run_task: Run a task

Description Usage Arguments

Description

Run a task. This is designed to run locally, rather than being run from a cluster.

Usage

1
2
3
run_task(experiment, task, id = NULL, parallel = TRUE, ...,
  dry_run = FALSE, backup = FALSE, capture_output = NULL,
  preschedule = FALSE)

Arguments

experiment

Name of the experiment

task

Name of the task

id

Id of the jobs to run. If omitted, all ids for this experiment will be run.

parallel

Run in parallel using parallel::mclapply?

...

Arguments passed through to mclapply.

dry_run

Logical, indicating if we should actually run things.

backup

Logical, backup any existing output

capture_output

Should we capture output? By default this happens if running noninteractively, or if more than one id is used.

preschedule

If running in parallel, should we preallocate jobs? If your jobs are at all computationally intensive, setting this to FALSE (the default) will allow better use of resources. For very quick jobs, the overhead involved here may outweigh savings, but that's unlikely if you're using experimentr...


traitecoevo/experimentr documentation built on May 31, 2019, 7:42 p.m.