I normally do work on a cluster computer. I have a few very large Rscripts that I use to run data. The full Rscript takes about 5 hours to run. To do this, I use the cluster by submitting a job with qsub. This allows me to leave my computer and wait for the cluster to run. However, I want to run this particular Rscript with minor changes 20 separate times.
What I have been doing is waiting for the cluster job to finish. Then going into the Rscript and tinkering with the file names, and then sending another job to the cluster.
My question is this: After I submit the job to the cluster, can I go ahead and tinker with the Rscript file while the cluster is computing (i.e. does the job send a copy of everything to the cluster at the time of submission)? OR do I need to wait for the job to finish before I can tinker with the Rscript (the cluster accesses my Rscript as needed throughout the job)?
I would ideally like scenario 1 to be the case because then I could send in many jobs one after the other after I make small adjustments, but I don't know how it works?
That depends on your cluster and how it is set up. The best thing you could do is save another instance of the script to a different name and play to your heart's content.