Characterizing Real-World Bugs in Tile Programs for Automated Bug Detection

TL;DR

This paper systematically studies code generation bugs in tile-based GPU programming frameworks, analyzing 301 bugs to inform future debugging and testing tools for these systems.

Contribution

It is the first comprehensive analysis of tile-program code generation bugs, providing insights into their causes, symptoms, and fixes based on 301 real bugs from GitHub.

Findings

Identified common root causes and symptoms of tile-program bugs

Categorized input patterns and test oracles that trigger bugs

Provided strategies used to fix these bugs

Abstract

Tile-based programming frameworks are increasingly adopted to write high-performance GPU kernels in domains such as deep learning and scientific computing. While these frameworks enhance productivity and hardware utilization, their multi-stage compilation pipelines introduce distinct code generation bugs that are tightly coupled to input shapes, data types, and backend targets. These bugs often manifest as silent correctness or performance issues, making them difficult to detect using existing compiler testing tools. Additionally, the unique programming conventions of tile domain-specific languages complicate root cause identification, while fixing such bugs demands specialized knowledge of tile abstractions and compilation pipelines. Despite the growing adoption of tile-based systems, their code generation bugs remain largely unexplored. This paper presents the first systematic study of tile-program code generation bugs. We curate 401 bug reports from GitHub and identify 301 tile-program codegen bugs for analysis, categorizing the root causes, symptoms, input patterns, test oracles that trigger these bugs, and the strategies used to fix bugs. Our study provides foundational insights for building debugging, testing, and repair tools tailored to tile-based compiler infrastructures.