https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91568
--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> --- Confirmed on the branch. internal compiler error: in exact_div, at poly-int.h:2139 0x121020e poly_int<1u, poly_result<unsigned long, if_nonpoly<unsigned long, unsigned long, poly_int_traits<unsigned long>::is_poly>::type, poly_coeff_pair_traits<unsigned long, if_nonpoly<unsigned long, unsigned long, poly_int_traits<unsigned long>::is_poly>::type>::result_kind>::type> exact_div<1u, unsigned long, unsigned long>(poly_int_pod<1u, unsigned long> const&, unsigned long) /space/rguenther/src/svn/gcc-9-branch/gcc/poly-int.h:2139 0x1210099 poly_int<1u, poly_result<unsigned long, unsigned long, poly_coeff_pair_traits<unsigned long, unsigned long>::result_kind>::type> exact_div<1u, unsigned long, unsigned long>(poly_int_pod<1u, unsigned long> const&, poly_int_pod<1u, unsigned long> const&) /space/rguenther/src/svn/gcc-9-branch/gcc/poly-int.h:2152 0x15df0e4 vect_get_num_vectors /space/rguenther/src/svn/gcc-9-branch/gcc/tree-vectorizer.h:1338 0x15e6e26 vect_slp_analyze_node_operations_1 /space/rguenther/src/svn/gcc-9-branch/gcc/tree-vect-slp.c:2522 trunk correctly computes a VF of two while the branch computes one.