Problem with the Rmpi package
It is related to MPI_INITIALIZE and MPI_FINALIZE. Either one can only be called once. Hence once detach() is used, R is no longer to be a MPI aware process and cannot be a MPI process again. Hao
Blanchette, Marco wrote:
Dear all, I just started to use the snow package to send multiple jobs on our cluster using MPI and the Rpmi package as the communication method. However, the Rmpi package have been behaving strangely. When I try to detach the Rmpi package I get the following error message:
library(Rmpi) detach()
Error in dyn.unload(file.path(libpath, "libs", paste("Rmpi",
.Platform$dynlib.ext, :
dynamic/shared library
'/Users/mab/Library/R/2.8/library/Rmpi/libs/Rmpi.so' was not loaded
Following that error, the snow package seems to be unable to initiate a
new cluster, whatever method is used. The fix is to kill and restart my R
session.
Any suggestion as to what is the problem?
--
Marco Blanchette, Ph.D.
Assistant Investigator
Stowers Institute for Medical Research
1000 East 50th St.
Kansas City, MO 64110
Tel: 816-926-4071
Cell: 816-726-8419
Fax: 816-926-2018
_______________________________________________ R-sig-hpc mailing list R-sig-hpc at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-hpc
Department of Statistics & Actuarial Sciences Fax Phone#:(519)-661-3813 The University of Western Ontario Office Phone#:(519)-661-3622 London, Ontario N6A 5B7 http://www.stats.uwo.ca/faculty/yu