AI Paper Atlas

Research map / Multi-agent systems and games / Cooperative multi-agent RL

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

Tian Lan, Sunil Srinivasa, Huan Wang, Stephan Zheng

JMLR 2022 · Multi-agent systems and games › Cooperative multi-agent RL · approach: MARL benchmarks and grouping

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
scale number of agents from 5 to 1000 in discrete TagWarpDrive · Discrete Tag (2000 envs)training iterations per second: 1.3 million -> 0.58 millionHurt
run RL on a single A100 GPU instead of a 16-CPU nodeWarpDrive · COVID-19 economic simulationsteps per second: 24x more steps per secondImproved
run continuous Tag training on GPU rather than 16-CPU nodeWarpDrive · Continuous Tag, 12 agents, 60 envstraining speed: 5x fasterImproved
use the WarpDrive CUDA action sampler instead of the PyTorch operatorWarpDrive · Discrete Tag, 2000 envssamples per second: 5 million -> 18 million; 3.6x fasterImproved
simulate environments on GPU instead of single-CPU NumPyWarpDrive · Discrete Tag per environmentthroughput: at least 50x fasterImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy