https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104722
--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Jakub Jelinek from comment #6) > Perhaps _GLIBCXX_PARALLEL should parallelize std::accumulate only if > binary_op is known to be associative? Or just don't parallelize it, since it breaks the guarantees of std::accumulate. That's why the C++17 parallel algos add std::reduce, which doesn't have the same guarantees.