AcMC²: Accelerated Markov Chain Monte Carlo for Probabilistic Models

Subho S. Banerjee, Zbigniew T. Kalbarczyk, and Ravishankar K. Iyer

ASPLOS 2019



Abstract

Probabilistic models (PMs) are ubiquitously used across a variety of machine learning applications. They have been shown to successfully integrate structural prior information about data and effectively quantify uncertainty to enable the development of more powerful, interpretable, and efficient learning algorithms. This paper presents AcMC², a compiler that transforms PMs into optimized hardware accelerators (for use in FPGAs or ASICs) that utilize Markov chain Monte Carlo methods to infer and query a distribution of posterior samples from the model. The compiler analyzes statistical dependencies in the PM to drive several optimizations to maximally exploit the parallelism and data locality available in the problem. We demonstrate the use of AcMC² to implement several learning and inference tasks on a Xilinx Virtex-7 FPGA. AcMC²-generated accelerators provide a 47−100× improvement in runtime performance over a 6-core IBM Power8 CPU and a 8−18× improvement over an NVIDIA K80 GPU. This corresponds to a 753−1600× improvement over the CPU and 248−463× over the GPU in performance-per-watt terms.

Citation

@inproceedings{Banerjee_ASPLOS2019,
  author = {Banerjee, Subho S. and Kalbarczyk, Zbigniew T. and Iyer, Ravishankar K.},
  title = {AcMC2 : Accelerating Markov Chain Monte Carlo Algorithms for Probabilistic Models},
  booktitle = {Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems},
  series = {ASPLOS '19},
  year = {2019},
  isbn = {978-1-4503-6240-5},
  location = {Providence, RI, USA},
  pages = {515--528},
  numpages = {14},
  url = {http://doi.acm.org/10.1145/3297858.3304019},
  doi = {10.1145/3297858.3304019},
  acmid = {3304019},
  publisher = {ACM},
  address = {New York, NY, USA},
} 

Related Projects

  • Powered by Hugo
  • Last updated 10/21/2021
  • Feed