The new order_parallel variable controls whether the ordering procedure for the direct sparse solver runs in parallel on compute clusters. For more information, see “Parallel ordering for the direct sparse solver,” Section 3.1.