Simulation usage definitions

The Simulation half of the hub's /usage page reports ten figures about the tools members have run. This chapter says what each one means.

It is a reference, not a task. Read it when a figure on /usage is about to go into a report and you need to be sure what it counts — simulation users and simulation runs in particular are easy to misread, and the difference between CPU time and wall time changes what a total means. If you are not quoting one of these numbers, you do not need this page.

Two of them — Simulation Users and Simulation Runs — are drawn as charts over the period you selected. The other eight are shown underneath as single values for one point in time; selecting a point on either chart moves them to that month.

The vocabulary

A simulation run — the page also calls it a job — is one unit of work submitted by a tool: a program, its inputs, and the instructions needed to run it. Starting a tool session is not a run; asking that tool to compute something is.

A simulation user is a registered account, identified by its own login, that submitted at least one run in the period.

The ten figures

Figure	Meaning
Simulation Users	Registered users who ran one or more simulation runs in the period.
Simulation Runs	Simulation runs submitted in the period.
Total CPU Time	Processor time spent executing those runs. A run using several cores at once accumulates CPU time on each of them, so this exceeds the elapsed time. It excludes anything the run was not computing during — queue waits, file I/O.
Total Wall Time	Elapsed time from a run starting to it finishing, as a stopwatch would measure it. Anything else the machine was busy with counts against it.
Total Interaction Time	Time users spent working inside tool sessions, as distinct from the compute time above.
Users with > 10 mins of CPU Time	Users whose runs consumed more than ten minutes of CPU time — a rough separation of real work from a first look.
Avg. Number of Simulation Runs/User	Simulation runs divided by simulation users, for the period.
Avg. Time between First and Last Simulation	Mean gap, per user, between their earliest and their latest run. A measure of how long people keep coming back.
Repeat Users with > 10 Simulation Jobs	Users who have submitted more than ten runs.
Repeat Users with > 3 Months	Users who came back after a gap of more than three months.

The labels themselves are rows in the statistics database, so a hub that maintains its own metrics tooling can rename them or report different things under the same names. If a label on your hub does not match this list, trust the hub.