Room: 370-370

MW 1:15 to 2:30

Instructor: William J. Dally

billd@csl.stanford.edu

Gates Room 301

(650)725-8945

Hours: MW 2:30 to 3:30 (except as noted – see announcements)

TA: Mattan Erez

Gates 303

(650)725-3726

On-Line Info: available via http://izio.stanford.edu (and http://www.stanford.edu/class/ee482)
 
 

Goal

This course will introduce students to current research topics in high-performance processor design. We will explore processor architecture from branch prediction and fetch strategies through execution and memory issues.

Organization

The course is organized as a discussion course with each class meeting allocated to the discussion of a particular topic. Before each meeting, all students are expected to read three or four papers describing recent research on that topic. One student will be responsible for leading each discussion and one student will be a scribe for each meeting, responsible for recording and writing up the discussion. There will be no homework assignments and no exams. There will be two written assignments and one project.

Assignments

Lead a Discussion and Summarize the Papers

Each student (or a pair of students if enrollment is over 16) will be expected to lead the discussion during one class meeting. In addition to reading the papers, which all students are expected to do, these students should come to class prepared to lead the discussion. This will involve:

1. Making a brief presentation summarizing and critiquing the readings. Please do bring visual aids (e.g., copies of figures from the papers) to assist with the presentation.
2. Place the research described by the papers in context and describe why it is (or isn’t) important.
3. Describe the major open research issues in this area.
4. Prepare a list of questions concerning the papers and the research area to kick off the class discussion.

Pseudo-DARPA Proposal

Proposal writing is an important research skill. To give you all practice with this, we will be handing out a simulated request for proposals in the area of high-performance processor design and asking each of you to respond to this request by writing a proposal. You may work on these proposals in groups of up to three students.

Project

The project will involve investigating some aspect of high-performance processor architecture via simulation. For example, you may develop and evaluate new branch prediction algorithms, prefetch strategies, or machine organizations. We suggest that the project investigate one of the areas that we will be discussing during class, but you are free to propose any research topic in this area. You may work on these projects in groups of up to three students.

Grading

EE282.