Given a qdo slurm log file, this computes the fraction of launched bricks that resulted in a tractor table to the total number of launched bricksself.
This is a useful test of scaling. More nodes with more time should create more tractor catalogues than less nodes with less time.
Functions
add_fits(text) |
|
trac_fns(slurm_fn) |
Map the string ‘logging to’ to the tractor catalogue name that Should be created for that job |