facebookresearch/CompilerGym
A library of reinforcement learning environments for compiler and program optimization tasks, developed by Meta/Facebook Research.
Not currently ranked — collecting fresh signals.
star history
CompilerGym provides easy-to-use and performant reinforcement learning environments that frame compiler optimization problems in a way familiar to ML practitioners. It wraps compiler toolchains (LLVM, GCC, etc.) as interactive gym-style environments where agents can propose optimizations and receive feedback on metrics like instruction count, compile time, or binary size. This enables ML researchers to apply reinforcement learning, search algorithms, and other ML techniques to practical compiler optimization challenges.
Frequently asked
- What is facebookresearch/CompilerGym?
- A library of reinforcement learning environments for compiler and program optimization tasks, developed by Meta/Facebook Research.
- Is CompilerGym open source?
- Yes — facebookresearch/CompilerGym is open source, released under the MIT license.
- What language is CompilerGym written in?
- facebookresearch/CompilerGym is primarily written in Python.
- How popular is CompilerGym?
- facebookresearch/CompilerGym has 1k stars on GitHub.
- Where can I find CompilerGym?
- facebookresearch/CompilerGym is on GitHub at https://github.com/facebookresearch/CompilerGym.