MUGEN: State-aware CPU fuzzing via mutation-integrated instruction generation

2026-08-08

PRE

Jinwei Hu

Zihui Guo *

Ruiqing Li

Zhijie Zhang

He Wang

崔

崔宁宁 (Ningning Cui)

Liwei Chen

Gang Shi

DOI：10.1016/j.cose.2026.105098

Original

Original request for help

Save

Abstract

En 中文

CPU fuzzing has emerged as a promising automated verification approach. However, our analysis shows that state-of-the-art CPU fuzzing techniques remain limited in exploring diverse CPU architectural behaviors, primarily because they apply CPU state information only outside the generation loop rather than during instruction construction itself. This limitation results in severe instruction semantic redundancy, imbalanced opcode distributions, and unstable memory operations, ultimately hindering the efficiency of architectural exploration.

MUGEN: State-aware CPU fuzzing via mutation-integrated instruction generation

Abstract

Journal

Organization

Cited Papers

Citing Papers