Performance Engineering

"Performance is obtained by good engineering" -- Your Instructor

Course Information

  • Course Number: CS 4910/7680
  • Semester: Summer 20
  • Piazza: Forum Board
      Use piazza instead of e-mail for all course related matters
    • This way all course staff can answer questions
    • Class as a whole can participate and learn in discussions
    • Code also formats better nicely pasted in
    • (Linking your github is also easy for private posts)
  • Section 1: CS 4910/7680
    • Schedule: Lectures released Monday and Thursday at 6pm
    • Location: Online!


  • Instructor: Mike Shah
  • E-mail: mikeshah( a t )Northeastern{dot}edu
    • Use e-mail primarily for personal matters
    • e.g. I have a co-op interview and need to reschedule XYZ
    • e.g. I am away for a conference on date X to date Y
    • e.g. I am away for a doctors visit.
    • (Read How to send an e-mail)
  • Office: Nightingale 132A
  • Student Hours: In Location TBD
    • Review course material with me and ask questions
    • Date/Time: By appointment
    • Sign up for a 15-30 minute appointments arranged by e-mail on Google Hangouts here
Teaching Assitant(s)
N/A n/a (a t)
  • Hours/Location listed on piazza under Resources/Staff

Schedule/Road Map

The following is our tentative syllabus for the course, some changes should be expected throughout the semester. I will announce in class lecture, piazza, or through e-mail any major changes.

  • To get all of the assignments/activities for the course, you must first click the following link: Performance Engineering Monorepo Do not do a 'git pull' until class starts (Occasionally I make changes/spelling corrections)
Lecture and Readings Assignments Note(s)
1 Monday - May 04, 2020
  • Lecture outline
    • Administrivia
    • Course Introduction
    • Key Issues in Performance
    • Begin Module!
Lab 1 Out Welcome back to class!
1 Thursday - May 07, 2020
  • Lecture outline
A1 Out (Due May 15 at 11:59 PM EST)
-- --
2 Monday - May 11, 2020
  • Lecture outline
    • Bently's Rules of Optimization
    • Compiler Optimizations
    • Begin Module!
Lab 2 Out, Lab 1 Due
-- --
2 Thursday - May 14, 2020
  • Lecture outline
    • Modern Computer Architecture
    • Pipelining
    • Vector Hardware
    • Begin Module!
A2 Out (Due May 27 at 11:59 PM EST)
-- --
3 Monday - May 18, 2020
  • Lecture outline
    • Caching
    • Measuring Performance
    • Small Data-Oriented Design Case Study
    • Begin Module!
Lab 3 Out, Lab 2 Due
-- --
3 Thursday - May 21, 2020
  • Lecture outline
    • Cache-Aware Algorithms
    • Cache-Oblivious Algorithms
    • Begin Module!
-- --
-- --
4 Monday - May 25, 2020
-- --
-- --
4 Thursday - May 28, 2020
  • Lecture outline
    • Multicore machines and performance issues
    • SIMD Programming
    • Begin Module!
-- --
-- --
4 Monday - June 01, 2020
Lab 4 out, lab 3 due
-- --
5 Thursday - June 04, 2020
A3 Out (Due June 11 at 11:59pm EST)

A4 Out (Due June 12 at 11:59pm EST)
Final Project Repo (Out)
5 Monday - June 08, 2020
Lab 5 Out, Lab 4 Due
-- --
6 Thursday - June 11, 2020
  • Lecture outline
-- --
-- --
6 Monday - June 15, 2020
-- --
-- --
7 Thursday - June 18, 2020
  • Lecture outline
    • Trace Evaluation
    • Final Project and Course Summary
    • Begin Module!
Project Due June 21st (Anywhere on Earth) Last day of Summer 1 classes

Course Description

Registrar Description: This course introduces program analysis techniques for understanding, measuring, and developing high performing software for modern CPU Architecture. The course consists of a lecture component and in-class laboratory where students implement algorithms and data structures to optimize run-time performance of software systems. Out of class activities include reading state-of-the-art research papers used in academic and industrial research and completing assignments which students will create their own performance analysis tools and systems. The course culminates with a final capstone project which students complete individually or in small groups. The skills learned from this class will help prepare students for careers as systems engineers, performance engineering, or otherwise performing empirical research in the domain of systems or software engineering.

Course Objectives

By the end of this course, you will:

  • Students will learn about modern multi-core computer architectures
  • Students will learn fundamental compiler optimizations and what prevents the compiler from optimizing code
  • Students will understand how to use profilers (e.g. perf) to empirically measure performance.
  • Students will learn how to build data structures in a ‘data-oriented’ mindset by creating cache-aware and cache-oblivious data structures.
  • Students will understand how to setup a repeatable experiment to measure metrics for performance (e.g. cache misses vs cache evictions).
  • Students will learn one program analysis framework (e.g. LLVM) to build a performance analysis tool


There will be no required textbook for this course. However, these resources are recommended.

Optimized C++: Proven Techniques for Heightened Performance Optimizing software in C++ An optimization guide for Windows, Linux, and Macplatforms by Agner Fog


  • CS 3650 or CS 5600 (or CS 5600 co-requisite)

Academic Integrity and Non-Discrimination

Students and instructors are to follow the Northeastern policies on these important issues.

  • Northeastern Non-Discrimination Policy - This classroom is a safe space for the instructor and students to talk about ideas, share viewpoints, and learn.
  • Northeastern Academic Integrity Policy - You only cheat yourself if you are not honest. Most often cheating occurs when an individual falls behind or perhaps has other circumstances occurring in their life. Please consult the instructor before ever considering cheating.
    • If you are caught cheating I have to report the violation. My official policy is you receive a 0 in the course. Always remember, if you use any external sources, you must cite them.
  • Student Code of Conduct: Students and instructors will follow the following guide for how we conduct ourselves. This is to create a respectful environment where everyone can learn.

Make-Up Policy

  • Students participating in varsity athletics(this does not include club sports or intramurals) or other University sanctioned events may have the need for a make-up. Please contact me in advance of such events, so that appropriate accommodations can be made.
  • Occasionally, other life events and circumstances occur that were not planned. If this is the case, please e-mail me privately.
  • E-mailing me asking for extensions just because is unfortunately not fair to your classmates. The 10% penalty for each day late has to be enforced so I do not get taken advantage of.


Part of what makes Northeastern University unique, is our diverse cohort of students, faculty, and staff. In order to support this, Northeastern is committed to providing equal access and support to all qualified students through the provision of reasonable accommodations so that each student may fully participate in the University experience.

  • If you have a disability that requires accommodations, please contact the Student Accessibility Services office at or (617) 373-2675 to make an appointment with the Disability Resource Center representatives in 20 Dodge Hall to determine appropriate accommodations.
  • We Care is another university resource for helping tackle challenges you may be facing. Asking for help is okay!

Lateness and Attendance Policy

Students who do well in this course tend to show up to the course consistently, participate, start coursework the day it is assigned, and engage with their peers. Come to class, come on time, and build good habits! In-Class activities that are not completed during attendance are a zero.

Wellness Policy

While university can be difficult at times, please do take care of yourself. It is okay to seek help and take a break. Please communicate with your instructor, advisor, and officials at the university if you just need a mental break.

Northeastern Univerity provides resources for student healthcare and general wellness:

Assessment/Course Polices

Please find below the grading distribution that will be used for this course. You will find the grade you earn in this course on Canvas.

  • In-Class Activity: (5%)
  • In-Class Labs:     (20%)(Each equally weighted)
    • Labs are due before class of the following week.
  • Assignments:       (40%)(Each equally weighted)
  • Final Project:     (35%)

  • The grade system follows the University Grading System.
    • A  = 95 – 100
    • A- = 91 – 94.99
    • B+ = 87 – 90.99
    • B  = 83 – 86.99
    • B- = 80 – 82.99
    • C+ = 77 – 79.99
    • C  = 73 – 76.99
    • C- = 70 – 72.99
    • D+ = 67 – 69.99
    • D  = 63 – 66.99
    • D- = 60 – 62.99
    • F  =  0 – 59.99
  • Please do not ask me for an A on the first day of class (It happens enough I have a bullet point here).
  • In the event of a sick/snow day (i.e. we miss a lab or in-class activity) the weight of each assignment increases (There may also be shuffling of course material if we are interrupted).
  • The expectation is that the assignments are fair but difficult, so you should start early!
  • Late Submissions of Assignments receive 10% off per day submitted late (up to 3 days max, then 0% received)
    • Unfortunately, with large classes I cannot make individual exceptions (i.e. "special deals") fairly to your classmates who are likely making other personal sacrifices to complete work on time.
    • Come speak to me about your wellness if something is otherwise impeding your progress so we can provide you available resources to succeed.
  • Assignments that do not compile/open receive no credit Simply put, programs that do not compile do not do anything.
  • There are no "re-grades" or points awarded one week after your grade is posted. "re-grades" may result in a higher, equal, or lower score.
    • There are no "re-grades" after the semester is over.
    • Do not ask multiple members of the course staff for "re-grades"
  • If you are currently waitlisted, you must submit your homework on time. That is the gamble! If you do not have Canvas access, you will submit by e-mail or other course mechanism (e.g. github).
  • There are no extra credit assignments just for you. I reserve the right to add points to assignments that do go above and beyond however. I reserve the right to give an extra credit assignment to the entire class, though this is highly unlikely.
  • I reserve the right to modify the grading scale in your favor if you show exemplary proficiency in any of the catagories. I will never modify the scale to lower a students grade.
  • In class work cannot be made up at a later date unless otherwise arranged with the instructor well in advance.
    • Course work completed after the date cannot be graded, as solutions will have been discussed (this includes if taking this course for an Incomplete).
    • Once again, "in-class" work must be completed in-class unless there is a documented emergency or you have prearranged with the instructor a make-up well in advance.
  • Lab time is meant for helping students with the lab, not completing homework. I have to prioritize lab first, then can answer homework questions.
  • No Facebook, no cell phones. Not only does it distract you, it distracts others!
  • Everyone needs to come see me in office hours (or by appointment) at least one time during the semester to introduce yourself. The purpose is so that you:
    • Know where my office is.
    • Get used to coming to office hours.
    • Let me know how I can help you achieve your goals.

Please do not redistribute or host any materials without e-mailing me first. I generally am happy to share the latest .pdf or slide presentation with those who ask. Thank you for your time!