Rhpc (version 0.14-71)

Rhpc-package: A package for High-Performance Computing

Description

Rhpc_lapply, Rhpc_lapplyLB and Rhpc_worker_call using MPI provides better HPC environment on R(works fast on HPC). Maybe start from the Rhpc batch command is better(faster). Because strong mutual concessions(slowly) often occur by sched_yield, we recommend the use with the batch mode...XD

Rhpc batch in package directory.

Usage

Rhpc_initialize()
Rhpc_finalize()
Rhpc_getHandle(procs)
Rhpc_worker_call(cl, FUN, ...)
Rhpc_worker_shy(cl, FUN, ...)
Rhpc_lapply(cl, X, FUN, ...)
Rhpc_lapplyLB(cl, X, FUN, ...)
Rhpc_numberOfWorker(cl)
Rhpc_Export(cl, variableNames, pos, envir)
Rhpc_EvalQ(cl, expr, envir)
Rhpc_setupRNG(cl, iseed)

Arguments

cl

external pointer to MPI communicator

procs

number of process, void if Rhpc

X

vector or list

FUN

string of function name or function object

...

arguments

variableNames

string vector

iseed

random number seed

expr

expression to evaluate

pos

default is current environment position

envir

where to look for the object

Details

http://prs.ism.ac.jp/~nakama/Rhpc

Examples

# use Rhpc command
Rhpc_initialize()
cl <- Rhpc_getHandle() # ommit number of worker if use Rhpc batch

# Rhpc set to options
opstr=list("Rhpc.mpi.rank","Rhpc.mpi.procs","Rhpc.mpi.c.comm","Rhpc.mpi.f.comm")
do.call("options",opstr)
Rhpc_worker_call(cl, "do.call","options", opstr)
# warning! : pointer not export, worker Rhpc.mpi.c.comm is (nil) on master.

Rhpc_worker_call(cl, Sys.getpid)
Rhpc_lapply(cl,1:10000,sqrt)
Rhpc_lapplyLB(cl,1:10000,sqrt)
Rhpc_finalize()
#
# command line example
#
# mpirun -n 4 Rhpc CMD BATCH --no-save `script file for exemplified above'
#