pith. sign in

arxiv: 2606.18421 · v1 · pith:YIINZT7Ynew · submitted 2026-06-16 · 💻 cs.SE

Finding Compiler-Platform Interaction Bugs in Deep Learning Pipelines via Cross-Layer Constraints

Yuxin Qiu , Jiyuan Wang , Ronak Badhe , Ben Limpanukorn , Miryung Kim , Qian Zhang This is my paper

Pith reviewed 2026-06-26 23:27 UTC · model grok-4.3

classification 💻 cs.SE
keywords deep learning compilerscompiler testingbug detectioncross-layer constraintsmodel generationbehavior monitoring
0
0 comments X

The pith

Extracting full-stack constraints across compilation passes and hardware platforms reveals 2,034 interaction bugs in deep learning compilers.

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

The paper seeks to establish that many bugs in deep learning compilers arise from violated assumptions in how compilation passes interact with each other and with target hardware platforms. It moves beyond type-based input restrictions to derive constraints that span the full stack, using them both to generate test models and to monitor distinct behaviors through inserted assertions. A sympathetic reader would care because existing approaches miss silent errors such as memory overflows and unexpected compilations that only surface under specific platform conditions. The evaluation on three compilers shows thousands of such cases triggered by prioritizing interaction-sensitive constraints.

Core claim

The central claim is that compiler-platform interaction bugs are caused by violated assumptions arising from interactions across compilation passes and hardware platforms, and that automatically extracting full-stack constraints to guide model generation, prioritize interaction-sensitive behaviors, and enable behavior equivalence partitioning via assertions will expose these bugs at scale.

What carries the argument

full-stack constraints extracted to jointly guide model generation and characterize compilation behaviors, with prioritization of those exposing interaction-sensitive behaviors and automatic insertion of assertions for monitoring

If this is right

  • DL compilers contain thousands of previously undetected bugs rooted in cross-pass and cross-platform interactions.
  • Behavior equivalence partitioning through assertions can detect symptoms that coverage metrics and pass/fail signals miss.
  • Prioritizing constraints that exercise deep compilation logic increases the ability to trigger interaction bugs during testing.

Where Pith is reading between the lines

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

  • The same constraint-extraction idea might help test other layered transformation systems where assumptions cross module boundaries.
  • If the extracted constraints prove stable across compiler versions, they could serve as a basis for regression testing suites focused on platform interactions.

Load-bearing premise

Bugs arise from violated assumptions in interactions across compilation passes and hardware platforms rather than from type mismatches alone.

What would settle it

Generating models with only type constraints and observing the same rate of memory overflows, integer overflows, and silent unexpected compilations as when full-stack constraints are used.

Figures

Figures reproduced from arXiv: 2606.18421 by Ben Limpanukorn, Jiyuan Wang, Miryung Kim, Qian Zhang, Ronak Badhe, Yuxin Qiu.

Figure 1
Figure 1. Figure 1: Overview of XCHECK. Unlike existing approaches [Liu et al.(2023a), Liu et al.(2023b), Wang et al.(2023), Ma et al.(2023), Deng et al.(2022), Mu et al.(2025)] that restrict inputs with local operator-level constraints and rely on coarse-grained pass/fail oracles, XCHECK extracts full-stack constraints (Section 3.1) to both drive deep compilation exploration (Section 3.2) and enable behavior equivalence part… view at source ↗
Figure 2
Figure 2. Figure 2: Constraint Extraction from Documentation. can be obtained from hardware specifications. Despite their importance, existing testing approaches focus only on operator-level constraints, leaving many overlooked. Observation 3: Existing testing oracles are coarse-grained and fail to differentiate compiler behaviors beyond crashes or acceptance, while constraints naturally define finer-grained outcome oracles. … view at source ↗
Figure 3
Figure 3. Figure 3: Example of Assertions to Insert. and output tensor properties. For example, to extract the input cardinality constraint, we count the occurrences of the pattern - <letter>+ ((<letter>*))? - T<digit>*: in Inputs section in Figure 2A, which is 1. Figure 2C shows the pattern matching results of Abs: the number of input tensors and output tensors should be both 1; the tensor type should be double, float, or in… view at source ↗
Figure 4
Figure 4. Figure 4: XCHECK injects assertions into the compiler for compilation behavioral monitoring. frequently used (LFU) one is chosen (Line 6). Next, XCHECK uses this constraint to guide the model generation. After that, XCHECK uses the target compiler to compile the generated model and records the number of compilation passes completed before a failure, which is denoted as p (Line 10). The rank r of the prioritized cons… view at source ↗
read the original abstract

The growing deployment of artificial intelligence (AI) necessitates robust deep learning (DL) compilers, such as TVM and ONNX-MLIR. These compilers take as input high-level AI models, lower them through multi-layer transformations, and specialize them to diverse hardware. Testing such compilers is uniquely challenging as correctness depends on implicit constraints embedded throughout the compilation stack. Existing testing approaches largely take type constraints to restrict input model generation and therefore emphasize type validation and monitor compilation crashes or coverage gains. This focus overlooks compiler-platform interaction bugs that arise from interleaved effects across compilation and execution environments. In this work, we propose a scalable, automated DL compiler testing framework for, in tandem, (1) finding compiler-platform interaction bugs and (2) enabling behavior equivalence partitioning. Our key insight is that these bugs are caused by violated assumptions arising from interactions across compilation passes and hardware platforms. Therefore, we move beyond constraining input generation and derive full-stack constraints. Our approach is three-fold. First, we design an automated approach to extract full-stack constraints that jointly guide model generation and characterize compilation behaviors. Second, we prioritize constraints that expose interaction-sensitive behaviors, so our generated models are capable of exercising deep compilation logic. Third, we enable behavior equivalence partitioning by automatically inserting assertions to monitor distinct compilation symptoms that coverage or pass/fail signals miss. We evaluated our tool, XCheck, on three widely-used DL compilers and found 2,034 bug-revealing cases, including memory overflows, integer overflows, and silent unexpected compilations that were rooted in compiler-platform interactions.

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 / 1 minor

Summary. The paper introduces XCheck, a testing framework for deep learning compilers (e.g., TVM, ONNX-MLIR) that extracts full-stack constraints across compilation passes and hardware platforms. These constraints guide model generation, prioritize interaction-sensitive behaviors, and insert assertions for behavior equivalence partitioning. The central empirical claim is that this approach revealed 2,034 bug-revealing cases (memory overflows, integer overflows, silent unexpected compilations) rooted in compiler-platform interactions on three widely-used compilers.

Significance. If the reported cases are confirmed as true positives with rigorous verification, the work would offer a practical advance over type-constraint-focused testing by targeting cross-layer interaction bugs that current methods miss. The emphasis on full-stack constraints and equivalence partitioning could improve coverage of deep compilation logic in DL pipelines, with potential for broader adoption in compiler testing.

major comments (2)
  1. [Abstract/Evaluation] Abstract and Evaluation: The central claim of 2,034 bug-revealing cases provides no details on the bug verification process, false-positive filtering criteria, or independent validation of the extracted constraints. This leaves the soundness of the reported bugs (and thus the effectiveness of the prioritization step) weakly supported by the presented evidence.
  2. [Approach] Approach description: The three-fold method (constraint extraction, prioritization, assertion insertion) is outlined at a high level, but without concrete examples of how a full-stack constraint is derived from a specific compilation pass and platform interaction or how prioritization scores are computed, it is difficult to assess whether the generated tests actually exercise the claimed deep logic.
minor comments (1)
  1. [Abstract] The abstract uses 'silent unexpected compilations' without defining the observable symptom or how it differs from a normal successful compilation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address the major points below and will revise the manuscript to improve clarity and evidence presentation.

read point-by-point responses

  1. Referee: [Abstract/Evaluation] Abstract and Evaluation: The central claim of 2,034 bug-revealing cases provides no details on the bug verification process, false-positive filtering criteria, or independent validation of the extracted constraints. This leaves the soundness of the reported bugs (and thus the effectiveness of the prioritization step) weakly supported by the presented evidence.

    Authors: We agree this is a valid concern and that the abstract and high-level evaluation summary would benefit from more explicit details. The full manuscript describes the verification process (manual inspection of symptoms, cross-checks against known compiler issues, and filtering of non-interaction bugs) in the evaluation section, but we will revise to expand the abstract with a brief mention of verification steps and add a dedicated paragraph detailing false-positive criteria (e.g., symptom-based manual review with inter-rater agreement) and any independent validation performed. This will better substantiate the reported cases. revision: yes

  2. Referee: [Approach] Approach description: The three-fold method (constraint extraction, prioritization, assertion insertion) is outlined at a high level, but without concrete examples of how a full-stack constraint is derived from a specific compilation pass and platform interaction or how prioritization scores are computed, it is difficult to assess whether the generated tests actually exercise the claimed deep logic.

    Authors: We acknowledge that concrete examples would aid assessment of the approach. We will revise Section 3 to include a running example deriving a full-stack constraint from a specific pass (e.g., in TVM or ONNX-MLIR) and its platform interaction, plus the exact computation of prioritization scores based on interaction sensitivity. This will demonstrate how tests target deep logic without altering the core method. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper describes an empirical testing framework (XCheck) that extracts cross-layer constraints from DL compilers, prioritizes them for test generation, inserts assertions, and evaluates the resulting cases on external compilers (TVM, ONNX-MLIR, etc.). No mathematical derivations, equations, fitted parameters, or predictions appear in the provided text. The central results (2,034 bug cases) are obtained by executing generated tests on third-party systems and are therefore independently falsifiable. No self-citations or ansatzes are invoked as load-bearing premises. The methodology is self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The work rests on domain assumptions about compiler behavior rather than new mathematical axioms or entities; no free parameters or invented entities are described.

axioms (1)
  • domain assumption Full-stack constraints can be automatically extracted to jointly guide model generation and characterize compilation behaviors
    Invoked in the three-fold approach description in the abstract.

pith-pipeline@v0.9.1-grok · 5833 in / 1105 out tokens · 32153 ms · 2026-06-26T23:27:52.338017+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

35 extracted references · 23 canonical work pages

  1. [1]

    [ten(2026)]

    Glow.https://github.com/pytorch/glow. [ten(2026)]

    2026

  2. [2]

    [onn(2026a)] 2026a

    NVIDIA TensorRT.https://developer.nvidia.com/tensorrt. [onn(2026a)] 2026a. ONNX Abs Operator Documentation. https://onnx.ai/onnx/operators/onnx__Abs. html#l-onnx-doc-abs. [onn(2026b)] 2026b. ONNX Documentation.https://onnx.ai/onnx/operators/index.html. [onn(2026c)] 2026c. ONNX: Open Neural Network Exchange.https://onnx.ai/. [tre(2026)]

    2026

  3. [3]

    Tree-sitter.https://tree-sitter.github.io/tree-sitter/. [Abadi et al.(2016)] Martín Abadi, Paul Barham, Jianmin Chen, Zhifeng Chen, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Geoffrey Irving, Michael Isard, Manjunath Kudlur, Josh Levenberg, Rajat Monga, Sherry Moore, Derek G. Murray, Benoit Steiner, Paul Tucker, Vijay Vasudevan, Pete Warde...

    2016

  4. [4]

    InProceedings of the 12th USENIX Conference on Operating Systems Design and Implementation(Savannah, GA, USA)(OSDI’16)

    TensorFlow: a system for large-scale machine learning. InProceedings of the 12th USENIX Conference on Operating Systems Design and Implementation(Savannah, GA, USA)(OSDI’16). USENIX Association, USA, 265–283. [Chen et al.(2018)] Tianqi Chen, Thierry Moreau, Ziheng Jiang, Lianmin Zheng, Eddie Yan, Meghan Cowan, Haichen Shen, Leyuan Wang, Yuwei Hu, Luis Cez...

    2018

  5. [5]

    InProceedings of the 13th USENIX Conference on Operating Systems Design and Implementation(Carlsbad, CA, USA)(OSDI’18)

    TVM: an automated end-to-end optimizing compiler for deep learning. InProceedings of the 13th USENIX Conference on Operating Systems Design and Implementation(Carlsbad, CA, USA)(OSDI’18). USENIX Association, USA, 579–594. [Chen et al.(2016)] Yu-Hsin Chen, Joel Emer, and Vivienne Sze

    2016

  6. [6]

    InProceedings of the 43rd International Symposium on Computer Architecture(Seoul, Republic of Korea)(ISCA ’16)

    Eyeriss: a spatial architecture for energy- efficient dataflow for convolutional neural networks. InProceedings of the 43rd International Symposium on Computer Architecture(Seoul, Republic of Korea)(ISCA ’16). IEEE Press, 367–379. doi: 10.1109/ISCA.2016. 40 [Danial(2021)] Albert Danial. 2021.cloc: v1.92. doi:10.5281/zenodo.5760077 [Das et al.(2020)] Sapta...

    work page doi:10.1109/isca.2016 2016

  7. [7]

    In2020 IEEE International Symposium on Circuits and Systems (ISCAS)

    A Systolic Dataflow Based Accelerator for CNNs. In2020 IEEE International Symposium on Circuits and Systems (ISCAS). 1–5. doi:10.1109/ISCAS45731.2020.9180403 [Deng et al.(2023)] Yinlin Deng, Chunqiu Steven Xia, Haoran Peng, Chenyuan Yang, and Lingming Zhang

    work page doi:10.1109/iscas45731.2020.9180403 2020

  8. [8]

    InProceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis(Seattle, W A, USA)(ISSTA 2023)

    Large Language Models Are Zero-Shot Fuzzers: Fuzzing Deep-Learning Libraries via Large Language Models. InProceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis(Seattle, W A, USA)(ISSTA 2023). Association for Computing Machinery, New York, NY , USA, 423–435. doi:10.1145/ 3597926.3598067 [Deng et al.(2024)] Yinlin Deng...

    arXiv 2023

  9. [9]

    InProceedings of the IEEE/ACM 46th International Conference on Software Engineering(Lisbon, Portugal)(ICSE ’24)

    Large Language Models are Edge-Case Generators: Crafting Unusual Programs for Fuzzing Deep Learning Libraries. InProceedings of the IEEE/ACM 46th International Conference on Software Engineering(Lisbon, Portugal)(ICSE ’24). Association for Computing Machinery, New York, NY , USA, Article 70, 13 pages. doi:10.1145/3597503.3623343 [Deng et al.(2022)] Yinlin...

    work page doi:10.1145/3597503.3623343 2022

  10. [10]

    InProceedings of the 30th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering(Singapore, Singapore) (ESEC/FSE 2022)

    Fuzzing deep-learning libraries via automated relational API inference. InProceedings of the 30th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering(Singapore, Singapore) (ESEC/FSE 2022). Association for Computing Machinery, New York, NY , USA, 44–56. doi:10.1145/3540250. 3549085 [Fowers et al.(2018...

    work page doi:10.1145/3540250 2022

  11. [11]

    In2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA)

    A Configurable Cloud-Scale DNN Processor for Real-Time AI. In2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA). 1–14. doi:10.1109/ISCA.2018.00012 [Genc et al.(2021)] Hasan Genc, Seah Kim, Alon Amid, Ameer Haj-Ali, Vighnesh Iyer, Pranav Prakash, Jerry Zhao, Daniel Grubb, Harrison Liew, Howard Mao, Albert Ou, Colin Schmidt, S...

    work page doi:10.1109/isca.2018.00012 2018

  12. [12]

    In2021 58th ACM/IEEE Design Automation Conference (DAC)

    Gemmini: Enabling 13 Finding Compiler-Platform Interaction Bugs in Deep Learning Pipelines via Cross-Layer Constraints Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration. In2021 58th ACM/IEEE Design Automation Conference (DAC). 769–774. doi:10.1109/DAC18074.2021.9586216 [Ghodrati et al.(2024)] Soroush Ghodrati, Sean Kinzer, Hanyan...

    work page doi:10.1109/dac18074.2021.9586216 2021

  13. [13]

    PyTorch 2: Faster Machine Learning Through Dynamic Python Bytecode Transformation and Graph Compilation

    Tandem Processor: Grappling with Emerging Operators in Neural Networks. InProceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2(La Jolla, CA, USA)(ASPLOS ’24). Association for Computing Machinery, New York, NY , USA, 1165–1182. doi:10.1145/3620665.3640365 [Gu et al.(2022)] J...

    work page doi:10.1145/3620665.3640365 2022

  14. [14]

    InProceedings of the 44th International Conference on Software Engineering (Pittsburgh, Pennsylvania)(ICSE ’22)

    Muffin: testing deep learning libraries via neural architecture fuzzing. InProceedings of the 44th International Conference on Software Engineering (Pittsburgh, Pennsylvania)(ICSE ’22). Association for Computing Machinery, New York, NY , USA, 1418–1430. doi:10.1145/3510003.3510092 [Guo et al.(2021)] Qianyu Guo, Xiaofei Xie, Yi Li, Xiaoyu Zhang, Yang Liu, ...

    work page doi:10.1145/3510003.3510092 2021

  15. [15]

    InProceedings of the 35th IEEE/ACM International Conference on Automated Software Engineering(Virtual Event, Australia)(ASE ’20)

    Audee: automated testing for deep learning frameworks. InProceedings of the 35th IEEE/ACM International Conference on Automated Software Engineering(Virtual Event, Australia)(ASE ’20). Association for Computing Machinery, New York, NY , USA, 486–498. doi:10.1145/3324884.3416571 [Jin et al.(2020)] Tian Jin, Gheorghe-Teodor Bercea, Tung D Le, Tong Chen, Gon...

    work page doi:10.1145/3324884.3416571 2020

  16. [16]

    arXiv preprint arXiv:2008.08272(2020)

    Compiling onnx neural network models using mlir. arXiv preprint arXiv:2008.08272(2020). [Kung et al.(2019)] H.T. Kung, Bradley McDanel, and Sai Qian Zhang

    arXiv 2008

  17. [17]

    Keckler, Christopher W

    Packing Sparse Convolutional Neural Networks for Efficient Systolic Array Implementations: Column Combining Under Joint Optimization. In Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems(Providence, RI, USA)(ASPLOS ’19). Association for Computing Machinery, New York, NY , USA...

    work page doi:10.1145/3297858.3304028 2023

  18. [18]

    In2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE)

    Graph- Based Fuzz Testing for Deep Learning Inference Engines. In2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE). 288–299. doi:10.1109/ICSE43902.2021.00037 [Lym and Erez(2020)] Sangkug Lym and Mattan Erez

    work page doi:10.1109/icse43902.2021.00037 2021

  19. [19]

    arXiv:2004.13027 https://arxiv.org/abs/ 2004.13027 [Ma et al.(2023)] Haoyang Ma, Qingchao Shen, Yongqiang Tian, Junjie Chen, and Shing-Chi Cheung

    FlexSA: Flexible Systolic Array Architecture for Efficient Pruned DNN Model Training.CoRRabs/2004.13027 (2020). arXiv:2004.13027 https://arxiv.org/abs/ 2004.13027 [Ma et al.(2023)] Haoyang Ma, Qingchao Shen, Yongqiang Tian, Junjie Chen, and Shing-Chi Cheung

    arXiv 2004

  20. [20]

    InProceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis(Seattle, WA, USA)(ISSTA 2023)

    Fuzzing Deep Learning Compilers with HirGen. InProceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis(Seattle, WA, USA)(ISSTA 2023). Association for Computing Machinery, New York, NY , USA, 248–260. doi:10.1145/3597926.3598053 [Mu et al.(2025)] Yanzhou Mu, Juan Zhai, Chunrong Fang, Xiang Chen, Zhixiang Cao, Peiran Yan...

    work page doi:10.1145/3597926.3598053 2023

  21. [21]

    ACM Softw

    Improving Deep Learning Framework Testing with Model-Level Metamorphic Testing.Proc. ACM Softw. Eng.2, ISSTA, Article ISSTA095 (June 2025), 23 pages. doi:10.1145/3728972 [Parashar et al.(2019)] Angshuman Parashar, Priyanka Raina, Yakun Sophia Shao, Yu-Hsin Chen, Victor A. Ying, Anurag Mukkara, Rangharajan Venkatesan, Brucek Khailany, Stephen W. Keckler, a...

    work page doi:10.1145/3728972 2025

  22. [22]

    In2019 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)

    Timeloop: A Systematic Approach to DNN Accelerator Evaluation. In2019 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS). 304–315. doi:10.1109/ISPASS.2019.00042 [Paszke et al.(2019)] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca A...

    work page doi:10.1109/ispass.2019.00042 2019

  23. [23]

    In2019 IEEE/ACM 41st International Conference on Software Engineering (ICSE)

    CRADLE: Cross-Backend Validation to Detect and Localize Bugs in Deep Learning Libraries. In2019 IEEE/ACM 41st International Conference on Software Engineering (ICSE). 1027–1038. doi:10.1109/ICSE.2019.00107 [Shen et al.(2021)] Qingchao Shen, Haoyang Ma, Junjie Chen, Yongqiang Tian, Shing-Chi Cheung, and Xiang Chen

    work page doi:10.1109/icse.2019.00107 2019

  24. [24]

    InProceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering (Athens, Greece)(ESEC/FSE 2021)

    A comprehensive study of deep learning compiler bugs. InProceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering (Athens, Greece)(ESEC/FSE 2021). Association for Computing Machinery, New York, NY , USA, 968–980. doi:10.1145/3468264.3468591 [Shen et al.(2025)] Qingchao Sh...

    work page doi:10.1145/3468264.3468591 2021

  25. [25]

    arXiv:2511.18918 [cs.SE]https://arxiv.org/abs/2511.18918 [Steinhöfel and Zeller(2022)] Dominic Steinhöfel and Andreas Zeller

    Optimization-Aware Test Generation for Deep Learning Compilers. arXiv:2511.18918 [cs.SE]https://arxiv.org/abs/2511.18918 [Steinhöfel and Zeller(2022)] Dominic Steinhöfel and Andreas Zeller

    arXiv 2022

  26. [26]

    InProceedings of the 30th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering

    Input invariants. InProceedings of the 30th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering. 583–594. [Steinhöfel and Smytzek(2022)] Dominic Steinhöfel and Marius Smytzek. 2022.rindPHI/islearn: ISLearn 0.2.13. doi:10.5281/zenodo.7035007 [Wang et al.(2024)] Haoyu Wang, Junjie Chen, Chuyue Xie, Shu...

    work page doi:10.5281/zenodo.7035007 2022

  27. [27]

    InProceedings of the 38th IEEE/ACM International Conference on Automated Software Engineering(Echternach, Luxembourg) (ASE ’23)

    MLIRSmith: Random Program Generation for Fuzzing MLIR Compiler Infrastructure. InProceedings of the 38th IEEE/ACM International Conference on Automated Software Engineering(Echternach, Luxembourg) (ASE ’23). IEEE Press, 1555–1566. doi:10.1109/ASE56229.2023.00120 [Wang et al.(2023)] Zihan Wang, Pengbo Nie, Xinyuan Miao, Yuting Chen, Chengcheng Wan, Lei Bu,...

    work page doi:10.1109/ase56229.2023.00120 2023

  28. [28]

    In Proceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis(Seattle, W A, USA)(ISSTA 2023)

    GenCoG: A DSL-Based Approach to Generating Computation Graphs for TVM Testing. In Proceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis(Seattle, W A, USA)(ISSTA 2023). Association for Computing Machinery, New York, NY , USA, 904–916. doi:10.1145/ 3597926.3598105 [Wang et al.(2020)] Zan Wang, Ming Yan, Junjie Chen, Sh...

    arXiv 2023

  29. [29]

    InProceedings of the 28th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering(Virtual Event, USA) (ESEC/FSE 2020)

    Deep learning library testing via effective model generation. InProceedings of the 28th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering(Virtual Event, USA) (ESEC/FSE 2020). Association for Computing Machinery, New York, NY , USA, 788–799. doi: 10.1145/ 3368089.3409761 [Xie et al.(2022)...

    arXiv 2020

  30. [30]

    InProceedings of the 31st ACM SIGSOFT International Symposium on Software Testing and Analysis(Virtual, South Korea)(ISSTA 2022)

    DocTer: documentation-guided fuzzing for testing deep learning API functions. InProceedings of the 31st ACM SIGSOFT International Symposium on Software Testing and Analysis(Virtual, South Korea)(ISSTA 2022). Association for Computing Machinery, New York, NY , USA, 176–188. doi:10.1145/3533767.3534220 [Xu et al.(2023)] Rui Xu, Sheng Ma, Yang Guo, and Dongsheng Li

    work page doi:10.1145/3533767.3534220 2022

  31. [31]

    Surv.56, 1, Article 20 (aug 2023), 37 pages

    A Survey of Design and Optimization for Systolic Array-based DNN Accelerators.ACM Comput. Surv.56, 1, Article 20 (aug 2023), 37 pages. doi:10.1145/3604802 [Yang et al.(2024)] Chenyuan Yang, Yinlin Deng, Runyu Lu, Jiayi Yao, Jiawei Liu, Reyhaneh Jabbarvand, and Lingming Zhang

    work page doi:10.1145/3604802 2023

  32. [32]

    ACM Program

    WhiteFox: White-Box Compiler Fuzzing Empowered by Large Language Models.Proc. ACM Program. Lang.8, OOPSLA2, Article 296 (Oct. 2024), 27 pages. doi:10.1145/3689736 [Yu et al.(2026)] Guangba Yu, Zirui Wang, Yujie Huang, Renyi Zhong, Yuedong Zhong, Yilun Wang, and Michael R. Lyu

    work page doi:10.1145/3689736 2024

  33. [33]

    arXiv:2601.13655 [cs.SE]https://arxiv.org/abs/2601.13655 [Zamudio Amaya et al.(2025)] José Antonio Zamudio Amaya, Marius Smytzek, and Andreas Zeller

    Why Does the LLM Stop Computing: An Empirical Study of User-Reported Failures in Open-Source LLMs. arXiv:2601.13655 [cs.SE]https://arxiv.org/abs/2601.13655 [Zamudio Amaya et al.(2025)] José Antonio Zamudio Amaya, Marius Smytzek, and Andreas Zeller

    arXiv 2025

  34. [34]

    ACM Softw

    FAN- DANGO: Evolving Language-Based Testing.Proc. ACM Softw. Eng.2, ISSTA, Article ISSTA040 (June 2025), 23 pages. doi:10.1145/3728915 [Zhou et al.(2024)] Chijin Zhou, Bingzhou Qian, Gwihwan Go, Quan Zhang, Shanshan Li, and Yu Jiang

    work page doi:10.1145/3728915 2025

  35. [35]

    ACM Program

    PolyJuice: Detecting Mis-compilation Bugs in Tensor Compilers with Equality Saturation Based Rewriting.Proc. ACM Program. Lang.8, OOPSLA2, Article 317 (Oct. 2024), 27 pages. doi:10.1145/3689757 15

    work page doi:10.1145/3689757 2024