/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x8c8-minmax-rndnu-neon-mull.c | 159 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local 203 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
D | 3x8c16-minmax-rndnu-neon-mlal.c | 183 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local 226 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
D | 4x8c8-minmax-rndnu-neon-mull.c | 190 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local 240 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
D | 3x8c8-minmax-rndnu-neon-mlal.c | 260 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local 304 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
D | 4x8c16-minmax-rndnu-neon-mlal.c | 222 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local 271 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 160 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c8-minmax-rndnu-neon-mlal.c | 317 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local 367 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
D | 3x16c8-minmax-rndnu-neon-mull.c | 247 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local 329 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
|
D | 3x16c16-minmax-rndnu-neon-mlal.c | 295 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local 376 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
|
D | 3x8c4-minmax-rndnu-neon-mull-dup.c | 206 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_dup() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld2r.c | 206 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld2r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 190 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld1r.c | 209 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); 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 | 201 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local 244 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
D | 3x8c8-minmax-rndnu-neon-mull.c | 177 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local 221 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
D | 4x8c16-minmax-rndnu-neon-mlal.c | 242 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local 291 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c16__neon_mlal() local
|
D | 4x8c8-minmax-rndnu-neon-mull.c | 210 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local 260 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mull() local
|
D | 3x8c8-minmax-rndnu-neon-mlal.c | 278 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local 322 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 178 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c8-minmax-rndnu-neon-mlal.c | 337 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local 387 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
D | 3x16c8-minmax-rndnu-neon-mull.c | 265 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local 347 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mull() local
|
D | 3x16c16-minmax-rndnu-neon-mlal.c | 313 const int32x4_t vsum2x23 = vpaddq_s32(vacc2x2, vacc2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local 394 const int32x2_t vsum2x23 = vpadd_s32(vpsum2x2, vpsum2x3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c16__neon_mlal() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 210 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld1r.c | 226 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld1r() local
|
D | 3x8c4-minmax-rndnu-neon-mull-dup.c | 223 const int32x2_t vsum2x23 = vpadd_s32(vget_low_s32(vacc2x23), vget_high_s32(vacc2x23)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mull_dup() local
|