pith. sign in

arxiv: 2606.23759 · v1 · pith:K2EHVWRGnew · submitted 2026-06-22 · 💻 cs.AR · cs.AI· cs.SE

VeriPilot: An LLM-Powered Verilog Debugging Framework

Yihan Wang , Cheng Liu , Jiazheng Zhang , Lei Zhang , Long Cheng , Xiaowei Li , Huawei Li This is my paper

Pith reviewed 2026-06-26 06:18 UTC · model grok-4.3

classification 💻 cs.AR cs.AIcs.SE
keywords Verilog debuggingLLM-based repairgolden reference modelcontrol-data-flow graphbug localizationautomated debuggingdigital circuit designsignal tracing
0
0 comments X

The pith

VeriPilot aligns internal variables between buggy Verilog and golden models then traces signals via CDFGs to give LLMs minimal suspicious regions for repair.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper claims that current LLM debugging for Verilog fails on complex bugs because root causes sit far from observable outputs and long contexts overwhelm reasoning. VeriPilot addresses this by using a golden reference model to match the meaning of internal variables and then walking control-data-flow graphs to isolate the smallest set of faulty code segments together with their correct counterparts. These targeted facts are handed to the LLM so it can reason step by step instead of guessing from test results alone. On NVIDIA's CVDP benchmark the approach lifts GPT-4o's repair success from 54.3 percent to 85.71 percent.

Core claim

VeriPilot uses golden reference models to align internal variable semantics between a buggy Verilog design and its correct version, then applies static analysis to produce control-data-flow graphs that identify a minimal set of suspicious code regions along with their correct counterparts; these structured insights are supplied to an LLM to guide localization and automated repair.

What carries the argument

LLM-based semantic alignment of internal variables combined with CDFG-derived step-by-step signal tracing that isolates minimal suspicious regions from a golden reference model.

If this is right

  • Bug localization accuracy rises because the method compares internal signals rather than only final outputs.
  • Repair effectiveness improves on designs with long dependency chains that defeat output-only prompting.
  • Large codebases become tractable because the framework reduces the effective context the LLM must reason over.
  • The same golden-model-plus-CDFG pattern can be reused across multiple LLM back-ends without retraining.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The technique could transfer to other hardware description languages if equivalent golden models and static-analysis front-ends are supplied.
  • Combining the CDFG traces with existing formal verification tools might further shrink the suspicious region set.
  • Design teams without ready golden models might first generate them via high-level synthesis or simulation to unlock the framework.

Load-bearing premise

Accurate golden reference models exist for the designs and the LLM can reliably match the semantics of internal variables between the buggy code and the reference.

What would settle it

Apply VeriPilot to the same benchmark designs but withhold the golden reference models and measure whether the repair success rate drops back to the 54.3 percent baseline.

Figures

Figures reproduced from arXiv: 2606.23759 by Cheng Liu, Huawei Li, Jiazheng Zhang, Lei Zhang, Long Cheng, Xiaowei Li, Yihan Wang.

Figure 1
Figure 1. Figure 1: Overview of the manual debugging methodology. The process initiates [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: VeriPilot is an LLM-powered multi-agent Verilog debugging framework. It consists of four phases: (1) simulation & verification to identify the earliest [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Progressive bug localization procedure in VeriPilot. (a) State frontier extraction identifies the earliest mismatched sequential states based on temporal [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Bug repair prompt template used by VeriPilot, which integrates [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Semantic alignment failure in the counter6bit design. The golden [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 5
Figure 5. Figure 5: The golden reference maintains a single binary counter [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Repair ambiguity after successful bug localization. The localized [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
read the original abstract

Verilog debugging remains one of the most time-consuming stages in digital circuit design. Recent advances in Large Language Models (LLMs) have enabled automated debugging; however, most existing approaches rely solely on test outputs and compiler feedback in an end-to-end manner, limiting their effectiveness on complex bugs. A key challenge is that the root cause of an error may be far removed from its observable outputs, making it difficult for LLMs to trace long dependency chains in code. This challenge is further exacerbated in large codebases, where long context lengths hinder efficient reasoning. To address these limitations, we propose VeriPilot, an LLM-powered debugging framework that leverages golden reference models to enable fine-grained bug localization and repair. VeriPilot goes beyond output-level comparison by aligning internal variable semantics between the Verilog design and its corresponding golden model through LLM-based analysis. It then performs step-by-step signal tracing using Control-Data-Flow Graphs (CDFGs) derived from static analysis, identifying a minimal set of suspicious code regions along with their correct counterparts from the golden model. These structured insights are subsequently provided to the LLM to guide reasoning and automated code repair. Experimental results on the Comprehensive Verilog Design Problems (CVDP) benchmark from NVIDIA demonstrate that VeriPilot improves the repair success rate of GPT-4o from 54.3\% to 85.71\%, significantly enhancing both bug localization accuracy and repair effectiveness for complex Verilog designs. The source code and benchmark are publicly available at Github https://github.com/YihanWn/VeriPilot.git.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces VeriPilot, an LLM-powered framework for Verilog debugging that aligns internal variables between a buggy design and a golden reference model via LLM analysis, performs CDFG-guided tracing to localize suspicious regions, and supplies the resulting structured insights to an LLM for automated repair. On the CVDP benchmark it reports raising GPT-4o repair success from 54.3% to 85.71%.

Significance. If the central performance claim is substantiated, the work would demonstrate a practical way to leverage golden models and static-analysis structure to improve LLM reasoning over long dependency chains in hardware designs. The public release of code and benchmark is a clear strength that supports reproducibility and follow-on work.

major comments (2)
  1. [Abstract and §4] Abstract and §4 (Experimental Results): the headline improvement (54.3% → 85.71%) is produced by a pipeline whose first two stages are LLM-based variable alignment and CDFG-guided tracing that depends on those alignments. No quantitative accuracy, precision, or failure-case analysis is supplied for the alignment step, so it is impossible to determine whether the measured gain can be attributed to the claimed mechanism rather than to noisy or incorrect alignments.
  2. [§4] §4 (Experimental Results): no ablation is reported that removes the golden-model path or the alignment/tracing components while keeping the final repair LLM fixed. Without such controls the contribution of each stage to the reported success rate remains unquantified.
minor comments (2)
  1. [Abstract] Abstract: the GitHub URL is given but the manuscript should state the exact commit or release tag used for the reported experiments.
  2. [§3] §3 (Method): the description of how CDFGs are extracted from Verilog and how variable alignments are injected into the repair prompt could be expanded with a short pseudocode or example to improve reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which highlight opportunities to strengthen the experimental section. We address each major comment below and will incorporate the requested analyses into a revised manuscript.

read point-by-point responses

  1. Referee: [Abstract and §4] Abstract and §4 (Experimental Results): the headline improvement (54.3% → 85.71%) is produced by a pipeline whose first two stages are LLM-based variable alignment and CDFG-guided tracing that depends on those alignments. No quantitative accuracy, precision, or failure-case analysis is supplied for the alignment step, so it is impossible to determine whether the measured gain can be attributed to the claimed mechanism rather than to noisy or incorrect alignments.

    Authors: We agree that the manuscript lacks a dedicated quantitative evaluation of the LLM-based variable alignment step. In the revision we will add a new subsection reporting alignment accuracy, precision, and recall on a manually annotated subset of CVDP cases, together with a discussion of observed failure modes and their downstream effect on repair success. revision: yes

  2. Referee: [§4] §4 (Experimental Results): no ablation is reported that removes the golden-model path or the alignment/tracing components while keeping the final repair LLM fixed. Without such controls the contribution of each stage to the reported success rate remains unquantified.

    Authors: We concur that component-wise ablations are necessary to isolate contributions. The revised manuscript will include controlled experiments that disable the golden-model alignment and CDFG tracing stages while retaining the identical repair LLM and test-feedback input, thereby quantifying the incremental benefit of each stage. revision: yes

Circularity Check

0 steps flagged

No circularity; empirical results on external benchmark

full rationale

The paper proposes an LLM-powered Verilog debugging framework and reports an empirical performance gain (54.3% to 85.71% repair success) on the external CVDP benchmark. No equations, fitted parameters, or derivations are present. The method description relies on golden reference models and LLM variable alignment as design choices, but these do not reduce any claimed result to a self-definition or self-citation chain. The central claim rests on reported experimental outcomes rather than internal re-labeling of inputs. This matches the default expectation of a non-circular empirical systems paper.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Based solely on the abstract; the approach relies on domain assumptions about reference models rather than new mathematical constructs.

axioms (1)
  • domain assumption Golden reference models exist and can be used for semantic alignment with the design under debug.
    Central to the fine-grained bug localization and repair approach.

pith-pipeline@v0.9.1-grok · 5822 in / 1211 out tokens · 40312 ms · 2026-06-26T06:18:57.475038+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

46 extracted references · 12 canonical work pages · 1 internal anchor

  1. [1]

    Verigen: A large language model for verilog code generation.ACM Transactions on Design Automation of Electronic Systems, 29(3):1–31, 2024

    Shailja Thakur, Baleegh Ahmad, Hammond Pearce, Benjamin Tan, Brendan Dolan-Gavitt, Ramesh Karri, and Siddharth Garg. Verigen: A large language model for verilog code generation.ACM Transactions on Design Automation of Electronic Systems, 29(3):1–31, 2024

    2024

  2. [2]

    Software/hardware co-design for llm and its application for design verification

    Lily Jiaxin Wan, Yingbing Huang, Yuhong Li, Hanchen Ye, Jinghua Wang, Xiaofan Zhang, and Deming Chen. Software/hardware co-design for llm and its application for design verification. In2024 29th Asia and South Pacific Design Automation Conference (ASP-DAC), pages 435–441. IEEE, 2024

    2024

  3. [3]

    Codev: Empowering llms for verilog generation through multi-level summariza- tion.arXiv preprint arXiv:2407.10424, 2024

    Yang Zhao, Di Huang, Chongxiao Li, Pengwei Jin, Ziyuan Nan, Tianyun Ma, Lei Qi, Yansong Pan, Zhenxing Zhang, Rui Zhang, et al. Codev: Empowering llms for verilog generation through multi-level summariza- tion.arXiv preprint arXiv:2407.10424, 2024

    work page arXiv 2024

  4. [4]

    Mingjie Liu, Yun-Da Tsai, Wenfei Zhou, and Haoxing Ren. Craftrtl: High-quality synthetic data generation for verilog code models with correct-by-construction non-textual representations and targeted code repair.arXiv preprint arXiv:2409.12993, 2024

    work page arXiv 2024

  5. [5]

    Mg-verilog: Multi-grained dataset towards enhanced llm- assisted verilog generation

    Yongan Zhang, Zhongzhi Yu, Yonggan Fu, Cheng Wan, and Yingyan Ce- line Lin. Mg-verilog: Multi-grained dataset towards enhanced llm- assisted verilog generation. In2024 IEEE LLM Aided Design Workshop (LAD), pages 1–5. IEEE, 2024

    2024

  6. [6]

    Hivegen–hierarchical llm-based verilog generation for scalable chip design

    Jinwei Tang, Jiayin Qin, Kiran Thorat, Chen Zhu-Tian, Yu Cao, Yang Katie Zhao, and Caiwen Ding. Hivegen–hierarchical llm-based verilog generation for scalable chip design. In2025 IEEE International Conference on LLM-Aided Design (ICLAD), pages 30–36. IEEE, 2025

    2025

  7. [7]

    From large to small: Transferring cuda optimization expertise via reasoning graph, 2025

    Junfeng Gong, Zhiyi Wei, Junying Chen, Cheng Liu, and Huawei Li. From large to small: Transferring cuda optimization expertise via reasoning graph, 2025

    2025

  8. [8]

    Verilogeval: Evaluating large language models for verilog code genera- tion

    Mingjie Liu, Nathaniel Pinckney, Brucek Khailany, and Haoxing Ren. Verilogeval: Evaluating large language models for verilog code genera- tion. In2023 IEEE/ACM International Conference on Computer Aided Design (ICCAD), pages 1–8. IEEE, 2023

    2023

  9. [9]

    Large language model for verilog generation with code- structure-guided reinforcement learning

    Ning Wang, Bingkun Yao, Jie Zhou, Yuchen Hu, Xi Wang, Zhe Jiang, and Nan Guan. Large language model for verilog generation with code- structure-guided reinforcement learning. In2025 IEEE International Conference on LLM-Aided Design (ICLAD), pages 164–170. IEEE, 2025

    2025

  10. [10]

    Verilogreader: Llm-aided hardware test gener- ation

    Ruiyang Ma, Yuxin Yang, Ziqian Liu, Jiaxi Zhang, Min Li, Junhua Huang, and Guojie Luo. Verilogreader: Llm-aided hardware test gener- ation. In2024 IEEE LLM Aided Design Workshop (LAD), pages 1–5. IEEE, 2024

    2024

  11. [11]

    Large language models (llms) for verification, testing, and design

    Chandan Kumar Jha, Muhammad Hassan, Khushboo Qayyum, Sallar Ahmadi-Pour, Kangwei Xu, Ruidi Qiu, Jason Blocklove, Luca Collini, Andre Nakkab, Ulf Schlichtmann, et al. Large language models (llms) for verification, testing, and design. In2025 IEEE European Test Symposium (ETS), pages 1–10. IEEE, 2025

    2025

  12. [12]

    Autochip: Automating hdl generation using llm feedback.arXiv preprint arXiv:2311.04887, 2023

    Shailja Thakur, Jason Blocklove, Hammond Pearce, Benjamin Tan, Sid- dharth Garg, and Ramesh Karri. Autochip: Automating hdl generation using llm feedback.arXiv preprint arXiv:2311.04887, 2023

    work page arXiv 2023

  13. [13]

    Rtlfixer: Automatically fixing rtl syntax errors with large language model

    YunDa Tsai, Mingjie Liu, and Haoxing Ren. Rtlfixer: Automatically fixing rtl syntax errors with large language model. InProceedings of the 61st ACM/IEEE Design Automation Conference, pages 1–6, 2024

    2024

  14. [14]

    Towards llm-powered verilog rtl assistant: Self-verification and self-correction.arXiv preprint arXiv:2406.00115, 2024

    Hanxian Huang, Zhenghan Lin, Zixuan Wang, Xin Chen, Ke Ding, and Jishen Zhao. Towards llm-powered verilog rtl assistant: Self-verification and self-correction.arXiv preprint arXiv:2406.00115, 2024

    work page arXiv 2024

  15. [15]

    Meic: Re-thinking rtl debug automation using llms

    Ke Xu, Jialin Sun, Yuchen Hu, Xinwei Fang, Weiwei Shan, Xi Wang, and Zhe Jiang. Meic: Re-thinking rtl debug automation using llms. InProceedings of the 43rd IEEE/ACM International Conference on Computer-Aided Design, pages 1–9, 2024

    2024

  16. [16]

    Veridebug: A unified llm for verilog debugging via con- trastive embedding and guided correction

    Ning Wang, Bingkun Yao, Jie Zhou, Yuchen Hu, Xi Wang, Zhe Jiang, and Nan Guan. Veridebug: A unified llm for verilog debugging via con- trastive embedding and guided correction. In2025 IEEE International Conference on LLM-Aided Design (ICLAD), pages 61–67. IEEE, 2025

    2025

  17. [17]

    Location is key: Leveraging llm for functional bug local- ization in verilog design

    Bingkun Yao, Ning Wang, Jie Zhou, Xi Wang, Hong Gao, Zhe Jiang, and Nan Guan. Location is key: Leveraging llm for functional bug local- ization in verilog design. In2025 62nd ACM/IEEE Design Automation Conference (DAC), pages 1–7. IEEE, 2025

    2025

  18. [18]

    Springer, 2000

    Janick Bergeron.Writing testbenches: functional verification of HDL models. Springer, 2000

    2000

  19. [19]

    Automated program repair.Communications of the ACM, 62(12):56–65, 2019

    Claire Le Goues, Michael Pradel, and Abhik Roychoudhury. Automated program repair.Communications of the ACM, 62(12):56–65, 2019

    2019

  20. [20]

    A survey of learning-based automated program repair

    Quanjun Zhang, Chunrong Fang, Yuxiang Ma, Weisong Sun, and Zhenyu Chen. A survey of learning-based automated program repair. ACM Transactions on Software Engineering and Methodology, 33(2):1– 69, 2023. 13

    2023

  21. [21]

    A critical review on the evaluation of automated program repair systems.Journal of Systems and Software, 171:110817, 2021

    Kui Liu, Li Li, Anil Koyuncu, Dongsun Kim, Zhe Liu, Jacques Klein, and Tegawend ´e F Bissyand ´e. A critical review on the evaluation of automated program repair systems.Journal of Systems and Software, 171:110817, 2021

    2021

  22. [22]

    Understanding bug fix patterns in verilog

    Sangeetha Sudakrishnan, Janaki Madhavan, E James Whitehead Jr, and Jose Renau. Understanding bug fix patterns in verilog. InProceedings of the 2008 international working conference on Mining software repositories, pages 39–42, 2008

    2008

  23. [23]

    Cirfix: Automatically repairing defects in hardware design code

    Hammad Ahmad, Yu Huang, and Westley Weimer. Cirfix: Automatically repairing defects in hardware design code. InProceedings of the 27th ACM International Conference on Architectural Support for Program- ming Languages and Operating Systems, pages 990–1003, 2022

    2022

  24. [24]

    Deheng Yang, Jiayu He, Xiaoguang Mao, Tun Li, Yan Lei, Xin Yi, and Jiang Wu. Strider: Signal value transition-guided defect repair for hdl programming assignments.IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 43(5):1594–1607, 2023

    2023

  25. [25]

    Rtl-repair: Fast symbolic repair of hardware design code

    Kevin Laeufer, Brandon Fajardo, Abhik Ahuja, Vighnesh Iyer, Borivoje Nikoli´c, and Koushik Sen. Rtl-repair: Fast symbolic repair of hardware design code. InProceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3, pages 867–881, 2024

    2024

  26. [26]

    Verifix: Verified repair of programming assign- ments.ACM Transactions on Software Engineering and Methodology (TOSEM), 31(4):1–31, 2022

    Umair Z Ahmed, Zhiyu Fan, Jooyong Yi, Omar I Al-Bataineh, and Abhik Roychoudhury. Verifix: Verified repair of programming assign- ments.ACM Transactions on Software Engineering and Methodology (TOSEM), 31(4):1–31, 2022

    2022

  27. [27]

    Llm-aided efficient hardware design automation

    Kangwei Xu, Ruidi Qiu, Zhuorui Zhao, Grace Li Zhang, Ulf Schlicht- mann, and Bing Li. Llm-aided efficient hardware design automation. arXiv preprint arXiv:2410.18582, 2024

    work page arXiv 2024

  28. [28]

    Llm-assisted bug identification and correction for verilog hdl.ACM Transactions on Design Automation of Electronic Systems, 2025

    Khushboo Qayyum, Chandan Kumar Jha, Sallar Ahmadi-Pour, Muham- mad Hassan, and Rolf Drechsler. Llm-assisted bug identification and correction for verilog hdl.ACM Transactions on Design Automation of Electronic Systems, 2025

    2025

  29. [29]

    Evaluating llm-based communicative agents for verilog design

    Parker Link and Benjamin Tan. Evaluating llm-based communicative agents for verilog design. In2025 26th International Symposium on Quality Electronic Design (ISQED), pages 1–8. IEEE, 2025

    2025

  30. [30]

    Lintllm: An open-source verilog linting framework based on large language models.arXiv preprint arXiv:2502.10815, 2025

    Zhigang Fang, Renzhi Chen, Zhijie Yang, Yang Guo, Huadong Dai, and Lei Wang. Lintllm: An open-source verilog linting framework based on large language models.arXiv preprint arXiv:2502.10815, 2025

    work page arXiv 2025

  31. [31]

    SysVCoder: An LLM-Driven Framework for Systematic Generation of System-Level Design

    Jian Zuo, Junzhe Liu, Xianyong Wang, Yicheng Liu, Navya Goli, Tong Xu, Hao Zhang, Umamaheswara Rao Tida, Zhenge Jia, and Mengying Zhao. Complexvcoder: An llm-driven framework for systematic gener- ation of complex verilog code.arXiv preprint arXiv:2504.20653, 2025

    work page internal anchor Pith review arXiv 2025

  32. [32]

    A generalize hardware debugging approach for large language models semi-synthetic, datasets.IEEE Transactions on Circuits and Systems I: Regular Papers, 72(2):623–636, 2024

    Weimin Fu, Shijie Li, Yifang Zhao, Kaichen Yang, Xuan Zhang, Yier Jin, and Xiaolong Guo. A generalize hardware debugging approach for large language models semi-synthetic, datasets.IEEE Transactions on Circuits and Systems I: Regular Papers, 72(2):623–636, 2024

    2024

  33. [33]

    Verimind: Agentic llm for automated verilog generation with a novel evaluation metric

    Bardia Nadimi, Ghali Omar Boutaib, and Hao Zheng. Verimind: Agentic llm for automated verilog generation with a novel evaluation metric. arXiv preprint arXiv:2503.16514, 2025

    work page arXiv 2025

  34. [34]

    Coopetitivev: Leveraging llm-powered coopetitive multi- agent prompting for high-quality verilog generation.arXiv preprint arXiv:2412.11014, 2024

    Zhendong Mi, Renming Zheng, Haowen Zhong, Yue Sun, Seth Kneeland, Sayan Moitra, Ken Kutzer, and Zhaozhuo Xu Shaoyi Huang. Coopetitivev: Leveraging llm-powered coopetitive multi- agent prompting for high-quality verilog generation.arXiv preprint arXiv:2412.11014, 2024

    work page arXiv 2024

  35. [35]

    Automatic high-quality verilog assertion generation through subtask-focused fine-tuned llms and iterative prompting.arXiv preprint arXiv:2411.15442, 2024

    Mohammad Shahidzadeh, Behnam Ghavami, Steve Wilton, and Lesley Shannon. Automatic high-quality verilog assertion generation through subtask-focused fine-tuned llms and iterative prompting.arXiv preprint arXiv:2411.15442, 2024

    work page arXiv 2024

  36. [36]

    Hdldebugger: Streamlining hdl debugging with large language models.ACM Transactions on Design Automation of Electronic Systems, 30(6):1–26, 2025

    Xufeng Yao, Haoyang Li, Tsz Ho Chan, Wenyi Xiao, Mingxuan Yuan, Yu Huang, Lei Chen, and Bei Yu. Hdldebugger: Streamlining hdl debugging with large language models.ACM Transactions on Design Automation of Electronic Systems, 30(6):1–26, 2025

    2025

  37. [37]

    hdl2v: A code translation dataset for enhanced llm verilog generation.arXiv preprint arXiv:2506.04544, 2025

    Charles Hong, Brendan Roberts, Huijae An, Alex Um, Advay Ratan, and Yakun Sophia Shao. hdl2v: A code translation dataset for enhanced llm verilog generation.arXiv preprint arXiv:2506.04544, 2025

    work page arXiv 2025

  38. [38]

    On hardware security bug code fixes by prompting large language models.IEEE Transactions on Information Forensics and Security, 19:4043–4057, 2024

    Baleegh Ahmad, Shailja Thakur, Benjamin Tan, Ramesh Karri, and Hammond Pearce. On hardware security bug code fixes by prompting large language models.IEEE Transactions on Information Forensics and Security, 19:4043–4057, 2024

    2024

  39. [39]

    Effective simulation and debugging for a high-level hardware language using software compilers

    Cl ´ement Pit-Claudel, Thomas Bourgeat, Stella Lau, Arvind, and Adam Chlipala. Effective simulation and debugging for a high-level hardware language using software compilers. InProceedings of the 26th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, pages 789–803, 2021

    2021

  40. [40]

    Vfo- cus: Better verilog generation from large language model via focused reasoning.arXiv preprint arXiv:2511.02285, 2025

    Zhuorui Zhao, Bing Li, Grace Li Zhang, and Ulf Schlichtmann. Vfo- cus: Better verilog generation from large language model via focused reasoning.arXiv preprint arXiv:2511.02285, 2025

    work page arXiv 2025

  41. [41]

    Mage: A multi-agent engine for automated rtl code generation

    Yujie Zhao, Hejia Zhang, Hanxian Huang, Zhongming Yu, and Jishen Zhao. Mage: A multi-agent engine for automated rtl code generation. In2025 62nd ACM/IEEE Design Automation Conference (DAC), pages 1–7. IEEE, 2025

    2025

  42. [42]

    Verilogcoder: Autonomous verilog coding agents with graph-based planning and abstract syntax tree (ast)-based waveform tracing tool

    Chia-Tung Ho, Haoxing Ren, and Brucek Khailany. Verilogcoder: Autonomous verilog coding agents with graph-based planning and abstract syntax tree (ast)-based waveform tracing tool. InProceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 300– 307, 2025

    2025

  43. [43]

    Buggen: A self-correcting multi-agent llm pipeline for realistic rtl bug synthesis

    Surya Jasper, Minh Luu, Evan Pan, Aakash Tyagi, Michael Quinn, Jiang Hu, and David Houngninou. Buggen: A self-correcting multi-agent llm pipeline for realistic rtl bug synthesis. In2025 ACM/IEEE 7th Symposium on Machine Learning for CAD (MLCAD), pages 1–9. IEEE, 2025

    2025

  44. [44]

    Pyverilog: A python-based hardware design processing toolkit for verilog hdl

    Shinya Takamaeda-Yamazaki. Pyverilog: A python-based hardware design processing toolkit for verilog hdl. InInternational Symposium on Applied Reconfigurable Computing, pages 451–460. Springer, 2015

    2015

  45. [45]

    pyslang: Systemverilog frontend

    Mike Popoloski. pyslang: Systemverilog frontend. https://github.com/ MikePopoloski/slang, 2022

    2022

  46. [46]

    cocotb: Coroutine-based cosimulation library for writing vhdl and verilog testbenches in python

    cocotb contributors. cocotb: Coroutine-based cosimulation library for writing vhdl and verilog testbenches in python. https://github.com/ cocotb/cocotb, 2024

    2024