zwang4/awesome-machine-learning-in-compilers

What it solves

This project addresses the challenge of finding and organizing research on how machine learning can improve compilers and program optimization. It provides a centralized repository for developers and researchers to discover how AI techniques can automate complex tasks like code tuning, instruction scheduling, and memory management.

How it works

It acts as a curated, categorized index of academic papers, books, tools, and datasets. The repository organizes these resources by specific sub-fields of compiler technology, such as iterative compilation, instruction-level optimization, and auto-tuning, allowing users to quickly navigate the landscape of ML-driven systems research.

Who it’s for

This list is designed for researchers, compiler engineers, and students interested in the intersection of machine learning and systems programming. It is a starting point for those looking to understand the history and current state of using AI to make software run faster and more efficiently.

Highlights

• A comprehensive collection of research papers categorized by technical focus (e.g., parallelism mapping, code size reduction, cost modeling).
• Links to relevant books, tutorials, and talks to support deep learning in this domain.
• A curated list of benchmarks and datasets used for evaluating ML-based compiler optimizations.
• Information on key conferences and journals where new developments in this field are published.