Diagnosis of Performance Faults in LargeScale MPI Applications via Probabilistic Progress-Dependence Inference.
Saved in:
| Title: | Diagnosis of Performance Faults in LargeScale MPI Applications via Probabilistic Progress-Dependence Inference. |
|---|---|
| Authors: | Laguna, Ignacio1, Ahn, Dong H.1, Supinski, Bronis R. de1, Bagchi, Saurabh2, Gamblin, Todd1 |
| Source: | IEEE Transactions on Parallel & Distributed Systems. May2015, Vol. 26 Issue 5, p1280-1289. 10p. |
| Subjects: | Debugging, High performance computing, Application software, Markov processes, Molecular dynamics |
| Abstract: | Debugging large-scale parallel applications is challenging. Most existing techniques provide little information about failure root causes. Further, most debuggers significantly slow down program execution, and run sluggishly with massively parallel applications. This paper presents a novel technique that scalably infers the tasks in a parallel program on which a failure occurred, as well as the code in which it originated. Our technique combines scalable runtime analysis with static analysis to determine the least-progressed task(s) and to identify the code lines at which the failure arose. We present a novel algorithm that infers probabilistically progress dependence among MPI tasks using a globally constructed Markov model that represents tasks’ control-flow behavior. In comparison to previous work, our algorithm infers more precisely the least-progressed task. We combine this technique with static backward slicing analysis, further isolating the code responsible for the current state. A blind study demonstrates that our technique isolates the root cause of a concurrency bug in a molecular dynamics simulation, which only manifests itself at 7,996 tasks or more. We extensively evaluate fault coverage of our technique via fault injections in 10 HPC benchmarks and show that our analysis takes less than a few seconds on thousands of parallel tasks. [ABSTRACT FROM AUTHOR] |
| Copyright of IEEE Transactions on Parallel & Distributed Systems is the property of IEEE 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: | Engineering Source |
| FullText | Text: Availability: 0 |
|---|---|
| Header | DbId: egs DbLabel: Engineering Source An: 102086813 AccessLevel: 6 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Diagnosis of Performance Faults in LargeScale MPI Applications via Probabilistic Progress-Dependence Inference. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Laguna%2C+Ignacio%22">Laguna, Ignacio</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Ahn%2C+Dong+H%2E%22">Ahn, Dong H.</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Supinski%2C+Bronis+R%2E+de%22">Supinski, Bronis R. de</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Bagchi%2C+Saurabh%22">Bagchi, Saurabh</searchLink><relatesTo>2</relatesTo><br /><searchLink fieldCode="AR" term="%22Gamblin%2C+Todd%22">Gamblin, Todd</searchLink><relatesTo>1</relatesTo> – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="JN" term="%22IEEE+Transactions+on+Parallel+%26+Distributed+Systems%22">IEEE Transactions on Parallel & Distributed Systems</searchLink>. May2015, Vol. 26 Issue 5, p1280-1289. 10p. – Name: Subject Label: Subjects Group: Su Data: <searchLink fieldCode="DE" term="%22Debugging%22">Debugging</searchLink><br /><searchLink fieldCode="DE" term="%22High+performance+computing%22">High performance computing</searchLink><br /><searchLink fieldCode="DE" term="%22Application+software%22">Application software</searchLink><br /><searchLink fieldCode="DE" term="%22Markov+processes%22">Markov processes</searchLink><br /><searchLink fieldCode="DE" term="%22Molecular+dynamics%22">Molecular dynamics</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: Debugging large-scale parallel applications is challenging. Most existing techniques provide little information about failure root causes. Further, most debuggers significantly slow down program execution, and run sluggishly with massively parallel applications. This paper presents a novel technique that scalably infers the tasks in a parallel program on which a failure occurred, as well as the code in which it originated. Our technique combines scalable runtime analysis with static analysis to determine the least-progressed task(s) and to identify the code lines at which the failure arose. We present a novel algorithm that infers probabilistically progress dependence among MPI tasks using a globally constructed Markov model that represents tasks’ control-flow behavior. In comparison to previous work, our algorithm infers more precisely the least-progressed task. We combine this technique with static backward slicing analysis, further isolating the code responsible for the current state. A blind study demonstrates that our technique isolates the root cause of a concurrency bug in a molecular dynamics simulation, which only manifests itself at 7,996 tasks or more. We extensively evaluate fault coverage of our technique via fault injections in 10 HPC benchmarks and show that our analysis takes less than a few seconds on thousands of parallel tasks. [ABSTRACT FROM AUTHOR] – Name: AbstractSuppliedCopyright Label: Group: Ab Data: <i>Copyright of IEEE Transactions on Parallel & Distributed Systems is the property of IEEE 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=egs&AN=102086813 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1109/TPDS.2014.2314100 Languages: – Code: eng Text: English PhysicalDescription: Pagination: PageCount: 10 StartPage: 1280 Subjects: – SubjectFull: Debugging Type: general – SubjectFull: High performance computing Type: general – SubjectFull: Application software Type: general – SubjectFull: Markov processes Type: general – SubjectFull: Molecular dynamics Type: general Titles: – TitleFull: Diagnosis of Performance Faults in LargeScale MPI Applications via Probabilistic Progress-Dependence Inference. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Laguna, Ignacio – PersonEntity: Name: NameFull: Ahn, Dong H. – PersonEntity: Name: NameFull: Supinski, Bronis R. de – PersonEntity: Name: NameFull: Bagchi, Saurabh – PersonEntity: Name: NameFull: Gamblin, Todd IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 05 Text: May2015 Type: published Y: 2015 Identifiers: – Type: issn-print Value: 10459219 Numbering: – Type: volume Value: 26 – Type: issue Value: 5 Titles: – TitleFull: IEEE Transactions on Parallel & Distributed Systems Type: main |
| ResultId | 1 |
