I want to repeat this question using python. Reason is I have access to 10 nodes in a cluster and each node is not identical. They range in performance and I want to find which is the best computer to use remotely based on memory and cpu-speed/cores available.
EDIT: Heck, even just a command line interface would be useful. Any quick and dirty solutions?
Take a look at the SIGAR library which has an extensive API for collecting system data cross-platform. It also has libraries available in many languages (Python, Java, Erlang, Ruby, etc).