Peregrine cluster fails to produce an incipient species tree with 1G of memory

[richelbilderbeek]

Zooming in on the results of the shorter article run, I've found when the Peregrine cluster fails to produce an incipient species tree in an hour:

This is when Speciation Initiation Rate is at the highest and when Extinction Rate is at the highest (yes, at the highest). This appears to be when the PBD::pbd_sim function needs more time to finish in creating a phylogeny. I can safely give this process more time (say, 10 hours instead of 1), as these incipient species trees will not clog up BEAST2.

So, I will set the time for scripts/add_species_trees.sh to 10 hours.

[richelbilderbeek]

This has not been a fix: the problem is not in the creation of species trees scripts/add_species_trees.sh, but in add_pbd_output.sh

[...]
/var/spool/slurmd/job7190238/slurm_script: line 10: 74568 Killed                  Rscript -e "library(Cer2016); add_pbd_output(\"$1\")"
[...]
slurmstepd: Exceeded step memory limit at some point.

I just need to add more memory to scripts/add_pbd_output.sh

[richelbilderbeek]

scripts/add_pbd_output.sh can now use 10G instead of 1G scripts/add_species_trees.sh.sh now uses 10 minutes instead of 10 hours.

[richelbilderbeek]

cat collected_n_taxa.csv | less

resulted in a CSV without NAs, so 10G appeared to be enough.