Hi,
I tried to merge 2 files :: 11Go and 8,5Go with run_telfile.py merge.
$ run_telfile.py merge --kind=time reestuarisation_part_p1.res reestuarisation_part_p2.res test.res
Interpreting command line options
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
~> Merging reestuarisation_part_p1.res, reestuarisation_part_p2.res into test.res kind=time
Skipping duplicate time 6566400.0 from reestuarisation_part_p2.res
Killed
I ran this on a machine with 32Go of RAM, it crashes.
I retried on 128Go RAM machine and it uses at maximum 75Go of RAM.
This is huge.
I think run_telfile.py merge could be improved by reading/writing by chunk (few timestep at a time) instead of first reading all files, store them in RAM and finally write to disk.