/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x8c8-minmax-rndnu-neon-mull.c | 161 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local 210 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
D | 3x8c16-minmax-rndnu-neon-mlal.c | 185 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local 233 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
D | 4x8c8-minmax-rndnu-neon-mull.c | 192 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local 247 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
D | 3x8c8-minmax-rndnu-neon-mlal.c | 262 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local 311 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
D | 4x8c16-minmax-rndnu-neon-mlal.c | 224 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local 278 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 163 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c8-minmax-rndnu-neon-mlal.c | 319 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local 374 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
D | 3x16c8-minmax-rndnu-neon-mull.c | 249 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local 336 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
|
D | 3x16c16-minmax-rndnu-neon-mlal.c | 297 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local 383 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
|
D | 3x8c4-minmax-rndnu-neon-mull-dup.c | 209 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_dup() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld2r.c | 209 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld2r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 193 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld1r.c | 212 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld1r() local
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 3x8c16-minmax-rndnu-neon-mlal.c | 203 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local 251 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
D | 3x8c8-minmax-rndnu-neon-mull.c | 179 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local 228 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
D | 4x8c16-minmax-rndnu-neon-mlal.c | 244 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local 298 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
D | 4x8c8-minmax-rndnu-neon-mull.c | 212 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local 267 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
D | 3x8c8-minmax-rndnu-neon-mlal.c | 280 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local 329 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 181 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c8-minmax-rndnu-neon-mlal.c | 339 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local 394 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
D | 3x16c8-minmax-rndnu-neon-mull.c | 267 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local 354 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
|
D | 3x16c16-minmax-rndnu-neon-mlal.c | 315 const int32x4_t vsum2x67 = vpaddq_s32(vacc2x6, vacc2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local 401 const int32x2_t vsum2x67 = vpadd_s32(vpsum2x6, vpsum2x7); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 213 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld1r.c | 229 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld1r() local
|
D | 3x8c4-minmax-rndnu-neon-mull-dup.c | 226 const int32x2_t vsum2x67 = vpadd_s32(vget_low_s32(vacc2x67), vget_high_s32(vacc2x67)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mull_dup() local
|