MC2: Multicore and cache analysis via deterministic and probabilistic jitter bounding

10Citations
Citations of this article
3Readers
Mendeley users who have this article in their library.
Get full text

Abstract

In critical domains, reliable software execution is increasingly involving aspects related to the timing dimension. This is due to the advent of high-performance (complex) hardware, used to provide the rising levels of guaranteed performance needed in those domains. Caches and multicores are two of the hardware features that have the potential to significantly reduce WCET estimates, yet they pose new challenges on current-practice measurement-based timing analysis (MBTA) approaches. In this paper we propose MC2, a technique for multilevel-cache multicores that combines deterministic and probabilistic jitter-bounding approaches to reliably handle both the variability in execution time generated by caches and the contention in accessing shared hardware resources. We evaluate MC2 on a COTS quad-core LEON-based board and our initial results show how it effectively captures cache and multicore contention in pWCET estimates with respect to actual observed values.

Cite

CITATION STYLE

APA

Díaz, E., Fernández, M., Kosmidis, L., Mezzetti, E., Hernandez, C., Abella, J., & Cazorla, F. J. (2017). MC2: Multicore and cache analysis via deterministic and probabilistic jitter bounding. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10300 LNCS, pp. 102–118). Springer Verlag. https://doi.org/10.1007/978-3-319-60588-3_7

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free