Yeah, the powershell scripts are woefully outdated. I was planning to create a new version of the script once I got around to adding the “scaffold a new exercise” feature in configlet (for which I haven’t yet found the time).
As a sidenote, thanks for all of your work on this platform and your exemplar solutions, I have found it very helpful! And as I’m on my phone the response here may be a touch brief.
I would like to add a problem I came across in my master’s studies. The ’ n body problem’ it’s a little physics heavy but when provided with the pseudocode it’s not too bad.
The main learning point would be to follow on from the other ‘parallel’ exercise, in showing that multi threading of a for loop is possible so long as loop dependencies are considered.
I would suggest students to engage in their own benchmarking and optimisations offlfine with some brief pointers, it’s a fairly studied topic so there’s a lot of approaches. I would like the auto feedback to include guidance if parallelism does not seem to be implemented.
High performance computing and parallelism in general I think is a topic not widely enough covered on exercism given how much one can learn.
Ehm, we have a policy of trying to limit the amount of math in our exercises, so I think this would be on the “too much math” side.
This is an excellent idea, it’s just that I don’t know if the n-body problem is the best way to go about it. Maybe we can try coming up with a different exercise?
If you’re creating a new exercise, consider if it should be added to the problem specs and/or if it is going into the C# track, how it can be worded to best allow it to eventually show up in the specs.