excerpt from a evaluation feedback regarding the first run of this workshop in March 2017
I think the scope of the course was too much. I would suggest breaking it into two. You could treat the section on parallel computation as an entirely different course. I don’t think you should have an introduction to "ls" and creating directories in UNIX on the same day as writing and executing a parallel code written in python for MPI. Anyone ready for MPI knowns basic UNIX and anybody learning UNIX needs more practice before worrying about MPI. I think you could make a very practical and helpful course that avoids parallel coding. There is a lot that people want to do on HPC in serial. Even for people that are competent, it is nice to do some revision on basics and to pick up new tricks. As for the MPI, I have already done some courses (I am by no means proficient) and I really don’t feel that I got too much out of this part of the course, like additional trick or revisions etc, simply because there was not enough time. I think a lot of new users will be very grateful for a course that sets them up to run jobs, work in the UNIX environment and understand outputs etc. The is time for more examples and even in course exercises. Maybe half an hour discussion of what parallelisation can offer. I strongly recommend removing the section on parallelisation and, in an ideal world, producing a second course (which could include OpenMP, probably even more helpful for beginners). But hey, that’s just my opinion. What do I know. Otherwise, great course. Thanks for having me.
excerpt from a evaluation feedback regarding the first run of this workshop in March 2017