Add a wrap-up episode

[tobyhodges]

Comparing what was merged in #4 with the outline described in #1, I think only one final episode is missing where the plotted results of the various Amdahl jobs run in the previous episode could be discussed.

Given that this would also be the last episode in the standard workshop you have planned, I suggest that it might also be a good place to pull some threads back together and reiterate the important concepts (nodes, jobs, parallelisation, I/O(?)) and key takeaways you hope to get across in an HPC Carpentry workshop. E.g. don't run things on the submission node; be a considerate user of shared resources; don't run things on the submission node; key resources to think about for cluster jobs are time, memory, and threads/cores; try to make a good estimate of these resource requirements for your job before you submit it; and don't run things on the submission node.