GPU-Accelerated Loopy Belief Propagation for Program Analysis

Haoyu Feng, Xin Zhang · 2025

Read GPU-Accelerated Loopy Belief Propagation for Program Analysis by Haoyu Feng, Xin Zhang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-22337.txt.

Description

Haoyu Feng and Xin Zhang speed up a key inference method for program analysis. Belief propagation passes messages on a graph to guess unknown values, but it runs slowly on large programs. Their system, FastLBP, moves the work to graphics chips. A shared format lets users state flexible update plans that older GPU tools cannot express.

A dependency check finds which updates can run in parallel. Tests on two fault-finding tools showed large gains. FastLBP ran many times faster than CPU versions and faster than a prior GPU tool. Accuracy stayed the same, and the tool handled update plans the older GPU tool could not.

Catalog record

CREATORHaoyu Feng, Xin Zhang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE93 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-22337

Browse the full library · About the Tower of Babel