756                     __m256 _sum04 = _mm256_loadu_ps(outptr0 + 32);  in conv3x3s2_pack1to8_avx()  local
759                     _sum04 = _mm256_fmadd_ps(_r09, _k00_0, _sum04);  in conv3x3s2_pack1to8_avx()
760                     _sum04 = _mm256_fmadd_ps(_r010, _k01_0, _sum04);  in conv3x3s2_pack1to8_avx()
761                     _sum04 = _mm256_fmadd_ps(_r011, _k02_0, _sum04);  in conv3x3s2_pack1to8_avx()
762                     _sum04 = _mm256_fmadd_ps(_r19, _k10_0, _sum04);  in conv3x3s2_pack1to8_avx()
763                     _sum04 = _mm256_fmadd_ps(_r110, _k11_0, _sum04);  in conv3x3s2_pack1to8_avx()
764                     _sum04 = _mm256_fmadd_ps(_r111, _k12_0, _sum04);  in conv3x3s2_pack1to8_avx()
765                     _sum04 = _mm256_fmadd_ps(_r29, _k20_0, _sum04);  in conv3x3s2_pack1to8_avx()
766                     _sum04 = _mm256_fmadd_ps(_r210, _k21_0, _sum04);  in conv3x3s2_pack1to8_avx()
767                     _sum04 = _mm256_fmadd_ps(_r211, _k22_0, _sum04);  in conv3x3s2_pack1to8_avx()
779                     _mm256_storeu_ps(outptr0 + 32, _sum04);  in conv3x3s2_pack1to8_avx()
1290                     __m256 _sum04 = _mm256_loadu_ps(outptr0 + 32);  in conv3x3s2_pack1to8_avx()  local
1292                     _sum04 = _mm256_fmadd_ps(_r09, _k00_0, _sum04);  in conv3x3s2_pack1to8_avx()
1293                     _sum04 = _mm256_fmadd_ps(_r010, _k01_0, _sum04);  in conv3x3s2_pack1to8_avx()
1294                     _sum04 = _mm256_fmadd_ps(_r011, _k02_0, _sum04);  in conv3x3s2_pack1to8_avx()
1295                     _sum04 = _mm256_fmadd_ps(_r19, _k10_0, _sum04);  in conv3x3s2_pack1to8_avx()
1296                     _sum04 = _mm256_fmadd_ps(_r110, _k11_0, _sum04);  in conv3x3s2_pack1to8_avx()
1297                     _sum04 = _mm256_fmadd_ps(_r111, _k12_0, _sum04);  in conv3x3s2_pack1to8_avx()
1298                     _sum04 = _mm256_fmadd_ps(_r29, _k20_0, _sum04);  in conv3x3s2_pack1to8_avx()
1299                     _sum04 = _mm256_fmadd_ps(_r210, _k21_0, _sum04);  in conv3x3s2_pack1to8_avx()
1300                     _sum04 = _mm256_fmadd_ps(_r211, _k22_0, _sum04);  in conv3x3s2_pack1to8_avx()
1302                     _mm256_storeu_ps(outptr0 + 32, _sum04);  in conv3x3s2_pack1to8_avx()