Simulations, including Monte Carlo simulations, can often take a long time to run. This makes them ideal candidates for background jobs. In this example, we estimate Pi using a Monte Carlo simulation.
With estimate-pi.R open, you can either click on the Jobs
pane and then Start Local Job or click on Source, then Source as Local
Job... or "Source as Launcher Job...". In this example, the job periodically
saves out results. This pattern works for both local and remote jobs, since
remote jobs have access to the specified working directory. If job scripts are
saving data, it is best practice to have each script save data to a unique
output file. This prevents concurrent scripts from attempting to write to the
same file.
Since the job periodically saves results, you can interactively load and work
with those results in the original R session. pi-results.R
provides an example of this pattern.
