VeriEQ: Finding Verilog Simulators and Synthesizers Bugs with Equivalence Circuit Transformation
Zhen Yan, Yuanliang Chen, Fuchen Ma, Zehong Yu, Dalong Shi, Yu Jiang
Abstract
Verilog simulators and synthesizers play a critical role in chip design and verification. However, due to the complexity of simulation and synthesis processes, they easily introduce various types of bugs. Among them, Behavioral Deviation Bugs (BDBs) are particularly severe, as they can cause incorrect results by introducing subtle semantic deviations that make the chip behave differently from its intended design, potentially enabling hardware backdoors.
In this work, we propose VeriEQ, an automated framework based on the idea of metamorphic testing, which detects BDBs by generating semantically equivalent Verilog programs. First, to increase the likelihood of triggering BDB, we analyze the structural patterns of historical BDB and design a Verilog code template. Second, we generate semantically equivalent variants by applying equivalence circuit transformation rules. These rules include constraints on bit-width and signedness to ensure logical consistency before and after the transformation. Finally, we design an inlined deviation checking mechanism that embeds multiple equivalent modules within a single testbench to improve testing efficiency. We implement and evaluate VeriEQ on four mainstream Verilog simulators and synthesizer. Experimental results show that VeriEQ achieves a 138.1% to 4161.9% speedup over state-of-the-art tools. In total, VeriEQ successfully detects 33 previously unknown bugs, including 29 BDBs, along with 4 hang bugs as additional findings. All discovered bugs have been confirmed, with 27 already fixed. In contrast, the other tools are able to detect only 1 to 7 bugs.
CCS Concepts: • Software and its engineering → Software testing and debugging.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext e1524c25-0547-4069-830f-2a1e9c4e4b0cBuilds on7
- Alive2: bounded translation validation for LLVMNuno P. Lopes, Juneyoung Lee, Chung-Kil Hur, Zhengyang Liu et al.PLDI 2021 · 109 citations
- A compiler infrastructure for accelerator generatorsRachit Nigam, Samuel Thomas, Zhijing Li, Adrian SampsonASPLOS 2021 · 54 citations
- LLHD: a multi-level intermediate representation for hardware description languagesFabian Schuiki, Andreas Kurth, Tobias Grosser, Luca BeniniPLDI 2020 · 35 citations
- SymRTLO: Enhancing RTL Code Optimization with LLMs and Neuron-Inspired Symbolic ReasoningYiting Wang, Wanghao Ye, Ping Guo, Yexiao He et al.NeurIPS 2025 · 24 citations
- Statfier: Automated Testing of Static Analyzers via Semantic-Preserving Program TransformationsHuaien Zhang, Yu Pei, Junjie Chen, Shin Hwei TanFSE 2023 · 15 citations
Related papers
- SEPE-SQED: Symbolic Quick Error Detection by Semantically Equivalent Program ExecutionYufeng Li, Qiusong Yang, Yiwei Ci, Enyuan TianDAC 2024 · 1 citation
- Towards Understanding the Bugs in Verilator, a Hardware Description Language CompilerSongyan Jiang, Maolin Sun, Kang Chen, Qingyang Li et al.ISSTA 2026
- Exploiting Sophisticated Static Analysis for VerilogQinlin Chen, Nairen Zhang, Jinpeng Wang, Jiacai Cui et al.PLDI 2026
- VerilogLAVD: LLM-Aided Pattern Generation for Verilog CWE DetectionXiang Long, Yingjie Xia, Li Kuang, Yao Wan et al.ACL 2026
- Location is Key: Leveraging LLM for Functional Bug Localization in Verilog DesignBingkun Yao, Ning Wang, Jie Zhou, Xi Wang et al.DAC 2025 · 9 citations
