https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122074
--- Comment #3 from rockeet <rockeet at gmail dot com> --- (In reply to Andrew Pinski from comment #1) > I dont see how this is wrong code as you do a full load via > *(__m256i_u*)data . Suffix "_u" in __m256i_u emphpasize we are using an unaligned vector which should be processed specially, with `_mm256_mask_cmpge_epi8_mask` which has the ability "Masked Fault Suppression", so the combination of the two expressed the complete intention. intel icc has no `__m256i_u`.
