Profile-guided System Optimizations for Accelerated Greybox Fuzzing

Abstract

Greybox fuzzing is a highly popular option for security testing, incentivizing tremendous efforts to improve its performance. Prior research has brought many algorithmic advancements, leading to substantial performance growth. However, less attention has been paid to the system-level designs of greybox fuzzing tools, despite the high impacts of such designs on fuzzing throughput.