Plan to complete the assigned readings prior to the start of class. The lectures will complement the readings and may cover different material, or offer a different perspective. I will expect you to be familiar with the assigned readings and, accordingly, to participate in class discussions: to ask and answer questions, offer commentary and insight. You may find it helpful to keep a journal of what you've read, especially research papers that will be assigned from time to time. Here are some tips on writing summaries. If you have questions about the readings, post them to Moodle.
Grading will be based on 3 programming labs. You'll learn how to optimize for the memory hierarchy, to program CUDA, MPI and OpenMP. The labs include a report, which will become progressively more important with each succeeding lab. Be sure to allow for time to complete this important part of the assignment.
All the programming labs may be done in teams of 2 or 3. I will permit a few teams of 1, but you must first obtain my permission. You may switch partners from assignment to assignment, but not once an assignment has started (unless your partner drops the course). If you do want to switch teams for the next assignment, you must inform your partner 3 days before the due date of the current assignment.
You will have a month to do the last lab which will give you the opportunity to conduct a mini research project. I will provide more details a couple of weeks into the quarter.
While I encourage you to discuss labs and projects with your classmates, the work done by your team must be independent of others’. If you take written notes while discussing someone else’s work, or anytime afterwords, then I cannot consider your work to be independent. You must not use someone else’s code, nor code you wrote for another course, unless I authorize it. In some cases it may be appropriate for you to use code you've previously authored. The general criteria I'll apply to considering such code is whether or not it functions in a support role for further exploration, but doesn't replace the assignment or significantly reduce the learning experience. See me if you have code you'd like to re-use; for code written jointly with others, you must obtain the co-author's permission.
Though I don’t expect there to be a problem
in a graduate course, Academic Integrity will be strictly enforced.
Anyone found
plagiarizing another’s work, or making their own work available to
others,
will receive a zero grade for the work
in question and face other possible consequences.
You are assumed to be familiar with the Academic Honestly Policies for this
course, as described in the following document,
which also discusses code usage.
If you aren’t sure about this policy be sure and see Professor Baden. Academic Integrity policies apply to teams as well as individuals.
I'll consider any reasonable request for a delayed turnin, but generally you must turn in the assignment on time to receive full credit. I'll accept regrade requests for one week after the assignment has been returned. After that time, the grading decision is final.
Maintained by [Sat Jan 4 18:59:16 PST 2014]