Second debriefing meeting May 21 at 14

Overall thoughts

  • BC

    • now there are new courses, relevant as prereqs
      • (enough to understand the material)
      • NAISS modules course
      • NAISS Slurm course
      • (linux 101)
      • (courses with login session)
      • Q: Should we keep having an optional login session?
        • [DECIDED] Yes: RB, RP, BC, BB
        • No:
      • Q: Should we have other courses as prequisites?
        • Yes:
        • No: RB
        • [DECIDED] No, but link to it from the prerequisites: RB, BC, LS
      • Q: Should we add fetching the tarball in the login (and preparation) session?
        • Yes: BC, PO
        • [DECIDED] No: BB (it would take away course time for people that do not do this, they will not read/do it, will give me more admin), LS, RP
      • Q: Should we have the Zoom link in the tarball?
        • Yes:
        • [DECIDED] No: RB, BB, BC, RP, LS
    • suggestions
      • cover HPC-specific topics (e.g. the module system, the job scheduler) enough to be confident and comfortable in doing the exercises
        • [BC] Knowing HPC clusters minimally is in the prereqs
        • [RB] Great idea! Let’s be focused!
      • [ ] Let prerequisites point more to other courses (e.g. the module system, the job scheduler)
        • [DECIDED, BC seems willing to do this] Yes: BC, BB, LS
        • No:
      • theory minimized, just get started!
        • like:
          • train writing a batch script
        • [Richel] Great idea! We know that lecturing affects learning negatively, based on multiple meta-analyses [Reference 3]
      • focus on specific needs for the specific language
      • cluster understanding in preparations

  • Evaluation results, with fancy graphs:

  • [Richel] I think following the pedagogic literature has a bigger impact than reading/discussing the evaluations and/or following our instincts [Reference 1]

    • BB: the evaluations are more important because they are from our actual participants
      • Yes, we sh
      • They also tell us some valuable things that we can agree on
    • LS: literature discussing university students, from anywhere in the world, at least from the second year onward should be applicable to PhD students
      • RB: Even though there are differences, I still feel following these guidelines is superior over following our instincts
    • PO: the literature referred to is about scholars
      • RB: This is false, the reference I cite is from cognitive science
      • RB: could you refer a paper to me?
    • PO: if you only focus on numbers, you should focus on the material
    • BC: If we can agree on that our courses are towards tutorial types more than information heavy, we should hav a focus on exercises. But of course having some WHYs and some intro to make good decisions -
  • [Richel] I think the camera recordings do more harm than good, as it kills interactivity during class, at the minimal gain of people outside of class being able to see these recordings [Reference 2].
  • [Richel] I think we should convert the GitHub repo of this course to MkDocs, over waiting for Erik Lindahl to pick an MkDocs theme
    • Yes: BC, RB
    • No:
  • [Richel] I think we should pick an MkDocs theme, over waiting for Erik Lindahl to do so
    • Yes: RB
    • No:

  • [Richel] I think we should pick the MkDocs theme most of us like:

  • Course dates

    • 6-8 October can we split or to late?
      • [Richel] Let’s keep it: we picked these for a reason
    • [Richel] Can we select an order?
      • R, MATLAB, Julia (like last time): RB
      • [other order]:

References

  • [Reference 1] For example ‘long lectures and presentations will fail because students stop viewing and listening after about six minutes’ [Nilson, Linda B., and Ludwika A. Goodson. Online teaching at its best: Merging instructional design with teaching and learning research. John Wiley & Sons, 2021]. From the same book, the 25 principes of learning from cognitive science (chapter 4, page 92) gives some great pointers, e.g. ‘2. Students learn […] better […] by engaging in an activity than when they passively watch […] an instructor’, ‘6. Some qualities […] help students learn […] better […]: human faces, […] instructor enthusiasm’.
  • [Reference 2] Using them as a supplementary tool may not be worth it in our settings, as ‘No performance advantage was found for viewing recordings despite students’ positive perception of their utility.’ [Lesser, Melissa, and Lora Packel. “Educational Lessons from a Pandemic: Lecture Recording and Learning Outcomes.” Internet Journal of Allied Health Sciences and Practice 22.2 (2024): 9., Link to paper] or ‘The main finding is that students using the online lectures as a substitute for attending lectures are ultimately at a fairly severe disadvantage in terms of their final marks’ [Williams, Andrew, Elisa Birch, and Phil Hancock. “The impact of online lecture recordings on student performance.” Australasian Journal of Educational Technology 28.2 (2012), Link to paper]. I found most experimental papers do not compare with live recording (it is usual pre-recorded versus an unrecorded/live lesson). The exception is a paper that does recommend the recording live webinars, although it admits that ‘[Webinars tend] to be less engaging and less interactive for the participant’ [Dailey-Hebert, Amber. “Maximizing interactivity in online learning: Moving beyond discussion boards.” Journal of Educators Online 15.3 (2018): n3., paper].
  • Reference 3 From [Hattie, John. Visible learning: The sequel: A synthesis of over 2,100 meta-analyses relating to achievement. Routledge, 2023, page 363]: the effect size of lecturing is -0.26 with a robustness index of 4 out of 5 and is based on 3 meta analyses using 273 studies using 27,296 people, measuring for 614 effects with a standard error of 0.08. One example open access study is [Knight, Jennifer K., and William B. Wood. “Teaching more by lecturing less.” Cell biology education 4.4 (2005): 298-310]

R

Comments by us

  • [Richel] The course is an introduction course. I feel the start-up parts are given too little time, to allow for things that I do not consider being beginner topics. I feel beginner things need/deserve their time, instead over being sacrificed it for more sexy topics.

    • [ ] Suggest to remove ‘Machine learning’
      • [Björn] more and more demand for this (right?)
        • [Richel] This may be, but should it come at a cost of the foundational topics? Putting it in an intermediate course could be a possible fix
    • [ ] Suggest to remove ‘Parallel’
      • [Björn] although not that common, this is on of the cores
        • [Richel] There is no other session that build on this one, hence I would not call it a foundational/beginner topic. Putting it in an intermediate course could be a possible fix
    • [ ] Suggest to remove ‘Isolated environments’
      • [Björn] To my knowledge this is not use extensively, although for reproducibility it maybe should (I don’t know the community well)
        • [Richel] This may be, but should it come at a cost of the foundational topics? Putting it in an intermediate course could be a possible fix

Given the extra time freed up by removing 2 out of the 3 topics above, below is a suggested schedule that would fit better. At 15:15 the 1 topic not removed can take place

```text
+-------+------------------------------+--------------+
| Time  | Topic                        | Teacher(s)   |
+=======+==============================+==============+
| 9:00  | (optional) First login       | BB + PO + RB |
+-------+------------------------------+--------------+
| 9:45  | Break                        | .            |
+-------+------------------------------+--------------+
| 10:00 | Introduction                 | RB           |
+-------+------------------------------+--------------+
| 10:10 | Syllabus                     | RB           |
+-------+------------------------------+--------------+
| 10:20 | Load modules and run         | RB           |
+-------+------------------------------+--------------+
| 11:00 | Break                        | .            |
+-------+------------------------------+--------------+
| 11:15 | Packages                     | BB           |
+-------+------------------------------+--------------+
| 12:00 | Lunch                        | .            |
+-------+------------------------------+--------------+
| 13:00 | Batch                        | BB           |
+-------+------------------------------+--------------+
| 14:00 | Break                        | .            |
+-------+------------------------------+--------------+
| 14:15 | Simultaneous session         | .            |
+-------+------------------------------+--------------+
| .     | HPC2N: ThinLinc, RStudio     | PO           |
+-------+------------------------------+--------------+
| .     | LUNARC: On-Demand, RStudio   | RP           |
+-------+------------------------------+--------------+
| .     | UPPMAX: Interactive, RStudio | RB           |
+-------+------------------------------+--------------+
| 15:00 | Break                        | .            |
+-------+------------------------------+--------------+
| 15:15 | To be decided by vote        | ?            |
+-------+------------------------------+--------------+
| 15:45 | Summary and evaluation       | RB           |
+-------+------------------------------+--------------+
| 16:00 | Done                         | .            |
+-------+------------------------------+--------------+
```

Acting points and responsible

MATLAB

  • Evaluations

  • Summary

    • 1 did not recommend
    • missing 5s
      • jupyter
      • GPU
      • client on desktop (is this the UPPMAX local)
      • parallel
    • large divergence in confidence
      • GPUS
      • view add-ons
      • parallel
    • no nead for real-time reading
    • future topics: 3rd party tools

  • Comments from us:

    • BC:
      • restructure the 2 batch sessions
      • more info-hiding (stuff they can cover themselves)

Schedule

  • is GPU removed?
Start Timing Topic Teacher
9:00 Log-in
9.45 Break
10:00 Syllabus Rebecca
10:15 Matlab in general Rebecca
10:25 Load modules and run + script Rebecca + person from each HPC center
10:50 Break .
11:05 Slurm job scheduler and Matlab Birgitte
12:00 LUNCH .
13:00 Matlab GUI and Slurm Pedro & Rebecca
13:30 Parallel and multithreaded functions 45 Pedro
13:50 Break .
14:00 Parallel and multithreaded functions 14:30
14:50 Break .
15:05 Simultaneous SESSIONS - desktop on demand (COSMOS) 30m Rebecca
- Simultaneous SESSIONS - local desktop (UPPMAX) 30 m Björn
- 30 m Simultaneous SESSIONS - (jupyter notebook) Pedro
15:35 interaction with other tools (skip this year) ?
15:35 Conclusion All
15:40 Evaluation .
15.50 Q/A
16:00 END

Acting points and responsible

  • Working group to meet when?

Julia

  • Evaluations

  • Summary

    • at least one puts a 5 all the time!
    • large divergence in confidence
      • install julia package
      • envs
      • interactive
      • batch script
    • good teaching
    • asks for ML and more GPU hands-on

  • Comments from us:

    • BC:
      • more info-hiding (stuff they can cover themselves)

Schedule

Start Timings Topic Teacher
9.00 45 Log-in help + tarball?
9.45 Break
10.00 15 Syllabus (exercises) Björn
10.15 10 Intro, Julia in general Pedro
10.25 30 Load modules and run Björn
10:55 Break .
11:10 50 Packages and isolated Pedro (and Björn)
12:00 LUNCH .
13.00 50 Batch and GPU 46 Pedro
13.50 Break .
14:05 45 Parallel+multithreaded functions 30 Pedro
14.50 Break .
15.00 25 UPPMAX: Jupyter and interactive 25 Björn
. 25 HPC2N: Jupyter Pedro
15.25 Conclusion and summary, Q/A Björn
15.30 Evaluation .
15.40 Q/A .
16.00 END .

Acting points and responsible

  • Working group to meet when?

Debriefing meeting April 10 at c. 15.40