(COL8085/861) Hardware System Design for AI
Officially known as 'Special Topics in Hardware Systems I'
Logistics
| Instructor | Sumit Diware (sdiware@cse.iitd.ac.in) |
| Academic Session | 2026–27 (Semester I) |
| Lecture Schedule | Tue, Fri 15:30–17:00 @ LH-620 |
| Office Hours | Wed 10:00–11:00 @ SIT-213 |
Policies
| Audits | Not allowed. Registration or sit-through only. |
| Attendance | Below 75% = one-grade penalty (e.g. A → A−), except for original D grade. |
| Plagiarism/cheating | Zero marks in the related evaluation component. |
| Classroom nuisance | Banned from attending further classes and visiting office hours. |
| Electronic devices | Usage not allowed, except for recording attendance. |
| Instructor slides | Not shared, attend and take notes in the class. |
| Late/no submission | Zero marks, no retakes under any circumstances. |
| Project PPT retake | Only with IITD hospital medical certificate, date/time/venue as per instructor. |
Prerequisites
- (COL215) Digital Logic Design
- (COL216) Computer Architecture
- Willingness to learn HDL (e.g. Verilog) via self-study.
Note: The instructor can adapt the content this point onwards as needed.
Course Description
This course aims to provide an overview of modern AI computing systems from a hardware perspective. It covers the characteristics of contemporary AI workloads, current AI hardware systems that execute them, and their evolution towards emerging AI hardware systems.
The tentative course outline is as follows:
- Fundamentals of AI Software
- Introduction to Hardware System Design
- Conventional AI Hardware Systems
- Emerging AI Hardware Systems (near/in-memory, neuromorphic, etc.)
- Software Considerations for Emerging AI Hardware
As the course involves research-based content, we will also discuss effective strategies for reading and understanding research papers.
Grading Scheme
The course involves three assignments (one of which includes an in-class quiz) and one project as follows:
| Component | Description | Max. Marks |
|---|---|---|
| Assignment-1 | Development and deployment of AI model. | 10 |
| Assignment-2 | Basics of hardware design, follow-up quiz. | 25 |
| Assignment-3 | Exploring common hardware accelerators. | 05 |
| Class Project | AI hardware design, demo, and presentation. | 60 |
Grading involves subjective decisions. These are at the instructor's discretion and no discussion will be entertained.