Hi all,
I use mpirun to execute an R script with MPI parallelization, using the snowfall package.
I set upt the parallelization by
'''
sfInit(parallel = TRUE)
'''
then I have
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
*** and potentially your MPI job)
[red075.cluster.local:190363] Local abort before MPI_INIT completed completed successfully, but am not able to aggregate error messages, and not able to guarantee that all other processes were killed!
when I launch the script with mpirun and add '''sfInit(parallel = TRUE, type = "MPI")'' in the script, I have the error
Error in makeMPIcluster(.sfOption$nodes, outfile = tmp, homogenous = TRUE, :
MPI cluster of size 39 already running
Error in sfInit(parallel = TRUE, type = "MPI") :
Starting of snow cluster failed! Error in makeMPIcluster(.sfOption$nodes, outfile = tmp, homogenous = TRUE, :
MPI cluster of size 39 already running
Error in makeMPIcluster(.sfOption$nodes, outfile = tmp, homogenous = TRUE, :
MPI cluster of size 39 already running
Execution halted
I tried not to use sfInit but the code will be run in a sequential mode instead of parallel.
'''
sfSource("CopulaFunctions.R")
'''
then I have :
Calling a snowfall function without calling 'sfInit' first or after sfStop().
'sfInit()' is called now.
snowfall 1.84-6.3 initialized: sequential execution, one CPU.
Any advise how to deal with that? Thank you