Fall 2015 :: CSE 610 - Parallel Computer Architectures

Paper Reading Assignments

The goals of this activity are threefold:

  • To make sure you come to class prepared by having read on the topic at hand a priori;
  • To practice and master the art of reading scientific papers, especially those on Computer Systems research;
  • And to engage in instructive discussions about the cource topics with other researchers (your classmates and the instructor) by asking good questions and subjecting your ideas to careful scrutiny.

You should post your paper analyses on the course newsgroup. For each assigned paper, you should make at least 2 posts on that paper. For any paper you wish to discuss start a new post titled “On <the paper title>” — if one does not already exist — and select “paper-reading” as its folder.

Your postings should demonstrate that you have read and carefully analyzed the paper. Do not summarize the papers. For example, your postings can include:

  • your questions, comments or critique of the paper,
  • answers to the questions, comments or critiques made by your classmates,
  • or discussions of other pieces of work that are closely related to the subject.

Seminar

For this assignment, you (in groups of 2) should pick one of the course topics marked as advanced on the Schedule page, prepare a short presentation to introduce that topic in one of the class meetings and then lead the class discussion on that topic for the rest of the meeting.

Term Project

You will undertake a semester-long research project in this course. The goals of this project are twofold:

  • to mimic the process of doing research in Computer Systems and Architecture and learn by doing, and
  • advance the state of the art in Parallel Computer Architecture and its applications by doing novel, original research.

The project will have four deliverables:

  • Proposal: to be formatted as a paper introduction
  • Midterm report: as a partial paper: intro, some implementation and results, some related work
  • Final report: formatted as a full paper
  • Final presentation: a conference-style presentation to be done in class
The deadlines are listed on the Schedule page.

Choosing the project topic

I have a number of research ideas that you can choose from. You are also welcome to choose a topic of your own as long as it is related to parallel computing and prallel architectures. You are especially encouraged to work on topics related to your thesis/dissertation research if there are parallelism-related aspects to them. In any case, please make sure you talk to me and get my approval on the choice of topic before writing your project proposal.