WarpDrive: Extremely Fast End-to-End Deep Multi-Agent Reinforcement Learning on a GPU

WarpDrive is a flexible, lightweight, and easy-to-use open-source reinforcement learning (RL) framework that implements end-to-end multi-agent RL on a single GPU (Graphics Processing Unit).

Using the extreme parallelization capability of GPUs, WarpDrive enables orders-of-magnitude faster RL compared to CPU simulation + GPU model implementations. It is extremely efficient as it avoids back-and-forth data copying between the CPU and the GPU, and runs simulations across multiple agents and multiple environment replicas in parallel. Together, this allows the user to run thousands of concurrent multi-agent simulations and train on extremely large batches of experience, achieving over 100x throughput over CPU-based counterparts.

Our current release includes several multi-agent environments based on the game of "Tag", where taggers are trying to run after and tag the runners. More environments will be added soon!

Below, we show multi-agent RL policies trained for different tagger:runner speed ratios using WarpDrive. These environments can run at millions of steps per second, and train in just a few hours, all on a single GPU!

WarpDrive also provides tools to build and train multi-agent RL systems quickly with just a few lines of code. Here is a short example to train tagger and runner agents:

# Create a wrapped environment object via the EnvWrapper# Ensure that use_cuda is set to True (in order to run on the GPU)env_wrapper = EnvWrapper( TagContinuous(**run_config["env"]), num_envs=run_config["trainer"]["num_envs"], use_cuda=True) # Agents can share policy models: this dictionary maps policy model names to agent ids.policy_tag_to_agent_id_map = { "tagger": list(env_wrapper.env.taggers), "runner": list(env_wrapper.env.runners), } # Create the trainer objecttrainer = Trainer( env_wrapper=env_wrapper, config=run_config, policy_tag_to_agent_id_map=policy_tag_to_agent_id_map, ) # Create and push data placeholders to the devicecreate_and_push_data_placeholders( env_wrapper, policy_tag_to_agent_id_map, training_batch_size_per_env=trainer.training_batch_size_per_env) # Perform training!trainer.train()

White Paper and Citing WarpDrive

You can find more details in our white paper: https://arxiv.org/abs/2108.13976.

If you're using WarpDrive in your research or applications, please cite using this BibTeX:

      title={WarpDrive: Extremely Fast End-to-End Deep Multi-Agent Reinforcement Learning on a GPU}, 
      author={Tian Lan and Sunil Srinivasa and Stephan Zheng},

Tutorials and Quick Start

Familiarize yourself with WarpDrive by running these tutorials on Colab!

Note: You may also run these tutorials locally, but you will need a GPU machine with nvcc compiler installed and a compatible Nvidia GPU driver. You will also need Jupyter. See https://jupyter.readthedocs.io/en/latest/install.html for installation instructions

You can find full reference documentation here.

Installation Instructions

To get started, you'll need to have Python 3.7+ and the nvcc compiler installed with a compatible Nvidia GPU CUDA driver.

CUDA (which includes nvcc) can be installed by following Nvidia's instructions here: https://developer.nvidia.com/cuda-downloads.

Docker Image

You can refer to the example Dockerfile to configure your system. In particular, we suggest you visit Nvidia Docker Hub to download the CUDA and cuDNN images compatible with your system. You should be able to use the command line utility to monitor the NVIDIA GPU devices in your system:


and see something like this

| NVIDIA-SMI 450.51.06    Driver Version: 450.51.06    CUDA Version: 11.0     |
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|   0  Tesla V100-SXM2...  Off  | 00000000:00:04.0 Off |                    0 |
| N/A   37C    P0    32W / 300W |      0MiB / 16160MiB |      0%      Default |
|                               |                      |                  N/A |

| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|  No running processes found                                                 |

In this snapshot, you can see we are using a Tesla V100 GPU and CUDA version 11.0.

Installing using Pip

You can install WarpDrive using the Python package manager:

pip install rl_warp_drive

Installing from Source

Clone this repository to your machine:

git clone www.github.com/salesforce/warp-drive

Optional, but recommended for first tries: Create a new conda environment (named "warp_drive" below) and activate it:

conda create --name warp_drive python=3.7 --yes
conda activate warp_drive

Install as an editable Python package:

cd warp_drive
pip install -e .

Testing your Installation

To test your installation, try running from the root directory:

conda activate warp_drive
cd warp_drive/cuda_includes
make compile-test

Running make compile-test will compile the core service source code into a CUDA binary and place it in a bin folder, and additionally, run some unit tests.