Home
last modified time | relevance | path

Searched refs:vacc2x10 (Results 1 – 12 of 12) sorted by relevance

/external/XNNPACK/src/qs8-igemm/gen/
D3x16c8-minmax-neon-mull-padal.c97 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal() local
204 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
269 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
357 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
D3x16c8-minmax-neon-mlal-padal.c97 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() local
256 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
393 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
458 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
546 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
D3x16c16-minmax-neon-mlal-padal.c97 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal() local
242 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
317 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
405 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
D4x16c8-minmax-neon-mull-padal.c101 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal() local
250 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
326 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
426 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
D4x16c8-minmax-neon-mlal-padal.c101 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
314 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
489 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
565 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
665 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
D4x16c16-minmax-neon-mlal-padal.c101 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
299 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
390 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
490 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
/external/XNNPACK/src/qs8-gemm/gen/
D3x16c8-minmax-neon-mull-padal.c98 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal() local
189 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
251 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
339 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
D3x16c8-minmax-neon-mlal-padal.c98 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() local
241 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
378 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
440 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
528 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
D3x16c16-minmax-neon-mlal-padal.c98 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal() local
227 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
299 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
387 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
D4x16c8-minmax-neon-mull-padal.c104 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal() local
233 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
306 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
406 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
D4x16c8-minmax-neon-mlal-padal.c104 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
297 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
472 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
545 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
645 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
D4x16c16-minmax-neon-mlal-padal.c104 int32x4_t vacc2x10 = vacc0x10; in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
282 vacc2x10 = vpadalq_s16(vacc2x10, vprod2x10); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
370 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
470 const int32x2_t vpsum2xA = vadd_s32(vget_low_s32(vacc2x10), vget_high_s32(vacc2x10)); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()