| /aosp_15_r20/external/XNNPACK/src/qs8-igemm/gen/ |
| H A D | 4x8c8-minmax-rndnu-neon-mull.c | 213 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
273 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
| H A D | 4x8c16-minmax-rndnu-neon-mlal.c | 245 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
304 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
| H A D | 4x8c8-minmax-rndnu-neon-mlal.c | 340 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
400 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
| H A D | 4x16c8-minmax-rndnu-neon-mull.c | 329 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mull() local
443 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mull() local
|
| H A D | 4x16c16-minmax-rndnu-neon-mlal.c | 393 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c16__neon_mlal() local
506 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c16__neon_mlal() local
|
| H A D | 4x8c4s2-minmax-rndnu-neon-mull.c | 215 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
| H A D | 4x16c8-minmax-rndnu-neon-mlal.c | 568 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local
682 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mull-dup.c | 271 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_dup() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mull-ld1r.c | 275 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld1r() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mull-ld2r.c | 271 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld2r() local
|
| H A D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 348 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 415 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
| /aosp_15_r20/external/XNNPACK/src/qs8-gemm/gen/ |
| H A D | 4x8c16-minmax-rndnu-neon-mlal.c | 225 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
284 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
| H A D | 4x8c8-minmax-rndnu-neon-mull.c | 193 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
253 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
| H A D | 4x8c8-minmax-rndnu-neon-mlal.c | 320 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
380 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
| H A D | 4x16c8-minmax-rndnu-neon-mull.c | 309 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mull() local
423 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mull() local
|
| H A D | 4x16c16-minmax-rndnu-neon-mlal.c | 373 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c16__neon_mlal() local
486 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c16__neon_mlal() local
|
| H A D | 4x8c4s2-minmax-rndnu-neon-mull.c | 195 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
| H A D | 4x16c8-minmax-rndnu-neon-mlal.c | 548 const int32x4_t vsum3x01 = vpaddq_s32(vacc3x0, vacc3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local
662 const int32x2_t vsum3x01 = vpadd_s32(vpsum3x0, vpsum3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mull-dup.c | 252 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_dup() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mull-ld2r.c | 252 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld2r() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mull-ld1r.c | 256 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld1r() local
|
| H A D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 328 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
| H A D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 396 const int32x2_t vsum3x01 = vpadd_s32(vget_low_s32(vacc3x01), vget_high_s32(vacc3x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/quantized/cpu/qnnpack/src/q8gemm/ |
| H A D | 4x-sumrows-neon.c | 106 const uint32x2_t vsum3x01 = in pytorch_q8sumrows_ukernel_4x__neon() local
|