Parallel Efficiency

The Parallel Efficiency (PE) reveals the inefficiency in splitting computation over processes and then communicating data between processes.

As with Global Effienciey, PE is a compound metric whose components reflects two important factors in achieving good parallel performance in code: 1) ensuring even distribution of computational work across processes; and 2) minimizing time communicating data between processes.

These are measured with Load Balance Efficiency (LBE) and Communication Efficiency (CE); and Parallel Efficiency is defined as the product of these two sub-metrics:

\[PE = LBE \times CE\]

In order to fully understand the formulas, you may also visit the glossary of the metrics terms.

Related programs: CalculiX solver · Communication computation trade-off · juKKR kloop · OMP Collapse ·