All the LLNL computing tutorials are great---they tend to be clearly written, with good examples, and cover a wide swath of use cases. The OpenMP https://computing.llnl.gov/tutorials/openMP/ and MPI https://computing.llnl.gov/tutorials/mpi/ are fantastic (when I google "mpi tutorial", the LLNL one is the top hit). Blaise is extremely knowledgable, but probably doesn't have time to respond to questions (unless you're a lab employee :P ).

Indeed, these tutorials were essentially my parallel processing class in college. We didn't have a mandatory textbook -- most of the content was in these tutorials.

I was recently a postdoc at LLNL, and took a real-life 1-week in-person crash-course on pthreads, OpenMP and MPI from Blaise. Afterwards I really felt like I could get going.

I guess it is not SUPER surprising---LLNL and LANL are basically where high performance computing was invented, and they have a tradition of excellence. I have accounts at a few other supercomputing facilities, and none compare in expertise and professionalism to LLNL.