Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

https://github.com/facebookresearch/CompilerGym

Reinforcement learning environments for compiler and program optimization tasks
https://github.com/facebookresearch/CompilerGym

Last synced: 3 months ago
JSON representation

Reinforcement learning environments for compiler and program optimization tasks

Host: GitHub
URL: https://github.com/facebookresearch/CompilerGym
Owner: facebookresearch
License: mit
Created: 2020-11-11T18:44:35.000Z (over 3 years ago)
Default Branch: development
Last Pushed: 2023-08-23T15:45:34.000Z (10 months ago)
Last Synced: 2024-03-04T15:33:43.932Z (4 months ago)
Language: Python
Homepage: https://compilergym.ai/
Size: 23.5 MB
Stars: 851
Watchers: 33
Forks: 120
Open Issues: 107
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
- Citation: CITATION.cff

Lists

awesome-production-machine-learning - CompilerGym - CompilerGym is a library of easy to use and performant reinforcement learning environments for compiler tasks. (Industry Strength RL)
my-awesome-stars - facebookresearch/CompilerGym - Reinforcement learning environments for compiler and program optimization tasks (Python)
awesome-machine-learning-in-compilers - CompilerGym - Reinforcement learning environments for compiler optimizations ([paper](https://arxiv.org/pdf/2109.08267.pdf)). (Software / Memory/Cache Modeling/Analysis)

README

        ![CompilerGym](https://github.com/facebookresearch/CompilerGym/raw/development/docs/source/_static/img/logo-padded.png)



  

  

      

  

  

  

      

  

  

  

      

  

  

  

      

  

  

  

      

  

  

  

      

  





  Reinforcement learning environments for compiler optimization tasks.





  

    Check

    the website

    for more information.

  



## Introduction

CompilerGym is a library of easy to use and performant reinforcement learning

environments for compiler tasks. It allows ML researchers to interact with

important compiler optimization problems in a language and vocabulary with which

they are comfortable, and provides a toolkit for systems developers to expose

new compiler tasks for ML research. We aim to act as a catalyst for making

compilers faster using ML. Key features include:

* **Ease of use:** built on the the popular [Gym](https://gym.openai.com/)

  interface - use Python to write your agent. With CompilerGym, building ML

  models for compiler research problems is as easy as building ML models to play

  video games.

* **Batteries included:** includes everything required to get started. Wraps

  real world programs and compilers to provide millions of instances for

  training. Provides multiple kinds of pre-computed program representations: you

  can focus on end-to-end deep learning or features + boosted trees, all the way

  up to graph models. Appropriate reward functions and loss functions for

  optimization targets are provided out of the box.

* **Reproducible:** provides validation for correctness of results, common

  baselines, and [leaderboards](#leaderboards) for you to submit your results.

For a glimpse of what's to come, check out [our

roadmap](https://github.com/facebookresearch/CompilerGym/projects/1).

## News

- November 2022: CompilerGym v0.2.5 adds support for Python 3.10 and drops

  support for 3.7. See [release

  notes](https://github.com/facebookresearch/CompilerGym/releases/tag/v0.2.5)

  for full details.

- April 2022: ⭐️ CompilerGym wins the **distinguished paper award** at

  [CGO'22](https://conf.researchr.org/home/cgo-2022)! You can read our work

  [here](https://arxiv.org/pdf/2109.08267.pdf).

- April 2022: 📖 Our **tutorial** at

  [CGO'22](https://conf.researchr.org/home/cgo-2022) was well attended. If you

  missed the event, you can work through the materials

  [here](https://chriscummins.cc/2022/cgo22-tutorial/).

- September 2021: 📄 CompilerGym was featured on the **Meta AI research blog**.

  You can read the post

  [here](https://ai.facebook.com/blog/compilergym-making-compiler-optimizations-accessible-to-all/).

## Installation

Install the latest CompilerGym release using:

    pip install -U compiler_gym

See

[INSTALL.md](https://github.com/facebookresearch/CompilerGym/blob/development/INSTALL.md)

for further details.

## Usage

Starting with CompilerGym is simple. If you not already familiar with the gym

interface, refer to the [getting started

guide](http://facebookresearch.github.io/CompilerGym/getting_started.html) for

an overview of the key concepts.

In Python, import `compiler_gym` to use the environments:

```py

>>> import compiler_gym                      # imports the CompilerGym environments

>>> env = compiler_gym.make(                 # creates a new environment (same as gym.make)

...     "llvm-v0",                           # selects the compiler to use

...     benchmark="cbench-v1/qsort",         # selects the program to compile

...     observation_space="Autophase",       # selects the observation space

...     reward_space="IrInstructionCountOz", # selects the optimization target

... )

>>> env.reset()                              # starts a new compilation session

>>> env.render()                             # prints the IR of the program

>>> env.step(env.action_space.sample())      # applies a random optimization, updates state/reward/actions

>>> env.close()                              # closes the environment, freeing resources

```

See the [examples](/examples) directory for agent implementations, environment

extensions, and more. See the [documentation

website](http://facebookresearch.github.io/CompilerGym/) for the API reference.

## Leaderboards

These leaderboards track the performance of user-submitted algorithms for

CompilerGym tasks. To submit a result please see

[this document](https://github.com/facebookresearch/CompilerGym/blob/development/CONTRIBUTING.md#leaderboard-submissions).

### LLVM Instruction Count

LLVM is a popular open source compiler used widely in industry and research. The

`llvm-ic-v0` environment exposes LLVM's optimizing passes as a set of actions

that can be applied to a particular program. The goal of the agent is to select

the sequence of optimizations that lead to the greatest reduction in instruction

count in the program being compiled. Reward is the reduction in instruction

count achieved scaled to the reduction achieved by LLVM's builtin `-Oz`

pipeline.

This leaderboard tracks the results achieved by algorithms on the `llvm-ic-v0`

environment on the 23 benchmarks in the `cbench-v1` dataset.

| Author                                           | Algorithm | Links | Date | Walltime (mean) | Codesize Reduction (geomean) |

|--------------------------------------------------| --- | --- | --- |-----------------|------------------------------|

 | Robin Schmöcker, Yannik Mahlau, Nicolas Fröhlich | PPO + Guided Search | [write-up](leaderboard/llvm_instcount/ppo/README.md), [results](leaderboard/llvm_instcount/ppo/results.csv) | 2022-02 | 69.821s  | **1.070×**                   |

| Facebook                                         | Random search (t=10800) | [write-up](leaderboard/llvm_instcount/random_search/README.md), [results](leaderboard/llvm_instcount/random_search/results_p125_t10800.csv) | 2021-03 | 10,512.356s     | 1.062×                       |

| Facebook                                         | Random search (t=3600) | [write-up](leaderboard/llvm_instcount/random_search/README.md), [results](leaderboard/llvm_instcount/random_search/results_p125_t3600.csv) | 2021-03 | 3,630.821s      | 1.061×                       |

| Facebook                                         | Greedy search | [write-up](leaderboard/llvm_instcount/e_greedy/README.md), [results](leaderboard/llvm_instcount/e_greedy/results_e0.csv) | 2021-03 | 169.237s        | 1.055×                       |

| Anthony. W. Jung                                 | GATv2 + DD-PPO | [write-up](leaderboard/llvm_instcount/gatv2_ddppo/README.md), [results](leaderboard/llvm_instcount/gatv2_ddppo/result.csv) | 2022-06 | 258.149s        | 1.047×                       |

| Facebook                                         | Random search (t=60) | [write-up](leaderboard/llvm_instcount/random_search/README.md), [results](leaderboard/llvm_instcount/random_search/results_p125_t60.csv) | 2021-03 | 91.215s         | 1.045×                       |

| Facebook                                         | e-Greedy search (e=0.1) | [write-up](leaderboard/llvm_instcount/e_greedy/README.md), [results](leaderboard/llvm_instcount/e_greedy/results_e10.csv) | 2021-03 | 351.611s        | 1.041×                       |

| Jiadong Guo                                      | Tabular Q (N=5000, H=10) | [write-up](leaderboard/llvm_instcount/tabular_q/README.md), [results](leaderboard/llvm_instcount/tabular_q/results-H10-N5000.csv) | 2021-04 | 2534.305        | 1.036×                       |

| Facebook                                         | Random search (t=10) | [write-up](leaderboard/llvm_instcount/random_search/README.md), [results](leaderboard/llvm_instcount/random_search/results_p125_t10.csv) | 2021-03 | 42.939s         | 1.031×                       |

| Patrick Hesse                                    | DQN (N=4000, H=10) | [write-up](leaderboard/llvm_instcount/dqn/README.md), [results](leaderboard/llvm_instcount/dqn/results-instcountnorm-H10-N4000.csv) | 2021-06 | 91.018s         | 1.029×                       |

| Jiadong Guo                                      | Tabular Q (N=2000, H=5) | [write-up](leaderboard/llvm_instcount/tabular_q/README.md), [results](leaderboard/llvm_instcount/tabular_q/results-H5-N2000.csv) | 2021-04 | 694.105         | 0.988×                       |

## Contributing

We welcome contributions to CompilerGym. If you are interested in contributing please see

[this document](https://github.com/facebookresearch/CompilerGym/blob/development/CONTRIBUTING.md).

## Citation

If you use CompilerGym in any of your work, please cite [our

paper](https://arxiv.org/pdf/2109.08267.pdf):

```

@inproceedings{CompilerGym,

      title={{CompilerGym: Robust, Performant Compiler Optimization Environments for AI Research}},

      author={Chris Cummins and Bram Wasti and Jiadong Guo and Brandon Cui and Jason Ansel and Sahir Gomez and Somya Jain and Jia Liu and Olivier Teytaud and Benoit Steiner and Yuandong Tian and Hugh Leather},

      booktitle={CGO},

      year={2022},

}

```