BOSpulse 5.2 Improved Performance On Multi-core Processors

While BOSpulse provides you with very fast flow solvers, the time required to run one or more analyses can never be short enough. The new BOSpulse release therefore comes with a new scheduling algorithm that ensures that all available processor cores are used in an optimal way, especially when transitioning from one simulation case to another.

The new scheduling algorithm breaks each simulation case into multiple, parallel tasks that are put in a single run queue. The algorithm removes a task from the queue whenever a processor core becomes available. The algorithm handles dependencies between tasks so that a task is not started before all its required input data have become available.

bosfluids, bosfluids 7.2, multi-core processors

The effectiveness of the new scheduling algorithm is especially apparent when executing multiple cases involving parameter studies. Indeed, the overall throughput time can be reduced by more than a third in some cases. The downside to this is that you might be left with less coffee breaks.