Diagnostic output from multiple runs within the same working directory

[cemitch99]

When HIPACE++ is run multiple times from within the same working directory, it appears that the diagnostic data from these runs is collected in a single diags folder.

• For output in diags/insitu/, the data from multiple runs appears to be superimposed.
• For output in diags/hdf5/, it appears that only the data from the final run is available.

Is there a way to parse the diagnostic output from the multiple runs separately?

If not, perhaps it would be beneficial to move any existing instance of diags to a newly-named folder such as diags.oldXXX to avoid confusion.

[MaxThevenet]

Thanks for your question! It is true that the results will be overwritten/interfere when re-running a simulation. For the moment, our typical way of addressing that is to specify a different output directory for different simulation through input parameters:

hipace.file_prefix = diags_2
beams.insitu_file_prefix = diags_2/beams_insitu
fields.insitu_file_prefix = diags_2/fields_insitu
plasmas.insitu_file_prefix = diags_2/plasmas_insitu

It is common to include it in a bash script

for i in 1 2 3; do
    tag=$i
    ./hipace inputs \
         hipace.file_prefix = diags_$tag \
         beams.insitu_file_prefix = diags_$tag/beams_insitu \
         fields.insitu_file_prefix = diags_$tag/fields_insitu \
         plasmas.insitu_file_prefix = diags_$tag/plasmas_insitu \
        &> output_$tag.txt
done

Would this work for you for the moment?

[cemitch99]

@MaxThevenet Thanks for the rapid reply! This looks like a good fix for now--I will give it a try. In the meantime, do you know if the characterization in the second bullet above is correct? (That is, the hdf5 data is simply overwritten after each run?)

[MaxThevenet]

Normally yes.