Academic Journal

Scuzer: A Scheduling Optimization Fuzzer for TVM.

Bibliographic Details
Title: Scuzer: A Scheduling Optimization Fuzzer for TVM.
Authors: Chen, Xiangxiang, Lin, Xingwei, Wang, Jingyi, Sun, Jun, Wang, Jiashui, Wang, Wenhai
Source: ACM Transactions on Software Engineering & Methodology; May2025, Vol. 34 Issue 4, p1-28, 28p
Subject Terms: Scheduling, Compilers (Computer programs), Mathematical optimization, Deep learning, Defect tracking (Computer software development), Optimizers (Computer software), Scientific experimentation
Abstract: The concept of Deep Learning (DL) compiler was proposed to deploy DL models more efficiently on diverse hardware through optimization techniques. As one of the most popular DL compilers, TVM incorporates three levels (high-level, schedule, and low-level) of optimizations, which can inadvertently introduce code logic bugs and build failure bugs. Among these optimizations, scheduling optimization is the core component of DL compilers, which ensures the acceleration of models on all devices. However, the existing works only focus on the testing of high-level and low-level optimizations in TVM, fail to take the most important and challenging intermediate scheduling optimization layer into consideration. To fill the gap, we propose a Scheduling Optimization Oriented Fuzzer (Scuzer) for TVM, which is specially designed to effectively detect bugs introduced by the scheduling optimization. In particular, Scuzer first proposes a set of schedule-triggering mutators to actively trigger many scheduling optimizations. Meanwhile, observing that scheduling optimization is closely coupled with program dataflow and operator type, Scuzer additionally proposes a set of structure-enriching mutators to enrich the structure of dataflows and operators. Based on these carefully designed mutators, Scuzer then devises a multi-objective algorithm that can adaptively select different combinations of objectives at each period to guide the selection of seeds and mutators during fuzzing. We conduct extensive experiments comparing with three state-of-the-art fuzzers that can be applied in testing scheduling optimization to evaluate the effectiveness of Scuzer. The experimental results demonstrate that Scuzer outperforms the 2nd-best state-of-the-art fuzzer by 7.4% in edge coverage and achieves 7 \(\times\) improvement in rule-operator coverage. Scuzer has successfully detected 17 previously unknown bugs (9 are inconsistent results and 5 are inconsistent compilations) in TVM, out of which 10 have been confirmed and 5 been fixed. [ABSTRACT FROM AUTHOR]
Copyright of ACM Transactions on Software Engineering & Methodology is the property of Association for Computing Machinery and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Complementary Index
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://resolver.ebsco.com/c/fiv2js/result?sid=EBSCO:edb&genre=article&issn=1049331X&ISBN=&volume=34&issue=4&date=20250501&spage=1&pages=1-28&title=ACM Transactions on Software Engineering & Methodology&atitle=Scuzer%3A%20A%20Scheduling%20Optimization%20Fuzzer%20for%20TVM.&aulast=Chen%2C%20Xiangxiang&id=DOI:10.1145/3705308
    Name: Full Text Finder (for New FTF UI) (ns324271)
    Category: fullText
    Text: Full Text Finder
    MouseOverText: Full Text Finder
Header DbId: edb
DbLabel: Complementary Index
An: 188947129
RelevancyScore: 994
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 993.548583984375
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Scuzer: A Scheduling Optimization Fuzzer for TVM.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Chen%2C+Xiangxiang%22">Chen, Xiangxiang</searchLink><br /><searchLink fieldCode="AR" term="%22Lin%2C+Xingwei%22">Lin, Xingwei</searchLink><br /><searchLink fieldCode="AR" term="%22Wang%2C+Jingyi%22">Wang, Jingyi</searchLink><br /><searchLink fieldCode="AR" term="%22Sun%2C+Jun%22">Sun, Jun</searchLink><br /><searchLink fieldCode="AR" term="%22Wang%2C+Jiashui%22">Wang, Jiashui</searchLink><br /><searchLink fieldCode="AR" term="%22Wang%2C+Wenhai%22">Wang, Wenhai</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: ACM Transactions on Software Engineering & Methodology; May2025, Vol. 34 Issue 4, p1-28, 28p
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Scheduling%22">Scheduling</searchLink><br /><searchLink fieldCode="DE" term="%22Compilers+%28Computer+programs%29%22">Compilers (Computer programs)</searchLink><br /><searchLink fieldCode="DE" term="%22Mathematical+optimization%22">Mathematical optimization</searchLink><br /><searchLink fieldCode="DE" term="%22Deep+learning%22">Deep learning</searchLink><br /><searchLink fieldCode="DE" term="%22Defect+tracking+%28Computer+software+development%29%22">Defect tracking (Computer software development)</searchLink><br /><searchLink fieldCode="DE" term="%22Optimizers+%28Computer+software%29%22">Optimizers (Computer software)</searchLink><br /><searchLink fieldCode="DE" term="%22Scientific+experimentation%22">Scientific experimentation</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: The concept of Deep Learning (DL) compiler was proposed to deploy DL models more efficiently on diverse hardware through optimization techniques. As one of the most popular DL compilers, TVM incorporates three levels (high-level, schedule, and low-level) of optimizations, which can inadvertently introduce code logic bugs and build failure bugs. Among these optimizations, scheduling optimization is the core component of DL compilers, which ensures the acceleration of models on all devices. However, the existing works only focus on the testing of high-level and low-level optimizations in TVM, fail to take the most important and challenging intermediate scheduling optimization layer into consideration. To fill the gap, we propose a Scheduling Optimization Oriented Fuzzer (Scuzer) for TVM, which is specially designed to effectively detect bugs introduced by the scheduling optimization. In particular, Scuzer first proposes a set of schedule-triggering mutators to actively trigger many scheduling optimizations. Meanwhile, observing that scheduling optimization is closely coupled with program dataflow and operator type, Scuzer additionally proposes a set of structure-enriching mutators to enrich the structure of dataflows and operators. Based on these carefully designed mutators, Scuzer then devises a multi-objective algorithm that can adaptively select different combinations of objectives at each period to guide the selection of seeds and mutators during fuzzing. We conduct extensive experiments comparing with three state-of-the-art fuzzers that can be applied in testing scheduling optimization to evaluate the effectiveness of Scuzer. The experimental results demonstrate that Scuzer outperforms the 2nd-best state-of-the-art fuzzer by 7.4% in edge coverage and achieves 7 \(\times\) improvement in rule-operator coverage. Scuzer has successfully detected 17 previously unknown bugs (9 are inconsistent results and 5 are inconsistent compilations) in TVM, out of which 10 have been confirmed and 5 been fixed. [ABSTRACT FROM AUTHOR]
– Name: Abstract
  Label:
  Group: Ab
  Data: <i>Copyright of ACM Transactions on Software Engineering & Methodology is the property of Association for Computing Machinery and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edb&AN=188947129
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1145/3705308
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 28
        StartPage: 1
    Subjects:
      – SubjectFull: Scheduling
        Type: general
      – SubjectFull: Compilers (Computer programs)
        Type: general
      – SubjectFull: Mathematical optimization
        Type: general
      – SubjectFull: Deep learning
        Type: general
      – SubjectFull: Defect tracking (Computer software development)
        Type: general
      – SubjectFull: Optimizers (Computer software)
        Type: general
      – SubjectFull: Scientific experimentation
        Type: general
    Titles:
      – TitleFull: Scuzer: A Scheduling Optimization Fuzzer for TVM.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Chen, Xiangxiang
      – PersonEntity:
          Name:
            NameFull: Lin, Xingwei
      – PersonEntity:
          Name:
            NameFull: Wang, Jingyi
      – PersonEntity:
          Name:
            NameFull: Sun, Jun
      – PersonEntity:
          Name:
            NameFull: Wang, Jiashui
      – PersonEntity:
          Name:
            NameFull: Wang, Wenhai
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 05
              Text: May2025
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 1049331X
          Numbering:
            – Type: volume
              Value: 34
            – Type: issue
              Value: 4
          Titles:
            – TitleFull: ACM Transactions on Software Engineering & Methodology
              Type: main
ResultId 1