Epistemic uncertainty analysis accounts for inaccurate input parameters and evaluates how such uncertainty propagates to output measures. In this work we will focus on Weibull distributions, in particular the one related to the reliability of multi-core systems-on-chips. We focus on a twocore system in which, when one of the cores fails, the other can take care of all the systems’ tasks. However, this results in an increased operational temperature that leads to a reduced lifetime. We first discuss the epistemic uncertainty distribution that we expect when we estimate the scale parameter of a Weibull distribution from a limited set of samples. We then extend the procedure to consider the epistemic uncertainty in the two core system-on-chip when we can measure the failure times of the first and of the second core.

Epistemic uncertainty propagation in a Weibull environment for a two-core system-on-chip

Pinciroli, R.;
2018-01-01

Abstract

Epistemic uncertainty analysis accounts for inaccurate input parameters and evaluates how such uncertainty propagates to output measures. In this work we will focus on Weibull distributions, in particular the one related to the reliability of multi-core systems-on-chips. We focus on a twocore system in which, when one of the cores fails, the other can take care of all the systems’ tasks. However, this results in an increased operational temperature that leads to a reduced lifetime. We first discuss the epistemic uncertainty distribution that we expect when we estimate the scale parameter of a Weibull distribution from a limited set of samples. We then extend the procedure to consider the epistemic uncertainty in the two core system-on-chip when we can measure the failure times of the first and of the second core.
2018
978-1-5386-3322-9
Epistemic uncertainty, Uncertainty propagation, Weibull distribution, MinMax Weibull distributions, MultiCore System on Chip
File in questo prodotto:
File Dimensione Formato  
2018_ICSRS_2018_Pinciroli.pdf

non disponibili

Tipologia: Versione Editoriale (PDF)
Licenza: Non pubblico
Dimensione 767.02 kB
Formato Adobe PDF
767.02 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12571/27429
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 4
  • ???jsp.display-item.citation.isi??? 3
social impact