Reflection¶

Author: Richel
Course: Intermediate Bianca
Date: 2024-11-11

Course schedule¶

When	Who	What
9:00	R	Introduction
9:10	R	Prevaluation
9:20	P	Transferring files p1
10:00	.	Break
10:15	P	Transferring files p2
10:45	R	Efficient jobs
11:00	.	Break
11:15	R	Efficient jobs
11:20	R	Complex jobs
11:40	R	Replicate jobs
12:00	.	Lunch
13:00	B	Software and packages installation p1
13:50	.	Break
14:00	B	Software and packages installation p2
14:40	BR	[IDEs on Bianca p1
14:50	.	Break
15:05	L	IDEs on Bianca p2
15:40	R	Summary
15:45	R	Evaluation
15:45	All	Optional Q/A
16:00	.	END

Who: B: Björn, P: Pavlin, R: Richèl

Reflection during the day¶

[x] [ADDED TO COURSE MATERIAL] Add more arguments for rsync, that differ from FileZilla:
- can transfer data from everywhere to everywhere, especially transfer sensitive data between HPC center
- (ignore) can transfer (soft) symbolic links

Method	Can transfer big files?	Can transfer between any HPC center?	Can transfer soft links?	Keeps file attributes?	Can do incremental file transfer?	Can send deletions?
`rsync`	Yes	Yes	Yes	Yes	Yes	Yes
FileZilla	?Yes	No	?No (P)	?	Yes	?
scp	?Yes	?Yes	?No (P)	?	?	?

[x] There is something in the documentation with rsync where it is missing that a command should be run on a local computer
- [x] Ask Pavlin where exactly
- Was already correct :-)

I imagined having all learners in the same main Zoom room. Then I saw Pavlin using breakout rooms. He was right! Thanks to him, I started using breakout rooms too. Thanks Pavlin!

Changes to schedule: - Prevaluation takes 3 minutes to fill in - Due to the prevaluation, 10 minutes is enough for the full introduction, so next session can start 10 minutes earlier - File transfer can be shortened by 15 mins - There was too little time for 'Efficient workflow', needs 15 more minutes - Installation part 1 can be from 13:00-13:45 (instead of to 13:50).

RStudio took only 15 minutes, as most learners had already done some

[ ] Suggest move RStudio to Basic
13:57: Hmm, I forgot to properly prepare the summary :-/ , let's do so now :-)
15:19 Evaluation started
15:20 A learner thanked me and Björn

None of teachers used a shared document.

[ ] Remove a link to a shared document from the course material

The learners were very enjoyable to work with, as they were very social. I guess some credit goes to us teachers for that, but they made it work on the other end of the conversation.

When	Who	What	Suggest
9:00	R	Introduction	9:00
9:10	R	Prevaluation	9:03
9:20	P	Transferring files p1	9:10
10:00	.	Break	10:00
10:15	P	Transferring files p2	10:15
10:45	R	Efficient jobs	10:30
11:00	.	Break	11:00
11:15	R	Efficient jobs	.
11:20	R	Complex jobs	11:30 (parallel)
11:40	R	Replicate jobs	11:30 (parallel)
12:00	.	Lunch
13:00	B	Software and packages installation p1	13:00
13:50	.	Break	13:45
14:00	B	Software and packages installation p2	14:00
14:40	BR	[IDEs on Bianca p1	14:30
14:50	.	Break	14:45
15:05	L	IDEs on Bianca p2 -> something else	15:00
15:40	R	Summary	15:30
15:45	R	Evaluation	15:35
15:45	All	Optional Q/A	15:45
16:00	.	END	16:00

Complex and replicate jobs are in parallel sessions, suggest to replace IDEs by one of these

Evaluation results¶

Evaluation can be found here.

I think the most interesting is:

question	mean_pre	mean_post	p_value	different	teacher
I can transfer files to/from Bianca using rsync	2.2	4.6	0.2857143	FALSE	P
I can see the CPU and memory usage of jobs	2.4	3.8	0.2857143	FALSE	R
I understand how to set up jobs efficiently	2.0	3.6	0.0476190	TRUE	R
I can schedule a simple workflow of jobs that depend on each other using Slurm	1.2	3.4	0.0793651	FALSE	R
I can schedule a simple workflow of jobs that depend on each other using Nextflow	0.2	1.2	0.2063492	FALSE	R
I can run replicate jobs using Slurm job arrays	1.4	2.2	0.6825397	FALSE	R
I understand how to install software myself	1.2	3.6	0.0476190	TRUE	B
I understand how to use packages and libraries for scripts	3.0	3.8	0.5238095	FALSE	B
I understand what containers are	2.0	3.0	0.2857143	FALSE	B
I understand how to build from source	0.8	1.8	0.1904762	FALSE	B
I can can run the voted-for IDE on Bianca	1.6	4.6	0.0476190	TRUE	BR
I understand how to deal with sensitive data	3.6	4.4	0.2857143	FALSE	None

Let's go trough mine:

question	mean_pre	mean_post	p_value	different	teacher
I can see the CPU and memory usage of jobs	2.4	3.8	0.2857143	FALSE	R

Nice improvement, keep it as it is.

question	mean_pre	mean_post	p_value	different	teacher
I understand how to set up jobs efficiently	2.0	3.6	0.0476190	TRUE	R

A significant positive effect! Keep it as it is!

question	mean_pre	mean_post	p_value	different	teacher
I can schedule a simple workflow of jobs that depend on each other using Slurm	1.2	3.4	0.0793651	FALSE	R

These was done by 3 out of the 5 learners, which has had a great impact on the confidences. Keep it as it is!

question	mean_pre	mean_post	p_value	different	teacher
I can schedule a simple workflow of jobs that depend on each other using Nextflow	0.2	1.2	0.2063492	FALSE	R

Great, I did not do that at all :-)

[ ] Remove this learning objective for next time

question	mean_pre	mean_post	p_value	different	teacher
I can run replicate jobs using Slurm job arrays	1.4	2.2	0.6825397	FALSE	R

These was done by 2 out of the 5 learners. There has been a 3 and 4 for confidence added, so it may have been a good session.

question	mean_pre	mean_post	p_value	different	teacher
I can can run the voted-for IDE on Bianca	1.6	4.6	0.0476190	TRUE	BR

A significant positive effect! Keep it as it is.

Judging all this, my findings are:

I don't care about the prevaluation much: let's not bother the learners with it
Remove the 'I can schedule a simple workflow of jobs that depend on each other using Nextflow' learning objective
I suggest to move RStudio to the Basic Bianca course
I suggest to remove a link to a shared document from the course material
If both complex and replicate jobs are important, I suggest to add 30 mins to that session, over going parallel
I suggest a simpler schedule