Unique identifiers for 263 users who received HPC training in 2022 was determined from collected attendee records. Note that users may enrol in multiple courses (e.g., Introduction to Spartan, Advanced Spartan, Parallel Processing, etc) and may return for revision. All these users are counted once only.

From the unique users a total of 212 usernames could be determined from email addresses. When enrolling for training users do not include their Spartan usernmae or their university ID; sometimes they don't even use a university email address, despite requests.

There were 97 users who established an account but did not use Spartan (compute hours = 0). Of the remaining 115 users the total of job hours was determined from trained users was 6280454, after they received training. This calculation ensured that users who had already run jobs on Spartan prior to receiving training was not counted. e.g.,

$ sreport cluster AccountUtilizationByUser cluster=spartan user=$username start=2022-11-01 end=2022-12-31 -t hours

The total allocated hours of cluster utilisation 11597951, from the command:

$ sreport cluster Utilization cluster=spartan start=2022-11-01 end=2022-12-31 -t hours

The means that at least 54.14% of cluster utilisation in 2022 was conducted by users after receiving training.

The following steps are recommended to improve record-keeping and utilisation.

1) Emphasising the need to enrollees to use University of Melbourne email addresses only, and rejecting applications that do not do this.

2) Contacting those who attended training but did not use Spartan to ascertain why this was the case.