Using MPIs Non-Blocking Allreduce for Health Checks in Dynamic Simulations

Dies ist die aktuellste Version dieses Eintrags.

Kurzfassung

Large-scale simulations often require frequent checks on global conditions that are not directly needed for the computation itself. While such health checks are not an integral part of the numerical algorithm, they serve an important role in controlling and coordinating the simulation. In distributed parallel computations their required communication may negatively impact the actual parallel computation and result in unnecessary synchronization points. By using a nonblocking reduction these synchronization requirements can be loosened and the impact on the actual computation minimized. Further, it enables us to shift the communication into the background and progress it during the MPI calls that are done during the computation anyway. This approach delays the decision to be made in response to those health checks. But as it is not vital for the correct computation itself such a delay is usually tolerable and could offer a more robust scaling to large process counts.

Verfügbare Versionen dieses Eintrags

• Using MPIs Non-Blocking Allreduce for Health Checks in Dynamic Simulations. (deposited 02 Nov 2023 15:29) [Gegenwärtig angezeigt]