Open bilby-bot opened 1 year ago
In GitLab by @git.ligo:colm.talbot on Aug 29, 2023, 19:50
Hi @git.ligo:alexandresebastien.goettel what version of Bilby/dynesty are you using?
In GitLab by @git.ligo:alexandresebastien.goettel on Aug 30, 2023, 08:53
Hi Colm, thanks! I'm using the 2.1.1 release Bilby version (26c84edf84fec6f3f3eb995d9f06257ad4eef180). "pip show dynesty" outputs Version 2.0.1.
In GitLab by @git.ligo:colm.talbot on Aug 30, 2023, 15:00
Thanks, please can you paste more of the log, I think this part of the trace is not informative, it looks like the job is trying to continue after being terminated.
Also what are the sampler kwargs you are using? Also, if there is anything else non-standard in your environment/job please say.
In GitLab by @git.ligo:alexandresebastien.goettel on Aug 31, 2023, 13:30
So I realised that the log itself contains many "checkpoint and exit on 77" followed by restarts (condor periodic restart), but other than that looks normal.
Should be worth mentioning that I am not running on an igwn cluster but on a separate machine. I am loading conda from cvmfs though and have installed bilby using "python setup.py install" (originally so I could modify some bilby code, but this install is clean out of master).
Here are my sampler settings:
{'nlive': 2048, 'naccept': 60, 'check_point_plot': True, 'check_point_delta_t': 1800, 'print_method': 'interval-60', 'sample': 'act-walk', 'bound': 'live', 'maxmcmc': 5000, 'nact': 4, 'verbose': True, 'npool': 17}
In GitLab by @git.ligo:alexandresebastien.goettel on Aug 28, 2023, 14:12
Some of my runs are running into an issue in the middle of sampling, crashing the entire thing: ---
The full traceback is below: