Performance-Reliability Trade-Off in Graphics Processing Units - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2022

Performance-Reliability Trade-Off in Graphics Processing Units

Abstract

We show that most performance improvements in GPUs increase the number of executions correctly completed before experiencing a failure. We consider four different performance improvements: architectural solutions, software implementations, compiler optimizations, and degree of parallelism.

Domains

Computer Science [cs]
Fichier principal
Vignette du fichier
radecs2022.pdf (2.54 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Fernando Fernandes dos Santos  : Connect in order to contact the contributor

https://inria.hal.science/hal-03680872

Submitted on : Sunday, May 29, 2022-4:39:32 PM

Last modification on : Tuesday, April 30, 2024-3:46:03 PM

Long-term archiving on: Tuesday, August 30, 2022-6:10:06 PM

Download

Dates and versions

hal-03680872 , version 1 (29-05-2022)

Licence

Attribution

Identifiers

  • HAL Id : hal-03680872 , version 1

Cite

Fernando Fernandes dos Santos, Paolo Rech. Performance-Reliability Trade-Off in Graphics Processing Units. RADiation Effects on Components and Systems (RADECS), Oct 2022, Venice, Italy. ⟨hal-03680872⟩

Export

BibTeX XML-TEI Dublin Core DC Terms EndNote DataCite

Collections

UNIV-RENNES1 CNRS INRIA INSA-RENNES IRISA CENTRALESUPELEC INRIA2 UR1-MATH-STIC UR1-UFR-ISTIC UNIV-RENNES UR1-MATH-NUM
88 View
132 Download

Share

Gmail Mastodon Facebook X LinkedIn More