Blog article proposals and requests

If you would like to:

  • request that a Chapel blog article be written on a specific topic, or
  • propose an article that you would like to write and submit

then please post a response on this topic. Thanks!
-Brad

1 Like

So a Blog is just a separate Discourse category?

@damianmoz : Good question, you are too on top of things. No, we are launching it on the Chapel website, probably today, and I filed this topic and category so that I could start linking to it in my drafts. Stay tuned.

-Brad

PS — I may retroactively delete this sidebar of the conversation once the blog is live to keep this topic on-topic—no offense intended if I do.

[edit: the blog is now launched at Chapel Language Blog]

couple of ideas:

  • how to get started with chapel
  • 5 new features of Chapel 1.29.0 you should know about

articles I could offer to write (but not sure what's the blog scope, so if you think some are off topic, do not hesitate to tell)

  • when the Chapel exercism track I've been working on is finally launched (by the end of this year hopefully), I'd be happy to write a short blog entry about it.
  • typical library development workflow. I can volunteer to write one about how I work with mason etc, but you probably want to add a giant warning sign that it's just how I do it and not necessarily how it should be done :wink:
  • Automatic Differentiation 101 in Chapel, basic motivation, principles and how to do it in Chapel with ForwardModeAD (the library I recently released)

few more ideas:

  • Why you should learn Chapel in 2023
  • what every HPC specialist should know about Chapel
  • what every researcher should know about Chapel
  • users favorite features about Chapel (e.g. open a topic here on discourse to collect favorite features and summarize in a blog post)

Vectorization and GPU use within Chapel

Vectorizatiuon

  • where is it currently at (in terms of maturity)
    -- it is just the foreach statement
  • guidelines as how to best write things to exploit it
    -- in general and on super scalar processors
  • some examples

GPU exploitation

  • Writing code which can exploit GPUs with examples
  • I assume it is not generic w.r.t. CPU vs GPU
  • want to at least achieve the functionality of THRUST

Sadly my knowledge on the above is zero but I would like to learn.

What is the purpose of the blog category of Discourse apart from asking questions like this?

Sorry about the edits. My first attempt was a bit of a mess.

@lucaferranti and @damianmoz — Thank you very much for jumping onto this thread and making your suggestions! We're currently completely underwater on the blog trying to keep up with Advent of Code 2022 and our kickoff series on that, but once we get past that and start figuring out next steps for the blog, we'll circle back around on these. For now, I'll say that I think that all of these are valid and interesting topics for the blog.

And Luca (or Damian), we would welcome submissions on your proposals, though you should hold off on spending too much time on any of them until we get through AoC in case we make technology changes based on our experience... and we also need to publish instructions on how to draft a blog article to the public (short answer for now is markdown or Chapel source annotated with comments using markdown). I'll also note that my first three AoC articles are much longer than I tend to envision the typical blog article as being, due to the number of features we're trying to teach each day. So don't feel as though there's an expectation that they be this long or detailed.

To Damian's question:

As it says on the label, this is intended to be a category for discussions about the Chapel blog: article proposals and requests, comments and questions on published articles, look-and-feel improvements, etc. Put another way, our blog framework does not (currently, at least) support a way for readers to comment on published blog articles on the article's page directly. So this category is meant to fulfill that purpose. For example, I could imagine someone posting a topic to the category like:

  • "I didn't understand this section of this article, could you clarify?"
  • "I think the approach in this article is poor for this reason and would improve it in this way"
  • "Did you know you can earn $13k/month working from the comfort of your own home? click here to find out how!"

(OK, just joking about that last one... it's just a frequent comment style I see when posts have open comment sections)

-Brad

1 Like

Matrix Multiplication. especially

  • how to exploit vectorization
  • how to exploit GPUs for same
  • both unblocked and blocked, single-locale and multi-locale variants
    -- obviously below a certain size, an unblocked, single-locale is fastest
  • other considerations
2 Likes

Another topic is the throwing of errors, best practice, suggested approaches.

What is done currently in things like Arkouda and CHARMS.

Thanks...

1 Like