ClimDown (version 1.0.8)

parallelization: Parallelization

Description

ClimDown uses the foreach package to support parallelization where possible. In general, most of ClimDown's computations contain an outer I/O chunk loop and an inner computational loop. The outer chunk loop serially reads in as much data from input files as it feasibly can, and then the inner loop will execute in parallel if the user has configured a parallel engine.

Users can configure a parallel engine before execution using either the doParallel or doMPI packages.

If the user does not configure a parallel backend, the inner loops will simply run serially and issue a warning.

See Also

doParallel and doMPI

Examples

# NOT RUN {
library(doParallel)
registerDoParallel(cores=4)
bccaq.netcdf.wrapper(...)
stopImplicitCluster()
# }
# NOT RUN {
# }