How To: How can I run jobs run per core, instead of per node, so that the data can be split into even smaller chunks?

The default setting for the rxExec function is to run a single instance on the cluster of your function for each available node. It is also possible to set the elemType argument to cores, in which case multiple instances of FUN will be run on a node, but the number won't exceed the max number of cores available on that node.