I currently have ~500 nested mixed-effects models to run, and one model takes ~2-3 hours to run using between-chain parallelization. However, I have access to a 64 core machine, thus I am looking to decrease the run time for my models using within-chain parallelization in brms. However, I am unsure how to set the number of threads and cores for each model. I am planning on using 4 chains, and want to maximize efficiency by reducing the run time as much as possible. Any guidance would be extremely helpful.
Asked
Active
Viewed 199 times
1
-
Please provide enough code so others can better understand or reproduce the problem. – Community Mar 05 '22 at 04:27
-
I think within-chain parallelization will be less efficient than running just more chains. So if you have 4 chains per model, fit 8 models at the same time. – Axeman Oct 05 '22 at 21:28