Process Efficiency considers sources of inefficiency which occur in the process parallelisation. It ignores the threads, by treating the following as useful:
Hence it measures inefficiency arising from:
Inefficiency due to thread MPI (i.e. inside OpenMP regions) is ignored, as it is treated as a thread inefficiency.
Process Efficiency sub-metrics are Process Load Balance Efficiency, and Process Communication Efficiency.
\[Process Efficiency = \frac{Avg(OpenMP + Serial Useful)}{TotalRuntime} = Process\_CE + Process\_LBE - 1\]In order to fully understand the formulas, you may also visit the glossary of the metrics terms.