cluster parallelization for R scripts

parallel_R.R

library(Rmpi)
 
## load the packages we'll need
RLIBS="~/R/x86_64-redhat-linux-gnu-library/2.13"
.libPaths(c(RLIBS, .libPaths()))
 
### Direct RMPI way:
mpi.spawn.Rslaves(nslaves=15)
slavefn <- function() { print(paste("Hello from", foldNumber)) }
mpi.bcast.cmd(foldNumber <- mpi.comm.rank())
mpi.bcast.Robj2slave(slavefn)
result <- mpi.remote.exec(slavefn())
print(result)
mpi.close.Rslaves()
 
library(snow)
## snow method
cluster <- makeCluster(16, type="MPI")
clusterEvalQ(cluster, library(utils)) # load a library
clusterExport(cluster, ls()) # export everything
out <- parSapply(cluster, 1:16, function(x) print(paste("snow hello from ", x)))
print(out)
stopCluster(cluster)
 
## SNOWFALL method
library(snowfall)
# default
sfInit( parallel=TRUE, cpus=16)
sfExportAll()
sfLibrary(utils)
out <- sfSapply(1:16, function(x) print(paste("snow hello from ", x)))
print(out)
sfStop()
 
# snowfall MPI
sfInit( parallel=TRUE, cpus=16, type="MPI" )
sfExportAll()
sfLibrary(utils)
out <- sfSapply(1:16, function(x) print(paste("snow hello from ", x)))
print(out)
sfStop()


# snow's close command, shuts down and quits from script 
mpi.quit(save = "no")