This may not be a Stan question but an R question, but when running MCMC on multiple models sequentially, is there a way to have the processors that are not being used go to the next model?
For example, let’s say I have models M1, M2, and M3 and am running MCMC sequentially. Some chains in M1 sample faster than others. Can I use the idling processors that finished sampling a chain in M1 proceed to sampling from M2?