Aging Assessment and Design Enhancement of Randomized Cache Memories
Visualitza/Obre
Cita com:
hdl:2117/103465
Tipus de documentArticle
Data publicació2017-01-17
EditorInstitute of Electrical and Electronics Engineers (IEEE)
Condicions d'accésAccés obert
Tots els drets reservats. Aquesta obra està protegida pels drets de propietat intel·lectual i
industrial corresponents. Sense perjudici de les exempcions legals existents, queda prohibida la seva
reproducció, distribució, comunicació pública o transformació sense l'autorització del titular dels drets
ProjecteCOMPUTACION DE ALTAS PRESTACIONES VII (MINECO-TIN2015-65316-P)
RYC-2013-14717 (MINECO-RYC-2013-14717)
PROBABILISTIC TIMING ANALYSIS OF HIGH-PERFORMANCE AND RELIABLE PROCESSORS (MINECO-TIN2014-60404-JIN)
RYC-2013-14717 (MINECO-RYC-2013-14717)
PROBABILISTIC TIMING ANALYSIS OF HIGH-PERFORMANCE AND RELIABLE PROCESSORS (MINECO-TIN2014-60404-JIN)
Abstract
Critical real-time systems require the estimation of the worst-case execution time (WCET) for scheduling purposes and resource budgeting. Measurement-based probabilistic timing analysis (MBPTA) has been shown recently as a powerful approach for WCET estimation. MBPTA builds upon time-randomized cache memories. While aging has been deeply analyzed for conventional time-deterministic caches (i.e., implementing modulo placement), little work has been done to assess the reliability of time-randomized caches. In this line, only the WCET analysis of faulty time-randomized caches has been introduced recently. However, the intrinsic robustness of randomized hardware designs has not been assessed yet. In this paper we perform, for the first time, an assessment of the agingrobustness of random placement cache designs: random modulo and hash-based random placement. We propose a new random modulo implementation preserving its key benefits in terms of low critical path impact, low miss rates and MBPTA compliance; while reducing hot-carrier injection aging by achieving a better, yet random, activity distribution across cache sets. On the other hand, we show that gains in terms of bias temperature instability aging are limited for random placement designs on their own.
CitacióTrilla, D. [et al.]. Aging Assessment and Design Enhancement of Randomized Cache Memories. "IEEE Transactions on Device and Materials Reliability", 17 Gener 2017, vol. 17, núm. 1, p. 32-41.
ISSN1530-4388
Versió de l'editorhttp://ieeexplore.ieee.org/document/7820072/
Col·leccions
Fitxers | Descripció | Mida | Format | Visualitza |
---|---|---|---|---|
Aging Assessment and Design Enhancement of.pdf | 611,0Kb | Visualitza/Obre |