mpiInit {ADaCGH}R Documentation

Initialized a cluster of workstations

Description

Use Rmpi and papply, and set a cluster of workstations ready to run ADaCGH.

Usage

mpiInit(wdir = getwd(), minUniverseSize = 15, universeSize = NULL,
        exit_on_fail = FALSE)

Arguments

wdir The common —e.g., NFS mounted— directory. We need a common directory for the graphics, so that they are all found in the same location.
minUniverseSize The minimal LAM/MPI universe for the function to return successfully. If the function determines that the available number of slaves is smaller than minUniverseSize it will fail (if exit_on_fail = TRUE) or give a warning.
universeSize Alternatively, you can specify the actual size of the LAM/MPI universe. If this many slaves cannot be started, the funcion will fail.
exit_on_fail If TRUE, kills R session if it cannot run succesfully.

Details

This function calls Rmpi and related functions to start a cluster of workstations, initializes the random number generator, and calls the required libraries on the slave nodes.

Value

This function is used to create a cluster.

Note

We have a somewhat similar function in basicClusterInit.Rd, but that is more sophisticated.

Author(s)

Ramon Diaz-Uriarte rdiaz02@gmail.com

References

Diaz-Uriarte, R. and Rueda, O.M. (2006). ADaCGH: an R package and web-based application for the analysis of aCGH data. Tech. report. http://www.ligarto.org/rdiaz/Papers/adacgh.pdf, http://adacgh.bioinfo.cnio.es.

Examples

## Not run: 
mpiInit()
## End(Not run)

[Package ADaCGH version 1.3-10 Index]