StartParallel: Calculate distances in parallel

View source: R/parallel.R

StartParallelR Documentation

Calculate distances in parallel

Description

Accelerate distance calculation by employing multiple CPU workers.

Usage

StartParallel(...)

SetParallel(cl)

GetParallel(cl)

StopParallel(quietly = FALSE)

Arguments

...

Parameters to pass to makeCluster().

cl

An existing cluster.

quietly

Logical; if TRUE, do not warn when no cluster was running.

Details

"TreeDist" parallelizes the calculation of tree to tree distances via the parCapply() function, using a user-defined cluster specified in options("TreeDist-cluster").

StartParallel() calls parallel::makeCluster() and tells "TreeDist" to use the created cluster.

SetParallel() tells "TreeDist" to use a pre-existing or user-specified cluster.

StopParallel() stops the current TreeDist cluster.

Value

StartParallel() and SetParallel() return the previous value of options("TreeDist-cluster").

GetParallel() returns the currently specified cluster.

StopParallel() returns TRUE if a cluster was destroyed, FALSE otherwise.

Author(s)

Martin R. Smith (martin.smith@durham.ac.uk)

Examples

if (interactive()) { # Only run in terminal
  library("TreeTools", quietly = TRUE)
  nCores <- ceiling(parallel::detectCores() / 2)
  StartParallel(nCores) # Takes a few seconds to set up processes
  GetParallel()
  ClusteringInfoDistance(as.phylo(0:6, 100))
  StopParallel() # Returns system resources
}

ms609/TreeDist documentation built on April 5, 2024, 12:07 a.m.