Resources
Essential Books
Machine Learning Systems
Introduction to Machine Learning Systems by Vijay Janapa Reddi
- Why: Provides the foundational understanding of how ML systems work
- Focus: Chapters on hardware-software co-design and system optimization
- Connection: Essential background for understanding AI-driven architecture design
Deep Learning Fundamentals
Understanding Deep Learning by Simon J.D. Prince
- Why: Solid mathematical foundations for understanding AI techniques used in architecture
- Focus: Optimization methods, neural networks, and learning theory
- Connection: Background for papers using neural approaches to architecture problems
Computer Architecture
Computer Architecture: A Quantitative Approach by Hennessy & Patterson
- Why: The definitive reference for architecture fundamentals
- Focus: Performance analysis, memory hierarchy, instruction-level parallelism
- Connection: Required background for understanding what AI agents are trying to optimize
Computer Organization and Design by Patterson & Hennessy
- Why: More accessible introduction to architecture concepts
- Use: Reference for students needing architecture review
Key Research Tools
ArchGym
ArchGym: An Open-Source Gymnasium for Machine Learning Assisted Architecture Design
- What: RL environment for architecture design space exploration
- Use: Hands-on experience with AI-driven architecture optimization
DREAMPlace
DREAMPlace: Deep Learning Toolkit-Enabled GPU Acceleration for Modern VLSI Placement
- What: GPU-accelerated placement engine using deep learning
- Use: Understanding AI applications in physical design
CompilerGym
CompilerGym: Robust, Performant Compiler Optimization Environments for AI Research
- What: RL environments for compiler optimization
- Use: Exploring AI agents for code optimization
Paper Access
Primary Sources
- ACM Digital Library: ISCA, MICRO, ASPLOS papers
- IEEE Xplore: DAC, ICCAD, HPCA papers
- arXiv: Latest preprints in cs.AR and cs.LG
Paper Discovery
- Semantic Scholar: AI-powered search with citation analysis
- Connected Papers: Visual exploration of paper relationships
“The goal isn’t to read everything, but to read the right things deeply and connect them meaningfully.”