Meeting 24 Jan

Agenda

• Background

• What to do differently this time

• What goes to what days & teachers

• Learning outcomes

• Resources

• How to advertise?

• Next meeting

• Next meeting R-matlab-julia

Background

As discussed in a hurry in December when setting dates

• New course/workhop with 4 days of python

• APPROX

  • 24 Apr: intro to Python 24 April

    • beginners or warm-up

    • ?the other days not recommended for complete beginners?

      • Too short of time maybe to digest?

      • Or let the users decide :+1: [name=RB] <—

    • separate registration? :+1: [name=RB]

      • Advertisement for all 4 days, learners pick on registration

  • 25 Apr packages, pandas, matplotlib

    • some higher analysis later april?

  • 28-29 Apr: slurm, parallel, big data, GPU and ML/DL/AI

    • including some analysis, like seaborn and ?

  • Is this a good way?

    • [name=RB] Yes, this is fine

• Proposal

  • 1st day with intro (similar or same as Richel’s ‘NAISS Intro to Python’ course on March 4th)

    • cluster modules: it is a NAISS course :-)

    • has a login session

  • 2nd day will be about packages and running on a cluster + basic analysis without slurm!!

    • morning: packages, virt env

      • conda is used in many clusters

      • pip and possibly pixi?

      • ?containers. We’ll decide later :+1:

    • afternoon

      • analysis without slurm(?)

      • what fits in here?

      • [VOTED] IDE’s

        • [name=RB] ?Jupyter

        • [name=JY] ?Spyder

        • [name=JY] ?VScode

      • need interactive session

  • 3rd day needs a introduction to slurm and GPUs

    • slurm, batch, interactive, on-demand

    • analysis requiring more than login resources

    • (parallel), (big data), (GPU)

  • 4th day for the most advanced material

    • Anders Hast (InfraVis) comes as a guest teacher to talk about dimensionality reduction techniques. (or first day): must be before ‘Machine learning’ in the schedule, ~30-60 mins

    • Data preparation

    • (GPU), ML/DL/AI

    • ?Professional Python development, e.g. testing

• Material, timing and interactivity

  • NAISS ‘Intro to Python’ –> Day 1 ≈ 1:1 relation

  • Combine

    • Python day in 4-day course

      • more packaging

      • less analysis

      • less ML

    • HPC-python in 2 days

      • less packaging

      • more IDEs

      • more analysis

      • more ML/DL

    • much common material

      • login, run, slurm, parallel, packaging

    • –> 3 days into 3 days –> more relaxed and airy!!

      • BUT: include Anders Hast as well?

What to do differently this time

• Evaluation outcome

  • Link to Björn’s summary

    • now in github

  • more exercises

• Reductions?

• Put evaluation results on website, a fancy result can be found at this UPPMAX course

What goes to which day? And Teachers

• Earlier sessions from HPC-python (to be placed out)

  • intro (just first day?) [name=RB]

  • load (just first day?) [name=RB]

  • install packages (expanded) [name=BC]

  • Compute nodes [name=BB]

    • batch [name=BB]

    • interactive [name=B]

    • Desktop on demand [name=RP]

    • IDEs [name=BB]

  • matplotlib [name=RP]

  • GPU [name=BB]

  • Pandas [name=RP]

  • seaborn etc [name=RP]

  • parallel [name=PO]

  • big data and formats [name=BC]

  • Dimensionality reduction? [name=AH]

  • ML/DL [name=JY]

• Day 1 (copied from NAISS ‘Intro to Python’)

:warning: Does not include loading a module

• BC: could we have “loading lmod module” here?

Time	Topic
09:00-10:00	Using the Python interpreter, includes login
10:00-10:15	Break
10:15-11:00	The way of the program, includes creating and running a script
11:00-11:15	Break
11:15-12:00	Working with Python scripts
12:00-13:00	Break
13:00-14:00	Variables, expressions and statements: variables
14:00-14:15	Break
14:15-15:00	Variables, expressions and statements: operators
15:00-15:15	Break
15:15-15:45	Variables, expressions and statements: user input
15:45-16:00	Evaluation

• Day 2

  • [VOTED] IDEs

Time	Topic
09:00-10:00	.
10:00-10:15	Break
10:15-11:00	.
11:00-11:15	Break
11:15-12:00	.
12:00-13:00	Break
13:00-14:00	.
14:00-14:15	Break
14:15-15:00	.
15:00-15:15	Break
15:15-15:45	?RB: Jupyter
15:45-16:00	Evaluation

• Day 3

• Day 4

Learning outcomes

• Useful

• We have most material, BUT

• what is the amount users can learn in 3-4 days?

• balance of

  • be confident in doing

    • many exercises

  • knowing big picture and knowing where look up details

    • showing

      • demos

    • mentioning

      • talk

      • discuss

    • link to good material for plunging into

  • BC’s view: all are needed, not just one or 2

• where to put the level in this course?

  • different in different sessions?

• Learning Outcomes

  • learner-centered

  • make material from these!

Resources

• NSC for non-users

• as usual + dardel

How to advertise?

• Each day having a name describing content?

How to work?

• issues

• meetings (how often)

• matrix (phase out slack)

• combination

ToDos

• All:

  • Look at evaluation and think about changes

  • (LOs)

  • schedule

  • suggestions of more required time for sessions

Next meeting

• Fri Feb 14 11-12

Next meeting R-matlab-julia

• Fri 31 Jan 11-12