OSU eCampus CS475 Parallel Programming review & recap

This post is part of an ongoing series recapping my experience in Oregon State University‚Äôs eCampus (online) post-baccalaureate Computer Science degree program. You can learn more about the program here.

Six-word summary: A well-designed and satisfying course

I feel like I sometimes write these recap articles to complain and warn other students of landmines, but CS475 was what every OSU online CS course should be: organized, succinct, and educational. I learned so much and enjoyed the course from start to finish.

CS475 Review

It only runs in the Spring quarters, but it’s worth holding out for! You’ll learn a lot, the material is interesting, the course is well-paced, and the exams felt fair.

I picked this class because parallel programming was something I thought sounded intimidating and was unlikely to be something I would attempt on my own. I’m very glad I chose it! It was way less intimidating than I feared, and now I feel entry-level competent in a subject I formerly knew nothing about.

Class structure

  • 10 weeks
  • 8 programming assignments where you code, run the code, and then write an analysis of it
  • I spent about 7-10 hours on this class each week, making it one of the lighter workloads in the program
  • Proctored midterm, proctored final
  • Weekly quizzes in Canvas (usually 5 questions with a full hour to answer them)
  • About 1-1.5 hours of video lectures each week (sadly they are without captions, but they are very high quality otherwise)
  • Some code is provided for you to use, saving time for focusing on what the class is really about
  • Instructor attends office hours! He’s really approachable, you should go even if you don’t have a specific question!

The instructor is SUPER active! You can get on a video chat with him – he’s just sitting there at pre-announced times, waiting for people to show up to his video chat room! He’ll look at your code! He’ll talk to you 1:1 about topics! I’ve never had such a productive “office hours” experience as I did in this class, and I didn’t even show up with any questions.

Contrast this with how most classes in this program hand you a problem and set you loose on figuring it yourself (via Google, Stack Overflow, old projects posted to GitHub, etc). There’s a time and a place for that, sure, but for $2000 in tuition I’m happiest when the class makes me feel like I learned something from the person teaching it. I want an experience I can’t get from Google and an independent project. I want an intelligent, skilled person to say “here’s what I know, and here’s how I think about it” and that’s how Professor Bailey ran this class.

Prof. Bailey also sends out an anonymous survey at the end of the course asking for feedback on each and every assignment as well as time spent. He genuinely cares about the class.

CS475 Project Structure

Unlike most classes, CS475 hands you some “starter code” for every assignment. You still have to put it in the right place, call it at the right time with the right data, and format and analyze the output yourself, but being provided this “starter code” was great for a few reasons:

  • Less time spent on things that aren’t the point of the assignment
  • No super fancy math skills needed (the bezier volume calculation formula is given to you, for example)
  • Less likely that you’ll go off into the weeds and do something ridiculously wrong and miss the point of the assignment
  • Less time needed overall for each assignment

This class has you throw nearly 100% of your efforts towards parallel programming related coding and analysis. For those of us with a finite number of hours in a week it was so good to just focus on the subject of the class and skip some of the boilerplate project setup. This class would pair well with another less-time consuming class, if you need (or prefer) to double up.

Most of this class’s projects go something like this: start a new C project, bring in the starter code, refactor/add to it to make it do what you need, write a bash script (or whatever you prefer) to run a loop and pass different parameters to the program each run, get all of the output in your Terminal (some people write it to a file), dump the output into a spreadsheet (I used Google Sheets), create graphs and write your analysis (in response to prompts).

Typical project output – here’s my output from Project 2

Bring these into your favorite spreadsheet software and create a pivot table. Then, generate the needed graphs for your analysis.

Somehow I made it this far in life without ever needing a pivot table. In this class, pivot tables made graphs a cinch.

And that’s pretty much how the assignments go.

I was a bit intimidated by this class before I took it. People say it’s dry and the subject has a reputation for being challenging, so I wanted to show other prospective students what’s actually involved and how approachable it really is.

Tips for CS475

NVIDIA CUDA support is limited on some versions of Mac computers. (For example, Apple just doesn’t support it for OS 10.14 at the time of this writing). If you’re on a Mac, I suggest using the school-provided rabbit server for your CUDA assignments and not wasting hours like I did trying to get it all working on a Mac.

Go to the office hours! Seriously, I went on a whim near the end of the quarter and was surprised to find myself alone in a video chat with the professor. I was just going to lurk but it ended up being a great chance to discuss a bunch of class topics one on one. Another student showed up and he stepped through her previous assignment code with her. You can’t beat the value of having an expert review your code with you!

The exams are straightforward. Some classes in this program use exams as an opportunity to test your speediness and your ability to keep track of registers through multiple iterations of multiple loops, but this class kept it fair and simple: watch the lectures, take notes, study those notes and you’ll do fine! He even gives a generous amount of time for the exams.

Overall, it’s just a high quality course. The lectures were freshly recorded, well-paced, on-topic, and relevant to the assignments and tests. The weekly assignments always looked challenging at first but were ultimately quite doable and enjoyable.

I’ve got just 4 classes left to complete my degree. I’ll be back in the fall quarter for Intro to Networks! See you all then!