search

ekirving / qpbrute

Heuristic search algorithm for fitting qpGraph models
MIT License
9 stars 3 forks source link

C stack usage 7970016 is too close to the limit #4

Closed angelesdecara closed 5 years ago

angelesdecara commented 5 years ago

Dear Evan,

A little hand accidentally closed my laptop lid, and when I try to restart the calculation of the Bayes factors, I get this message:

INFO: There are 784 graphs to compute Bayes factors for. INFO: Bayes factor done for graph 204a55b Traceback (most recent call last): File "qpbayes.py", line 223, in calculate_bayes_factors(argv.geno, argv.snp, argv.ind, argv.prefix, argv.pops, argv.out, nthreads=argv.threads) File "qpbayes.py", line 201, in calculate_bayes_factors qpb.calculate_bayes_factors() File "qpbayes.py", line 154, in calculate_bayes_factors self.model_likelihood(graph) File "qpbayes.py", line 172, in model_likelihood MCMC_NUM_BURN], env={'OMP_NUM_THREADS': '1'}, stdout=open(log_file, 'w')) File "/Users/angeles/Downloads/qpbrute/utils.py", line 50, in run_cmd raise RuntimeError(err) RuntimeError: Registered S3 methods overwritten by 'ggplot2': method from [.quosures rlang c.quosures rlang print.quosures rlang Something went wrong, trying again. Error: C stack usage 7970016 is too close to the limit Execution halted

Any clue what's going on? Many thanks,

Angeles

ekirving commented 5 years ago

That error is caused by R trying to access more system resources than your operating system allows by default.

This answer on Stack Overflow shows how you can permanently increase the ulimit in OS X https://unix.stackexchange.com/questions/108174/how-to-persistently-control-maximum-system-resource-consumption-on-mac#answer-221988

However, it might make more sense to run the Bayes Factor calculations on a linux server, rather than on your laptop. The code will run to completion on OS X (provided systems limits are increased), but for non-trivial datasets each model can take > 1 week to complete.