Running simulations in parallel

Once you successfully create a simulation specification and branch file it is time to use the distributed runner. Recall, we ran a single simulation with a model specification file in this way

simulate run /path/to/your/model/specification

Very similar to this, vivarium-cluster-tools includes a command for simulating in parallel

psimulate run /path/to/your/model/specification  /path/to/your/branch

By default, output will be saved in /share/costeffectiveness/results. If you want to save the results somewhere else you can specify your output directory as an optional argument

psimulate run /path/to/your/model/specification /path/to/your/branch -o /path/to/output

Another optional argument is the cluster project under which to run the simulations. By default, the cluster project used is proj_cost_effect. To use a different project, specify it with the -P flag

psimulate run /path/to/your/model/specification /path/to/your/branch -P proj_csu

Currently, the projects that simulation science has access to are proj_cost_effect, proj_cost_effect_diarrhea, proj_cost_effect_dcpn, proj_cost_effect_conic, and proj_csu. Only these projects may be used.

If your psimulate run has failed to complete you can restart the failed jobs by specifying which output directory includes the partially completed jobs using restart

psimulate restart /path/to/the/previous/results/

For psimulate restart you can also choose a project with optional flag -P.

If you wish to expand a previous psimulate run by adding additional input draws and/or random seeds, you can do so using expand.

psimulate expand /path/to/the/previous/results/ --add-draws 10 --add-seeds 5

You can use one or both of --add-draws and --add-seeds to expand your simulation. Any previous results will not be overwritten, but any additional simulations resulting from the new input draws and/or random seeds will be run.

psimulate expand also supports choosing a project via the option flag -P.