CS 5220

Applications of Parallel Computers

Please click the play button below.

Welcome to CS 5220, Applications of Parallel Computers, the Fall 2020 deluxe edition. I am your host, David Bindel. This slide deck covers some of the logistical aspects of the course. It is a strange semester, so the logistics are complex. I have taught at Cornell for the past decade, but I have not had a semester like this before, either. Also, this is a one-man band; I have no course staff this semester. So, please: bear with me! If you want to start with an overview and introduction to the technical part of the course, feel free to back out and look at that slide deck instead. It should be hyperlinked from the same course schedule page that brought you here. But the logistics are important, too, so I urge you to go through this deck carefully and ask questions, either on Piazza or in the course meeting.

Slide logistics

Slides produced with reveal.js
Narration via the audio-slideshow plugin
Press the S key to open speaker notes for a text version of the narration.

Enrollment

Initial enrollment was restricted to non-CIS majors
Plenty of room for the waitlist
Plenty of room generally, and open add/drop
Please do not register for a formal audit

For several years, we have restricted enrollment in upper-division CS classes. This is to make sure that our students, particularly the MEng students (who register late and who are only hear for one year), can take the classes they need. But parallel scientific computing is not ML, and we are not in danger of running out of room. If you were on the waitlist, you should be automatically enrolled in the next couple days. And if you want to add the class now, the add/drop period is not restricted. You can enroll through September 16, though sooner is better. In some semesters, I allow students to formally audit 5220. This semester, if you want to just follow along, I recommend keeping up with the material on the web page. Otherwise, I expect you to participate in the homework, the group projects, and other course activities.

Course community

A class is not just the lectures and HW!
We want group interaction
This can be synchronous or asynchronous
Several things planned, will adapt as we go

Course community

The typical mix of 5220 students includes:

CS undergrad and MEng students (and a few PhD)
PhD students from other disciplines

This means we assume you will:

Know a C family languages.
Not know some background.
Be a teacher and a learner.

Let's start with who you are. Usually, the two main audiences for the class are CS undergraduates and MEng students, and PhD students from other disciplines who need to do parallel computing for their research. There are also always a few CS PhD students, non-traditional students, or undergraduates from other disciplines. I assume you know how to program in some C family language, and I assume you know enough to learn how to navigate in a Unix environment without too much lecturing from me -- though I will provide some references for those of you who haven't done this much. But otherwise I try not to assume much about what you already know. What this means is that I expect some of you to be confused when I talk about computer architecture, and some of you to be confused when I talk about fluid dynamics, and some of you to be confused when I talk about numerical methods. And I may say some things that confuse everyone! Fortunately, you also have a lot of collective knowledge, and students in this class have worked together to do some pretty remarkable things. There are at least two PhD theses in the past few years that I know arose largely from projects in this class! So plan to learn from each other and teach each other as much as you learn from me (and teach me). And be prepared to ask lots of questions.

Course meeting

Main meeting: TR 8:00-9:30
I will help coordinate alternate meetings, too
This is not about lecture (class is flipped)
Goal is discussion and small group activities
We may also do some programming together!
Zoom recordings posted to Canvas

The main meeting for the class is Tuesdays and Thursdays at 8:00 AM, Eastern. This is a very awkward time for students who are coming in from California, or even for Ithaca students who happen to be night owls. So we will coordinate alternate meeting times, too, though these may involve only students and not me. I call this a meeting rather than a lecture because I will not be going through the narrated lecture slides with you, for the most part. Instead, we will use the time for question-and-answer, discussion, and small group activities. We may also walk through code or tool demos together, partly so that you can ask questions as we go. The Zoom recordings of these meetings will be posted to Canvas, though any activity that happens in breakout rooms will not be recorded.

Course buddies

Role: Make sure nobody falls through cracks
One or two buddies per student
You can nominate, if you know people
I may also do some buddy assignments

Project groups

Most course work is done in small groups
We will assign groups for initial projects
Final project is at your discretion
Active collaboration and pair programming are encouraged!

Collaboration and AI

5220 is supposed to be collaborative!
Want you to use online resources effectively
But for academic integrity: cite!

Information systems

Course home page
GitHub
Canvas
CMS
Piazza
Other (Discord, Slack, Gather, ...)?
Computing systems

Finally, let's talk about the class information systems. All of the systems we use will be linked from the course home page, though those that are restricted may be listed in the Cornell-only section. The non-restricted part of the home page is all built from a public Github repository, and you should feel free to file issues or pull requests (or just send email) if I make mistakes! There is a Canvas page for the class, but I mostly plan to use it for sharing Zoom recordings -- we have to restrict access to these because of FERPA. I will use our home-grown course management system (CMS) for homework submission and keeping track of grades. We will use Piazza for asynchronous course question-and-answer things, but I am also looking at a few options to support ad hoc synchronous interactions, including Discord, Slack, and Gather.town. I also encourage you to explore for yourself; for example, I know that VisualStudio has support for pair programming, though I haven't tried it out myself. If you come across tools that you find useful, let us all know! Finally, we will be using various external computing resources over the course of the class, and I'll distribute information on those when possible.

Homework Zero

Your netid
How you like to be called
Your background (degree program, major, etc)
Whether you can make the main meeting time
What times might work for you for synchronous meetings (with the class or project groups)
Any proposed "course buddies"
What you hope to get out of the course
Any concerns you have at this point

CS 5220

Applications of Parallel Computers

Slide logistics

Enrollment

Course community

Course community

Course meeting

Course buddies

Project groups

Collaboration and AI

Information systems

Homework Zero

Good luck!