Benchmarking functionalities that are not part of the main workflow

[sfmig]

We are defining workflows that are most representative of the users experience with the idea of benchmarking them.

However some functions are not part of the basic workflows (e.g. loading the output XML file, containing the result of the analysis).

Should we include these as simpler workflows here (like a XML loading one) or should we benchmark these modules / functions individually (maybe following the structure of the Python modules as in our initial asv work)?

[adamltyson]

It would be great if these could be benchmarked, but it's maybe not a top priority. If they're straightforward to do though, then great. Ideally as much would be benchmarked in the repo (all the quick stuff), as with your previous asv work.