A Spatially Correlated Competing Risks Time-to-Event Model for Supercomputer GPU Failure Data.
Saved in:
| Title: | A Spatially Correlated Competing Risks Time-to-Event Model for Supercomputer GPU Failure Data. |
|---|---|
| Authors: | Min, Jie1 (AUTHOR) jiem@usf.edu, Hong, Yili1 (AUTHOR), Meeker, William Q.2 (AUTHOR), Ostrouchov, George3 (AUTHOR) |
| Source: | Technometrics. Aug2025, Vol. 67 Issue 3, p531-545. 15p. |
| Subjects: | Supercomputers, Graphics processing units, Failure analysis, High performance computing, Statistical models, Competing risks, Statistical correlation, Bayesian analysis |
| Abstract: | Graphics processing units (GPUs) are widely used in many high-performance computing (HPC) applications such as imaging/video processing and training deep-learning models in artificial intelligence. GPUs installed in HPC systems are often heavily used, and GPU failures occur during HPC system operations. Thus, the reliability of GPUs is of interest for the overall reliability of HPC systems. The Cray XK7 Titan supercomputer was one of the top ten supercomputers in the world. The failure event times of more than 30,000 GPUs in Titan were recorded and previous data analysis suggested that the failure time of a GPU may be affected by the GPU's connectivity location inside the supercomputer among other factors. In this article, we conduct in-depth statistical modeling of GPU failure times to study the effect of location on GPU failures under competing risks with covariates and spatially correlated random effects. In particular, two major failure types of GPUs in Titan are considered. The connectivity locations of cabinets are modeled as spatially correlated random effects, and the positions of GPUs inside each cabinet are treated as covariates. A Bayesian framework is used for statistical inference. We also compare different methods of estimation such as the maximum likelihood, which is implemented via an expectation-maximization algorithm. Our results provide interesting insights into GPU failures in HPC systems. This article has online supplementary materials. [ABSTRACT FROM AUTHOR] |
| Copyright of Technometrics is the property of Taylor & Francis Ltd 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 |
|
Full text is not displayed to guests.
Login for full access.
|
|
| FullText | Links: – Type: pdflink Text: Availability: 1 |
|---|---|
| Header | DbId: egs DbLabel: Engineering Source An: 187023484 AccessLevel: 6 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: A Spatially Correlated Competing Risks Time-to-Event Model for Supercomputer GPU Failure Data. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Min%2C+Jie%22">Min, Jie</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> jiem@usf.edu</i><br /><searchLink fieldCode="AR" term="%22Hong%2C+Yili%22">Hong, Yili</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Meeker%2C+William+Q%2E%22">Meeker, William Q.</searchLink><relatesTo>2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Ostrouchov%2C+George%22">Ostrouchov, George</searchLink><relatesTo>3</relatesTo> (AUTHOR) – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="JN" term="%22Technometrics%22">Technometrics</searchLink>. Aug2025, Vol. 67 Issue 3, p531-545. 15p. – Name: Subject Label: Subjects Group: Su Data: <searchLink fieldCode="DE" term="%22Supercomputers%22">Supercomputers</searchLink><br /><searchLink fieldCode="DE" term="%22Graphics+processing+units%22">Graphics processing units</searchLink><br /><searchLink fieldCode="DE" term="%22Failure+analysis%22">Failure analysis</searchLink><br /><searchLink fieldCode="DE" term="%22High+performance+computing%22">High performance computing</searchLink><br /><searchLink fieldCode="DE" term="%22Statistical+models%22">Statistical models</searchLink><br /><searchLink fieldCode="DE" term="%22Competing+risks%22">Competing risks</searchLink><br /><searchLink fieldCode="DE" term="%22Statistical+correlation%22">Statistical correlation</searchLink><br /><searchLink fieldCode="DE" term="%22Bayesian+analysis%22">Bayesian analysis</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: Graphics processing units (GPUs) are widely used in many high-performance computing (HPC) applications such as imaging/video processing and training deep-learning models in artificial intelligence. GPUs installed in HPC systems are often heavily used, and GPU failures occur during HPC system operations. Thus, the reliability of GPUs is of interest for the overall reliability of HPC systems. The Cray XK7 Titan supercomputer was one of the top ten supercomputers in the world. The failure event times of more than 30,000 GPUs in Titan were recorded and previous data analysis suggested that the failure time of a GPU may be affected by the GPU's connectivity location inside the supercomputer among other factors. In this article, we conduct in-depth statistical modeling of GPU failure times to study the effect of location on GPU failures under competing risks with covariates and spatially correlated random effects. In particular, two major failure types of GPUs in Titan are considered. The connectivity locations of cabinets are modeled as spatially correlated random effects, and the positions of GPUs inside each cabinet are treated as covariates. A Bayesian framework is used for statistical inference. We also compare different methods of estimation such as the maximum likelihood, which is implemented via an expectation-maximization algorithm. Our results provide interesting insights into GPU failures in HPC systems. This article has online supplementary materials. [ABSTRACT FROM AUTHOR] – Name: AbstractSuppliedCopyright Label: Group: Ab Data: <i>Copyright of Technometrics is the property of Taylor & Francis Ltd 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=187023484 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1080/00401706.2025.2475783 Languages: – Code: eng Text: English PhysicalDescription: Pagination: PageCount: 15 StartPage: 531 Subjects: – SubjectFull: Supercomputers Type: general – SubjectFull: Graphics processing units Type: general – SubjectFull: Failure analysis Type: general – SubjectFull: High performance computing Type: general – SubjectFull: Statistical models Type: general – SubjectFull: Competing risks Type: general – SubjectFull: Statistical correlation Type: general – SubjectFull: Bayesian analysis Type: general Titles: – TitleFull: A Spatially Correlated Competing Risks Time-to-Event Model for Supercomputer GPU Failure Data. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Min, Jie – PersonEntity: Name: NameFull: Hong, Yili – PersonEntity: Name: NameFull: Meeker, William Q. – PersonEntity: Name: NameFull: Ostrouchov, George IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 08 Text: Aug2025 Type: published Y: 2025 Identifiers: – Type: issn-print Value: 00401706 Numbering: – Type: volume Value: 67 – Type: issue Value: 3 Titles: – TitleFull: Technometrics Type: main |
| ResultId | 1 |
