GRID - Applications - Efficiency

This graph shows the efficiency of all jobs that are associated with an application profile over a time period. Efficiency is calculated as CPU time / (run time * # CPU). Essentially it is a measure of the percentage of the compute capacity that is being used. Average, minimum and maximum values are shown.

The following example shows a relatively idle cluster, which is why the efficiency is low. Operating at higher levels of efficiency is desirable since it allows organizations to be more productive and spend less on infrastructure.

Figure 1. GRID - Applications - Efficiency