Skip to content
STIMSMITH

Hardware-Based Mutation

Technique WIKI v1 · 5/27/2026

Hardware-Based Mutation is a technique reported in the Instiller RTL fuzzing work. The evidence describes it as part of Instiller’s hardware-based seed selection and mutation strategies, using hardware-related heuristics and mutation operations to improve fuzzing performance for RTL fuzzing.

Hardware-Based Mutation

Hardware-Based Mutation is a technique described in the Instiller RTL fuzzing work. In the cited paper, it appears as part of “hardware-based seed selection and mutation strategies” intended to improve fuzzing performance in RTL fuzzing by using hardware-related heuristics and mutation operations.[1]

Context

Instiller is presented as an RTL fuzzer for CPU-oriented hardware verification. The paper motivates RTL fuzzing by noting that hardware bugs must be detected before deployment, and that previous CPU fuzzing approaches face issues such as growing RTL input-instruction length and ineffective longer inputs.[2]

Technique role

Within Instiller, Hardware-Based Mutation is one of several proposed techniques. The paper lists three major technique areas: input-instruction distillation based on a variant of ant colony optimization, handling multiple interruptions and exceptions with priorities, and hardware-based seed selection and mutation strategies.[3]

The available evidence does not provide the detailed mutation operators themselves, but it does state the design intent: hardware-related heuristics and mutation operations are used to improve fuzzing performance in the RTL-fuzzing setting.[1]

Reported system-level results

The Instiller paper reports that the prototype outperforms prior work in experiments on real-world target CPU cores, including 29.4% more coverage than DiFuzzRTL and 17.0% more detected mismatches. These results are reported for Instiller as a whole, not isolated specifically to Hardware-Based Mutation.[4]

[1]: Hardware-Based Mutation description. [2]: Instiller RTL fuzzing context. [3]: Instiller contribution list. [4]: Instiller reported evaluation results.

LINKED ENTITIES

1 links

CITATIONS

4 sources
4 citations
[1] Hardware-Based Mutation is described as part of Instiller’s hardware-based seed selection and mutation strategies for RTL fuzzing, using hardware-related heuristics and mutation operations to improve fuzzing performance. [2401.15967] Instiller: Towards Efficient and Realistic RTL Fuzzing
[2] Instiller is an RTL fuzzer for hardware/CPU bug detection, and the paper identifies growing RTL input-instruction length and ineffective longer inputs as problems in prior CPU fuzzing work. [2401.15967] Instiller: Towards Efficient and Realistic RTL Fuzzing
[3] The Instiller paper lists hardware-based seed selection and mutation strategies as one of its contributions, alongside input-instruction distillation and realistic interruption/exception handling. [2401.15967] Instiller: Towards Efficient and Realistic RTL Fuzzing
[4] Instiller reports 29.4% more coverage than DiFuzzRTL, 17.0% more detected mismatches, 79.3% shorter input instructions with VACO, and a 6.7% average execution-speed increase from distillation. [2401.15967] Instiller: Towards Efficient and Realistic RTL Fuzzing