Memuse bug with hydro_sort
Doing some tests to see where the increased memory over time with hydro runs comes from.
Running the code with memory use reports (note however I'm only writing the summary at the end as the files are too big otherwise).
Bit of a strange output after a certain point.
See e.g, /cosma7/data/dp004/rttw52/swift_runs/runs/TABULA/50BoxTest/TABULA_L0050N0752_MemJMal/memuse_report-rank0-step149.dat
The earlier steps looks ok in their reporting.