jeromekelleher
commented
2 weeks ago That would be excellent. I think a simple script that runs some commands, and outputs timings and output rate (Mb/s) for vcztools and bcftools would be sufficient. We don't need anything elaborate.
It would be good to have some tests to check the performance of various vcztools commands. The thing to check is that running highly-selective queries (i.e. ones that just return a few records) on large datasets is fast.
They could be like the ones @jeromekelleher ran in https://github.com/sgkit-dev/vcztools/pull/8. These are not amenable to running on CI, so it's probably fine to have some instructions to run them manually on 1000G chromosome 2 or 20 data.