MAGMA

This is the implementation of the paper MAGMA, youtube.

Installation

• Install requirement

pip install -r requirements.txt

• Download cost model and build symbolic link

python build.py

Example Usage

• Run MAGMA: sh run/runGA.sh
• Run RL: sh run/runRL.sh
  • Available RLs: A2C, ACKTR, PPO2, DQN, TRPO, ACER, SAC DDPG
• Run Blackbox: sh run/run_blackbox.sh
  • Avaliable Blackbox: PSO, Portfolio, OnePlusOne,CMA, DE, NaiveTBPSA, cGA, CauchyLHSSearch, HaltonSearch, HammersleySearch, MetaRecentering

Contributor

• Sheng-Chun (Felix) Kao
• Tushar Krishna

Citation

@inproceedings{kao2022magma,
  title={MAGMA: An Optimization Framework for Mapping Multiple DNNs on Multiple Accelerator Cores},
  author={Kao, Sheng-Chun and Krishna, Tushar},
  booktitle={2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA)},
  pages={814--830},
  year={2022},
  organization={IEEE}
}