Run MCMC chains in parallel with callr

Run MCMC chains in parallel (at the same time). This runner uses the callr package to distribute your chains over a number of worker processes on the same machine. If you have used mcstate, this is the same as "worker" processes. Unless your chains take a few seconds to run, this will be slower than running with the default serial runner (monty_runner_serial), however for long running chains, the speedup will typically scale with workers added, so long as your chains divide neatly over workers.

Usage

monty_runner_callr(n_workers, progress = NULL)

Arguments

n_workers

The number of workers to use. This should be no larger than the number of chains (though this is harmless) and no larger than the total number of cores available on your computer. Ideally the number of chains you want to run is a multiple of this number (for example, if you had 8 chains, then 1, 2, 4, and 8 are good choices of n_workers, and 7 workers will likely be no faster than 4).

progress

Optional logical, indicating if we should print a progress bar while running. If NULL, we use the value of the option monty.progress if set, otherwise we show the progress bar (as it is typically wanted). Alternatively, you can provide a string indicating the progress bar type. Options are fancy (equivalent to TRUE), none (equivalent to FALSE) and simple (a very simple text-mode progress indicator designed play nicely with logging; it does not use special codes to clear the line).

Value

A runner of class monty_runner that can be passed to monty_sample()