Feature Requests

[mathisloevenich]

At the current state there is already some bits of information that is really useful.

I opened this discussion to get more thought on which information is useful and which is not.
This should also help future development to understand the thoughts behind some feature.

[mathisloevenich]

Some bits that I think are useful to add for a summarized representation of log files:

• earliest submission, execution and termination date and latest submission , execution and termination date
  • maybe additiional job times (time differences between these events)
• visualize the variation of resource usage into a graph
• maybe move to use pandas dataframes for some type of data ?
• summarize log files with different states together

[mathisloevenich]

I am now merging a PR that can also hold the
assigned GPU name (if exists) for example: CUDA.

This is shown if you analyze a single log but file,
but is not yet included for the summary as I did not know yet how to include it.

I have two main thoughts on that:

• summarize the jobs based on the Gpus they have been executed on.
• Otherwise you could also just mention the names of the GPU's that have been used
  (and how many jobs have been executed on them)

Be aware that this value: AssignedGpus does not necessarily exist.
There are for me three known scenarios when it comes to resources: only cpu resource data, cpu and gpu resource data without assigned value for gpu, cpu and gpu resource data with assigned value for gpu