Description: A new parallel ordering procedure for the direct sparse solver will be used when the direct sparse solver is executed on multiple host machines of a computer cluster. This is an MPI-only feature and will not be invoked when run in threaded-only parallel mode. The parallel ordering procedure will produce different orderings on different numbers of host machines, and performance may degrade in some cases. It is expected that the parallel ordering procedure will improve overall performance of Abaqus/Standard in the majority of large analyses. For cases where this is not true and performance is negatively impacted, this feature can be disabled using the command line or environment file parameter order_parallel=OFF.