A Custom Precision Based Architecture for Accelerating Parallel Tempering MCMC on FPGAs without Introducing Sampling Error

Markov Chain Monte Carlo (MCMC) is a method used to draw samples from probability distributions in order to estimate - otherwise intractable - integrals. When the distribution is complex, simple MCMC becomes inefficient and advanced, computationally intensive MCMC methods are employed to make sampli...

Full description

Saved in:

Bibliographic Details
Published in:	2012 IEEE 20th International Symposium on Field-Programmable Custom Computing Machines pp. 153 - 156
Main Authors:	Mingas, G., Bouganis, C-S
Format:	Conference Proceeding
Language:	English
Published:	IEEE 01-04-2012
Subjects:	Acceleration Bayesian methods Computer architecture Field programmable gate arrays FPGA Markov Chain Monte Carlo Parallel Tempering Pipeline processing Pipelines Throughput
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Markov Chain Monte Carlo (MCMC) is a method used to draw samples from probability distributions in order to estimate - otherwise intractable - integrals. When the distribution is complex, simple MCMC becomes inefficient and advanced, computationally intensive MCMC methods are employed to make sampling possible. This work proposes a novel streaming FPGA architecture to accelerate Parallel Tempering, a widely adopted MCMC method designed to sample from multimodal distributions. The proposed architecture demonstrates how custom precision can be intelligently employed without introducing sampling errors, in order to save resources and increase the sampling throughg put. Speedups of up to two orders of magnitude compared to software and 1.53x-76.88x compared to a GPGPU implementation are achieved when performing Bayesian inference for a mixture model.
ISBN:	9781467316057 1467316059
DOI:	10.1109/FCCM.2012.34