()
284 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c16__neon_mlal()
D3x16c8-minmax-rndnu-neon-mlal.c87 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() local
290 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
412 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
434 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
509 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
D3x16c8-minmax-rndnu-neon-mull.c87 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
223 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull()
245 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull()
320 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull()
D3x16c16-minmax-rndnu-neon-mlal.c87 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
271 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal()
293 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal()
367 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal()
D4x16c8-minmax-rndnu-neon-mull.c93 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mull() local
277 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mull()
300 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mull()
387 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mull()
D4x16c8-minmax-rndnu-neon-mlal.c93 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local
361 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
516 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
539 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
626 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
D4x16c16-minmax-rndnu-neon-mlal.c93 int32x4_t vacc1x15 = vacc0x15; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c16__neon_mlal() local
341 vacc1x15 = vpadalq_s16(vacc1x15, vprod1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c16__neon_mlal()
364 const int32x4_t vsum1xEF = vpaddq_s32(vacc1x14, vacc1x15); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c16__neon_mlal()
450 const int32x2_t vpsum1xF = vadd_s32(vget_low_s32(vacc1x15), vget_high_s32(vacc1x15)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c16__neon_mlal()