Energy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection

To meet the performance demands of chip multiprocessors, chip designers have increased the capacity and hierarchy of cache memories. Accordingly, a shared lower-level cache reduces conflict misses by adopting a multi-way set-associative structure with high associativity. This structure allows fast a...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Yun-Seok Oh, Eui-Young Chung
Formato: article
Lenguaje:EN
Publicado: IEEE 2021
Materias:
Acceso en línea:https://doaj.org/article/651c5058aa7b4724be14ea67f2319fd5
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:651c5058aa7b4724be14ea67f2319fd5
record_format dspace
spelling oai:doaj.org-article:651c5058aa7b4724be14ea67f2319fd52021-11-26T00:00:58ZEnergy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection2169-353610.1109/ACCESS.2021.3126739https://doaj.org/article/651c5058aa7b4724be14ea67f2319fd52021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9606910/https://doaj.org/toc/2169-3536To meet the performance demands of chip multiprocessors, chip designers have increased the capacity and hierarchy of cache memories. Accordingly, a shared lower-level cache reduces conflict misses by adopting a multi-way set-associative structure with high associativity. This structure allows fast access because it allows access to all the ways in the cache set in parallel. However, it consumes a large amount of dynamic energy. Therefore, various schemes have been proposed to increase the energy efficiency of the cache memory. These schemes use <italic>way prediction</italic> or <italic>partial comparison</italic> to reduce unnecessary way access. This paper proposes a <italic>way prediction</italic> algorithm suitable for a shared second-level cache with high associativity. This algorithm is based on real-time <italic>way access dominance detection</italic> (<italic>WADD</italic>). Through this detection, the proposed algorithm can determine the number and location of <italic>way candidates</italic> suitable for each <italic>partial access pattern</italic> among the fragmented access patterns owing to the first-level cache replacement policy and intermingled accesses by multiple cores. Through this process, the proposed algorithm can implement an efficient <italic>way prediction</italic>. Simulation results show that the <italic>WADD</italic> exhibits the highest energy efficiency among the comparison groups, thus reducing the energy-delay product by 13.5&#x0025; compared with the conventional cache without <italic>way prediction</italic>. This result is achieved by reducing the <italic>way prediction</italic> penalty through fast detection and high prediction accuracy.Yun-Seok OhEui-Young ChungIEEEarticleCache memorycomputer architectureenergy efficiencymultiprocessing systemsway predictionElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 155048-155057 (2021)
institution DOAJ
collection DOAJ
language EN
topic Cache memory
computer architecture
energy efficiency
multiprocessing systems
way prediction
Electrical engineering. Electronics. Nuclear engineering
TK1-9971
spellingShingle Cache memory
computer architecture
energy efficiency
multiprocessing systems
way prediction
Electrical engineering. Electronics. Nuclear engineering
TK1-9971
Yun-Seok Oh
Eui-Young Chung
Energy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection
description To meet the performance demands of chip multiprocessors, chip designers have increased the capacity and hierarchy of cache memories. Accordingly, a shared lower-level cache reduces conflict misses by adopting a multi-way set-associative structure with high associativity. This structure allows fast access because it allows access to all the ways in the cache set in parallel. However, it consumes a large amount of dynamic energy. Therefore, various schemes have been proposed to increase the energy efficiency of the cache memory. These schemes use <italic>way prediction</italic> or <italic>partial comparison</italic> to reduce unnecessary way access. This paper proposes a <italic>way prediction</italic> algorithm suitable for a shared second-level cache with high associativity. This algorithm is based on real-time <italic>way access dominance detection</italic> (<italic>WADD</italic>). Through this detection, the proposed algorithm can determine the number and location of <italic>way candidates</italic> suitable for each <italic>partial access pattern</italic> among the fragmented access patterns owing to the first-level cache replacement policy and intermingled accesses by multiple cores. Through this process, the proposed algorithm can implement an efficient <italic>way prediction</italic>. Simulation results show that the <italic>WADD</italic> exhibits the highest energy efficiency among the comparison groups, thus reducing the energy-delay product by 13.5&#x0025; compared with the conventional cache without <italic>way prediction</italic>. This result is achieved by reducing the <italic>way prediction</italic> penalty through fast detection and high prediction accuracy.
format article
author Yun-Seok Oh
Eui-Young Chung
author_facet Yun-Seok Oh
Eui-Young Chung
author_sort Yun-Seok Oh
title Energy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection
title_short Energy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection
title_full Energy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection
title_fullStr Energy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection
title_full_unstemmed Energy-Efficient Shared Cache Using Way Prediction Based on Way Access Dominance Detection
title_sort energy-efficient shared cache using way prediction based on way access dominance detection
publisher IEEE
publishDate 2021
url https://doaj.org/article/651c5058aa7b4724be14ea67f2319fd5
work_keys_str_mv AT yunseokoh energyefficientsharedcacheusingwaypredictionbasedonwayaccessdominancedetection
AT euiyoungchung energyefficientsharedcacheusingwaypredictionbasedonwayaccessdominancedetection
_version_ 1718409973949530112