I am currently work on a research project which requires
parallel technique. I have successfully adapted an Rmpi code from the
following website. http://math.acadiau.ca/ACMMaC/Rmpi/task_pull.R and be able to
run my own code now.
However, since I am developing my own code which involves many
calculations, each time I had an error with my Rmpi code, I got an error
message like the following:
Error in inherits(x, "data.frame") : object 'mean_chain' not
found
Calls: write.table -> is.data.frame -> inherits
Execution halted
I therefore correct the error and re-run the code, the odd thing is that
in our cluster, the previous version of my R code is loaded and the new
one is just ignored, so I end up with the same error message. I tried
to use
#
R CMD BATCH --no-restore /home/sph/jma/Rmpi/ext2/ext2_pull.R to prohibit the
previous version, it does not work.
Basically, how can I clean up the cluster if I have an error in my Rmpi code?
Thank you very much and I appreciate your help!
Best Regards,
[[alternative HTML version deleted]]
