search

COSMIC-PopSynth / COSMIC

COSMIC (Compact Object Synthesis and Monte Carlo Investigation Code)
GNU General Public License v3.0
47 stars 59 forks source link

Issues with generating black holes with cosmicpop at near-Solar metallicity #487

Closed michaelzevin closed 3 years ago

michaelzevin commented 3 years ago

There are a couple issues I've had in some runs, I'll try to list them logically below. The population run directory is on Quest at /projects/b1095/michaelzevin/cosmic/populations/acc_efficiency/acclim_1/0.017

  1. When I ran this population using the multidim sampler, it wasn't finding any black holes. This seems odd to me, but could maybe be due to the initial conditions? So I ran with the independent sampler using q_min=0.1. However, I hope this will instead get fixed from addressing the issues below, since in versions of COSMIC from last summer I was still getting BBHs at Solar metallicity.

  2. When rerunning using the independent sampler, the model generated 54 matching BBH systems on the first step (which I saw in the log). However, it did not acquire any more systems in the rest of the run, which had a wall time of 5 days. I've attached the .dat file with these 54 systems.

  3. When the run hit its wall time of 5 days, rather than writing this in the .out file, it seems to have overwritten the log and .out files completely (i.e., they're empty).

dat_kstar1_14_kstar2_14_SFstart_13700.0_SFduration_0.0_metallicity_0.017.h5.zip

katiebreivik commented 3 years ago

This was happening because cosmic-pop had a few continue statements which bumped the code out of the loop before Nstep was updated. This means that the same seed was run over and over forever and failed to produce any new systems. This has been fixed in PR: #493