R

From HPC users
Revision as of 12:51, 9 September 2013 by Albensoeder (talk | contribs)
Jump to navigationJump to search

There are two R releases available on FLOW and HERO. The actual relase is available by

 module load r/3.0.1

With this command the corresponding Intel compilers and Intel MPI will be automatically loaded.

Note: The release 2.15.2 is deprecated and only for legacy on the system.


Additional installed packages

The actual release contains a lot of additional R packages, for example

  chron, abind, ade4, akima, bitops, caTools, circular, doMC, doMPI, ensembleBMA, foreach, gdata, geosphere,
  gplots, gtools, hht, iterators, lattice, latticeExtra, ncdf, raster, RColorBrewer, RCurl, RMpi, signal, solaR,
  stringr, timeDate, truncnorm, VGAM, zoo

and many more...

Usage of R and MPI

For parallelization the packages doMPI and Rmpi are installed. To launch an parallel R script inside an SGE script please use command line

  mpirun -bootstrap sge -np $NSLOTS -env I_MPI_FABRICS shm:ofa R --slave -f SCRIPTNAME SCRIPT_CMDLINE_OPTIONS

to enable SGE to control all processes of your script. Please do not use the batch starting sequence R CMD BATCH.

The corresponding parallel environment in the SGE submission script is specified by

 #$ -pe impi41 NUMBER_OF_CORES
 #$ -R y

Note (only!) for Rmpi

  • Before you start R with the mpirun command you have to set the environment variable R_PROFILE in your SGE-Script. Otherwise the MPI processes were not spawned. Please add following line to your SGE-script
  export R_PROFILE="/cm/shared/apps/r/3.0.1/intel/64/2013.0.028/lib64/R/library/Rmpi/Rprofile"
  • The MPI processes were spawned by the mpirun command. The Rmpi command mpi.spawn.Rslaves() is not necessary and should not be used within the script!


External links