Course Policy

Course Policy

Textbook H&P | Hennessy & Patterson | Computer Architecture: A Quantitative Approach | 6th Edition

Required readings are posted on the course schedule and they include chapters from this required textbook, excerpts from other H&P editions, as well as research papers.

Course Description

The objective of this course is to learn the fundamental aspects of computer architecture design and analysis. Topics include processor design, pipelining, superscalar, out-of-order execution, caches (memory hierarchies), virtual memory, storage systems, simulation techniques, parallel architectures, warehouse scale computing, domain-specific accelerators, and technology trends and future challenges.

Prerequisite

You must have already taken CompSci/ECE 250 at Duke or equivalent. If you are an ECE MS student, you must have already taken ECE 550 (concurrent enrollment is not permitted).

Office Hours

Professor Wills’s office hours

D304 LSRC — Tuesdays 2.45 pm - 3.45 pm or by appointment. Please sign up on my Calendly for a 15-minute slot. I will not debug students’ code.

TAs’ office hours

D Wing 3rd Floor LSRC — Tuesdays and Wednesdays 4-5 pm; Thursdays and Fridays 3-4 pm.

Course Resources

These are the course resources. Students are expected to keep up with any information disseminated through these resources, i.e., course website, Canvas website, and Ed discussion site.

Course Website

This is the main course website. Course schedule is posted here along with required textbook chapters, additional excerpts, and research paper readings.

Canvas

CompSci 550/ECE 550 Fall 2026 Canvas Website is a supplement to the main course website. Students will find lecture slides and assignments on Canvas. Students will also be able to access the Announcements from this site.

Ed Discussion

This course uses Ed Discussion for students to post questions about the course. The authors of the posts will be visible to the instructor and TAs even when they are anonymous to students.

Academic Misconduct

Students are expected to adhere to the Duke Community Standard.

I will not tolerate any academic dishonesty. This includes cheating on the assignments, project, exams, summaries and plagiarism on summaries or project. Be especially careful on your project to cite prior work and give proper credit to others’ research. I will refer all suspected cases of cheating to the Gradaute School or Duke Undergraduate Judicial Board as appropriate. Please ask the instructor if you have any questions about misconduct.

Grading

  • Class Participation (up to 1% of Course Grade extra credit at the instructor’s discretion)
  • 5% Research Paper Summaries
  • 20% Assignments (up to 5% of Assignments extra credit; i.e., up to 1% of Course Grade extra credit)
  • 20% Midterm Exam – Closed Book, Closed Notes, Closed Internet (No Calculators, Phones, and Smart Watches)
  • 30% Final Exam – Closed Book, Closed Notes, Closed Internet
  • 25% Semester Project

Class participation

Students are expected to attend and actively participate in lectures. To give an incentive to attend lectures and foster interaction, you will earn up to 1% of Course Grade extra credit for actively answering and asking questions in class, and participating in discussions.

Research Paper Summaries

This course requires readings from the textbook and from selected research papers (there is a lot of reading for this class, be prepared…). You will improve your ability to read research papers by the end of this semester. The contents of the textbook readings and research papers will be on exams even if we don’t explicitly cover something in lecture. There will be several class sessions devoted to discussing approximately 12-16 research papers (1-2 each class). Students will write short summaries for each paper prior to the class discussion.

Paper summaries are due at midnight the day before lecture.

Complete each paper summary individually, concisely, in two to three paragraphs. The summary must describe the problem and its motivation, the paper’s main ideas or approach (such as the solution, method, or technique), and an analysis of its strengths and weaknesses. Also include at least one discussion question. Strengthen summaries by contextualizing the paper’s ideas in relation to concepts covered in the course.

The summaries must be entirely your own words, do not copy abstracts, paragraphs, sentences from the paper or elsewhere.

Summaries are graded as 2, 1, or 0 (2: includes all components and is decent, 1: incomplete or poor, 0: not turned in or turned in late).

Assignments

There are a total of five assignments for this course. Each assignment has a conceptual part and a practical part. The pratical part includes modifying a simulator, running simulation, performing experiments, and analyzing results. Assignments are out 1-2 weeks before the due date and it is due before midnight on Fridays. Please turn in your assignments via Gradescope and .pdf when there is a written part.

There is a one-day late policy on the assignments. 50% of the credit is given if the assignment is one day late. After that, 0 points will be given for the assignment. The time unit for the late policy is on a day granularity — not hours and not minutes. The late penalties are applied to the entire assignment and not by question. For example, if one question is turned in 1 day late for Assignment 2, you will get 50% * <your score of Assignment 2>. We will always grade your latest submission, regardless of if you scored higher on earlier submissions.

To give an incentive to start your assignments early, you will earn 5% extra credit for turning in your assignments one day early. Again, the extra credit is on a day granularity — not hours and not minutes. For example, if you turned in Assignment 2 by 11:59pm on Thursday of the week that Assignment 2 is due, you will get 105% * <your score of Assignment 2>.

You are expected to complete the assignments individually, unless otherwise stated.

Exams

There will be one midterm exam (in class 75 minutes) and a cumulative final exam (3 hours) in this course.

Regrade Requests

Regrade requests for any assignments/midterms have to be submitted within one week of grade release. Please submit your regrade requests via GradeScope.

Semester Project

The course project is expected to be completed in groups of 2-3 students. A project is graded based on a project proposal (10% of the project grade) and a project status report (15% of the project grade), with the remaining grade from a project presentation (25% of the project grade) and a final written report (50% of the project grade). Your project presentation will have a peer-graded portion and your attendance is required for both presentation days. Your project can be combined with ongoing research or a project in another course; either of these requires approval from all instructors and advisors involved.

All project documents should conform to the following formatting: 11pt font, single space with 1 inch margins.

Project Proposal

The project proposal should be one to two pages and include a statement of the project goal/hypothesis and an argument about why you think the idea will work (i.e., a well thought out justification). Details will be posted in the assignment section on Canvas.

Project Status Report

Your project status should be 3 pages maximum and include the following items.

  • Statement of project goal/hypothesis
  • Proposed solution and feasibility study to back up the solution
  • Evaluation methodology
  • List of milestones for project (e.g., simulator development, benchmarks, experiments, etc.)
  • Timeline for completion of milestones
  • Indication of progress on various milestones (e.g., completed, not started, x% complete, etc.)
  • Related work with appropriate citations

Details will be posted in the assignment section on Canvas.

Project Final Report

The standard for projects is high. The final report should be 4 to 8 pages long. I expect you to spend significant time on this project. Your final document should be conference/workshop caliber in presentation and style (the results don’t need to be publishable). Use the papers we read during the semester as guides for style and format. Clearly articulate the problem, your proposed solution, methodology, and interpretation of results (not just presenting numbers). You should also include sufficient related work, with an explanation and understanding of how it fits together.

No late projects or paper reports will be accepted.