Resources_used.cpupercent and resources_used.mem

Had a question on the values reported for the attributes resources_used.cpupercent and resources_used.mem in qstat -f :

When a job finishes execution , do these values represent the maximum value or average value for these attributes? Or are these the point in time values. i.e the cpu, memory consumed by the job just before it finished execution?


When the job is in running state and it has run more than 120 seconds: then it is nearly point in time values of nearest mom poll cycle. Please check these sections from this guide

3.6.1 Configuring MoM Polling Cycle
8.5.5 How MoM Polling Affects Accounting

When the job has completed: then the values are aggregated values.

Thanks for the information.
Need one clarification on what is meant by aggregated values? In our use case the job is running on one vnode.

Aggregated memory, cputime , walltime , vmem with respect to a job (running on one node or multiple nodes)