Machine Learning for Timing Analysis: The Good, the Bad and the Ugly - Inria - Institut national de recherche en sciences et technologies du numérique
Document Associé À Des Manifestations Scientifiques Année : 2024

Machine Learning for Timing Analysis: The Good, the Bad and the Ugly

Résumé

The microarchitecture of processors is becoming increasingly complex and less documented, making the design of timing models for WCET calculation increasingly complicated, if not impossible. We have recently experimented with the use of machine learning techniques (ML) to predict the WCET of basic blocks. Predicted WCETs can then be integrated into static WCET calculation tools, resulting in a hybrid WCET calculation. In this keynote, we present our experience using ML for WCET calculation, across a range of architectures, from very simple ones (MSP430, Cortex M4) to more complex architectures. Rather than presenting only what worked, we also discuss in this keynote the bad, and even very bad, surprises encountered during the process, and how we overcame (most of) them.
Fichier principal
Vignette du fichier
OASIcs.WCET.2024.7.pdf (299.78 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04672502 , version 1 (19-08-2024)

Licence

Identifiants

Citer

Isabelle Puaut. Machine Learning for Timing Analysis: The Good, the Bad and the Ugly. Thomas Carle. WCET 2024 - 22nd International Workshop on Worst-Case Execution Time Analysis, Jul 2024, Lille, France. Open Access Series in Informatics (OASIcs), Open Access Series in Informatics (OASIcs), pp.1-1, 2024, Workshop on Worst-Case Execution Time Analysis (WCET). ⟨10.4230/OASIcs.WCET.2024.7⟩. ⟨hal-04672502⟩
22 Consultations
13 Téléchargements

Altmetric

Partager

More