When reading multiple time steps, it is not efficient to open, read, and close many ASCII files repeatedly. We should think about adding an optional first step to dump all output to something like HDF5 or Zarr to make repeated access much more efficient. Then, the Profile objects could read from that file (if it exists) and fall back to the usual ASCII file if not.
When reading multiple time steps, it is not efficient to open, read, and close many ASCII files repeatedly. We should think about adding an optional first step to dump all output to something like HDF5 or Zarr to make repeated access much more efficient. Then, the
Profileobjects could read from that file (if it exists) and fall back to the usual ASCII file if not.