The scripting method is one device one line at a time. I have always felt it was slow and do not use it. If I had to make a recommendation, use PyExpect or Perl, or Expect with fork enabled in TCL. Any one of these will fork the pid's and spawn children and hence run tasks across multiple devices concurrently.
In production though it's best to err on the side of caution, I have been scripting for years and the learning curve is steep unless you have a very strong computer science background.