/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x8c8-minmax-rndnu-neon-mull.c | 158 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local 202 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
D | 3x8c16-minmax-rndnu-neon-mlal.c | 182 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local 225 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
D | 4x8c8-minmax-rndnu-neon-mull.c | 189 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local 239 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
D | 3x8c8-minmax-rndnu-neon-mlal.c | 259 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local 303 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
D | 4x8c16-minmax-rndnu-neon-mlal.c | 221 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local 270 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 159 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c8-minmax-rndnu-neon-mlal.c | 316 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local 366 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
D | 3x16c8-minmax-rndnu-neon-mull.c | 246 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local 328 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
|
D | 3x16c16-minmax-rndnu-neon-mlal.c | 294 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local 375 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
|
D | 3x8c4-minmax-rndnu-neon-mull-dup.c | 205 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_dup() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld2r.c | 205 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld2r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 189 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld1r.c | 208 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); 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 | 200 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local 243 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
D | 3x8c8-minmax-rndnu-neon-mull.c | 176 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local 220 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
D | 4x8c16-minmax-rndnu-neon-mlal.c | 241 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local 290 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
D | 4x8c8-minmax-rndnu-neon-mull.c | 209 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local 259 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
D | 3x8c8-minmax-rndnu-neon-mlal.c | 277 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local 321 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 177 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c8-minmax-rndnu-neon-mlal.c | 336 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local 386 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
D | 3x16c8-minmax-rndnu-neon-mull.c | 264 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local 346 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
|
D | 3x16c16-minmax-rndnu-neon-mlal.c | 312 const int32x4_t vsum2x01 = vpaddq_s32(vacc2x0, vacc2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local 393 const int32x2_t vsum2x01 = vpadd_s32(vpsum2x0, vpsum2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 209 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld1r.c | 225 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld1r() local
|
D | 3x8c4-minmax-rndnu-neon-mull-dup.c | 222 const int32x2_t vsum2x01 = vpadd_s32(vget_low_s32(vacc2x01), vget_high_s32(vacc2x01)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mull_dup() local
|