Schedule & Readings
Course Schedule
“The goal isn’t to read everything, but to read the right things deeply and connect them meaningfully.”
AI agents designing the complete computing stack - from code to silicon
📖 Reading Reflection Due: Before each class session
🎯 Discussion Leadership: Sign up by Week 3
The Thematic Flow:
- AI for Software: Agents understand what needs to be computed efficiently
- AI for Architecture: Agents design how to compute it efficiently in hardware
- AI for EDA: Agents implement the architecture physically in silicon
| Week | Date | Topic | Key Papers |
|---|---|---|---|
| 1 | Sep 3 | Course Introduction | |
| Sep 3 | The AI Agent Revolution in Architecture | Course overview and vision | |
| Sep 4 | Architecture 2.0: Foundational Vision | A graph placement methodology for fast chip design Chip Placement with Deep Reinforcement Learning |
AI for Software
Software optimization and compilation
| Week | Date | Topic | Key Papers |
|---|---|---|---|
| 2 | Sep 9 | AI for Compiler Optimization | |
| Sep 9 | CompilerGym and RL for Compilation | TBD | |
| Sep 11 | LLVM and ML-Driven Passes | TBD | |
| 3 | Sep 16 | AI for Software Generation & Autotuning | |
| Sep 16 | TVM and Tensor Compilation | TBD | |
| Sep 18 | Halide and Domain-Specific Languages | TBD | |
| 4 | Sep 23 | AI for Performance Prediction | |
| Sep 23 | Performance Modeling with ML | TBD | |
| Sep 25 | Workload Characterization | TBD | |
| 5 | Sep 30 | AI for Memory & Data Optimization | |
| Sep 30 | Memory Access Pattern Learning | TBD | |
| Oct 2 | Data Layout and Prefetching | TBD | |
| 6 | Oct 7 | AI for Parallel Code Optimization | |
| Oct 7 | Parallel Scheduling and Load Balancing | TBD | |
| Oct 9 | GPU Code Optimization | TBD |
AI for Architecture
Hardware design and system architecture
| Week | Date | Topic | Key Papers |
|---|---|---|---|
| 7 | Oct 14 | AI for Design Space Exploration | |
| Oct 14 | ArchGym Fundamentals | TBD | |
| Oct 16 | Reinforcement Learning for Architecture | TBD | |
| 8 | Oct 21 | AI for Cache & Memory Hierarchy | |
| Oct 21 | Cache Replacement and Prefetching | TBD | |
| Oct 23 | Memory System Design | TBD | |
| 9 | Oct 28 | AI for Branch Prediction & Control Flow | |
| Oct 28 | Neural Branch Prediction | TBD | |
| Oct 30 | Control Flow Optimization | TBD | |
| 10 | Nov 4 | AI for Processor Microarchitecture | |
| Nov 4 | Superscalar Design and Scheduling | TBD | |
| Nov 6 | Out-of-Order Execution Optimization | TBD | |
| 11 | Nov 11 | AI for Accelerator Design | |
| Nov 11 | Neural Architecture Search for Accelerators | TBD | |
| Nov 13 | Domain-Specific Architecture Generation | TBD |
AI for EDA
Physical implementation and chip design
| Week | Date | Topic | Key Papers |
|---|---|---|---|
| 12 | Nov 18 | AI for Physical Design | |
| Nov 18 | DREAMPlace and GPU-Accelerated Placement | TBD | |
| Nov 20 | Routing and Physical Synthesis | TBD | |
| 13 | Nov 25 | AI for Logic Synthesis & Timing | |
| Nov 25 | Logic Synthesis and Optimization | TBD | |
| Nov 27 | No class - Thanksgiving Break | ||
| 14 | Dec 2 | Integration & Future Directions | |
| Dec 2 | Industry Perspectives + Project Presentations | Student presentations |
📝 Projects Due: December 2
Reading Access
- Primary Sources: Papers available through Harvard Library, ACM Digital Library, IEEE Xplore
- Course Reserves: Key papers uploaded to course management system
- Supplemental Materials: Additional resources linked from course website
Discussion Format
Each session follows a structured format:
- Opening (10 min): Key takeaways and questions from readings
- Deep Dive (45 min): Structured discussion led by students or instructor
- Matrix Mapping (15 min): How do these papers fit in our framework?
- Synthesis (5 min): Connections to previous sessions and upcoming topics
Schedule subject to adjustment based on guest speaker availability and emerging research developments.