GRAPHSPY: Fused Program Semantic-Level Embedding via Graph Neural Networks for Dead Store Detection

TL;DR

This paper introduces GRAPHSPY, a graph neural network-based method for detecting unnecessary memory operations in software, achieving high accuracy with low overhead by leveraging program semantics.

Contribution

It presents a novel hybrid program embedding approach using graph neural networks to identify dead store operations efficiently.

Findings

Achieves 90% accuracy in dead store detection.

Reduces detection overhead to about half of existing tools.

Effectively models program semantics for low-overhead analysis.

Abstract

Production software oftentimes suffers from the issue of performance inefficiencies caused by inappropriate use of data structures, programming abstractions, and conservative compiler optimizations. It is desirable to avoid unnecessary memory operations. However, existing works often use a whole-program fine-grained monitoring method with incredibly high overhead. To this end, we propose a learning-aided approach to identify unnecessary memory operations intelligently with low overhead. By applying several prevalent graph neural network models to extract program semantics with respect to program structure, execution order and dynamic states, we present a novel, hybrid program embedding approach so that to derive unnecessary memory operations through the embedding. We train our model with tens of thousands of samples acquired from a set of real-world benchmarks. Results show that our model achieves 90% of accuracy and incurs only around a half of time overhead of the state-of-art tool.