I am using doFuture, future.batchtools, and the foreach function to parallelize some code over 128 CPUs on one node of a HPC that runs on a PBS scheduler. What I would like my code to do is send out one batch script that allocates my tasks over those 128 CPUs. What my code does instead is send out one batch script per task that sets all 128 CPUs on the task, and from what I can see in the logs in ~/.future, it is sending out incorrectly formatted PBS scripts at that.