Notes from the Parallel computation course – 2026-10-22

Events: Intro about the participant  Schedule/agenda  Stage setting (background/basic check)  Terminology  Motivation for topic of the day  Hand calc. of speedup  connection with Amdahl's law ….

Obs:

• Too much personal attention??? o Is it because there’s only 1 student?
• Space to think?
• The picture showing the process flow might be a bit hard to infer, especially when you are put on the spot
• The content is well-designed to serve as self-study as well o Could some parts of this be designed as (pre-session) self-study so that the TLO can be used to focus on more advanced topics/exercises?
• Answer to the exercise exists right there…. Experiences of resisting urge to just look at the answer?
• Example from cosmos – nice to see real world scenario in contrast to theoretical calculation
• Perhaps touching up on how to check the efficiency/parallelizability of a code might be useful or could act as a good continuation for this course? Current impression (me): it is just not possible to determine how much or which parts of the code can be parallelized
• Very practical in terms of how to get things running on the cluster, not so much on the deeper inner workings – but perhaps this is not the goal of this course either – should that be stated/distinguished more clearly in the objectives? (maybe I just missed it)
• Exercises are too tiny?
• Why this HTML format for the content?
• Too much emphasis on the 5 min. late part?
• Limitations of Zoom screen sharing mode of presentation?