https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111494
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> --- Created attachment 60250 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=60250&action=edit patch OK, so it at least got re-broken when refactoring reduction vectorization to rely on vectorizable_* for vectorizing the reduction stmt. So like the attached, generating vect__1.10_27 = MEM <vector(4) int> [(int *)vectp_a.8_25]; vect_sum_6.13_29 = VIEW_CONVERT_EXPR<vector(4) unsigned int>(vect__1.10_27); vect_sum_6.13_30 = VIEW_CONVERT_EXPR<vector(4) unsigned int>(vect_sum_10.11_28); vect_sum_6.13_31 = vect_sum_6.13_29 + vect_sum_6.13_30; vect_sum_6.12_32 = VIEW_CONVERT_EXPR<vector(4) int>(vect_sum_6.13_31); _1 = a[i_12]; but it's a bit of a local hack and it might be better to promote the whole reduction variable to unsigned.
