The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
À medida que o número de núcleos em um processador aumenta, as hierarquias de cache contêm mais níveis de cache e um cache de último nível (LLC) maior. Assim, o consumo de energia e energia da hierarquia de cache torna-se não negligenciável. Enquanto isso, como os comportamentos de uso do cache de aplicativos individuais podem ser diferentes, é possível obter maior eficiência energética do sistema de computação determinando as configurações de cache apropriadas para aplicativos individuais. Este artigo propõe um mecanismo de controle de cache para melhorar a eficiência energética ajustando uma hierarquia de cache para cada aplicação. Nosso mecanismo primeiro ignora e desativa um nível de cache menos significativo, depois desativa parcialmente o LLC e, finalmente, ajusta a associatividade se sofrer um grande número de falhas de conflito. O mecanismo pode alcançar uma economia de energia significativa sacrificando uma pequena degradação do desempenho. Os resultados da avaliação mostram que nosso mecanismo melhora a eficiência energética em 23.9% e 7.0%, em média, em relação à linha de base e aos mecanismos de desvio no nível de cache, respectivamente. Além disso, mesmo que ocorra a contenção de recursos da LLC, o mecanismo proposto ainda é eficaz para melhorar a eficiência energética.
Jiaheng LIU
Tohoku University
Ryusuke EGAWA
Tokyo Denki University
Hiroyuki TAKIZAWA
Tohoku University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Jiaheng LIU, Ryusuke EGAWA, Hiroyuki TAKIZAWA, "A Conflict-Aware Capacity Control Mechanism for Deep Cache Hierarchy" in IEICE TRANSACTIONS on Information,
vol. E105-D, no. 6, pp. 1150-1163, June 2022, doi: 10.1587/transinf.2021EDP7201.
Abstract: As the number of cores on a processor increases, cache hierarchies contain more cache levels and a larger last level cache (LLC). Thus, the power and energy consumption of the cache hierarchy becomes non-negligible. Meanwhile, because the cache usage behaviors of individual applications can be different, it is possible to achieve higher energy efficiency of the computing system by determining the appropriate cache configurations for individual applications. This paper proposes a cache control mechanism to improve energy efficiency by adjusting a cache hierarchy to each application. Our mechanism first bypasses and disables a less-significant cache level, then partially disables the LLC, and finally adjusts the associativity if it suffers from a large number of conflict misses. The mechanism can achieve significant energy saving at the sacrifice of small performance degradation. The evaluation results show that our mechanism improves energy efficiency by 23.9% and 7.0% on average over the baseline and the cache-level bypassing mechanisms, respectively. In addition, even if the LLC resource contention occurs, the proposed mechanism is still effective for improving energy efficiency.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2021EDP7201/_p
Copiar
@ARTICLE{e105-d_6_1150,
author={Jiaheng LIU, Ryusuke EGAWA, Hiroyuki TAKIZAWA, },
journal={IEICE TRANSACTIONS on Information},
title={A Conflict-Aware Capacity Control Mechanism for Deep Cache Hierarchy},
year={2022},
volume={E105-D},
number={6},
pages={1150-1163},
abstract={As the number of cores on a processor increases, cache hierarchies contain more cache levels and a larger last level cache (LLC). Thus, the power and energy consumption of the cache hierarchy becomes non-negligible. Meanwhile, because the cache usage behaviors of individual applications can be different, it is possible to achieve higher energy efficiency of the computing system by determining the appropriate cache configurations for individual applications. This paper proposes a cache control mechanism to improve energy efficiency by adjusting a cache hierarchy to each application. Our mechanism first bypasses and disables a less-significant cache level, then partially disables the LLC, and finally adjusts the associativity if it suffers from a large number of conflict misses. The mechanism can achieve significant energy saving at the sacrifice of small performance degradation. The evaluation results show that our mechanism improves energy efficiency by 23.9% and 7.0% on average over the baseline and the cache-level bypassing mechanisms, respectively. In addition, even if the LLC resource contention occurs, the proposed mechanism is still effective for improving energy efficiency.},
keywords={},
doi={10.1587/transinf.2021EDP7201},
ISSN={1745-1361},
month={June},}
Copiar
TY - JOUR
TI - A Conflict-Aware Capacity Control Mechanism for Deep Cache Hierarchy
T2 - IEICE TRANSACTIONS on Information
SP - 1150
EP - 1163
AU - Jiaheng LIU
AU - Ryusuke EGAWA
AU - Hiroyuki TAKIZAWA
PY - 2022
DO - 10.1587/transinf.2021EDP7201
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E105-D
IS - 6
JA - IEICE TRANSACTIONS on Information
Y1 - June 2022
AB - As the number of cores on a processor increases, cache hierarchies contain more cache levels and a larger last level cache (LLC). Thus, the power and energy consumption of the cache hierarchy becomes non-negligible. Meanwhile, because the cache usage behaviors of individual applications can be different, it is possible to achieve higher energy efficiency of the computing system by determining the appropriate cache configurations for individual applications. This paper proposes a cache control mechanism to improve energy efficiency by adjusting a cache hierarchy to each application. Our mechanism first bypasses and disables a less-significant cache level, then partially disables the LLC, and finally adjusts the associativity if it suffers from a large number of conflict misses. The mechanism can achieve significant energy saving at the sacrifice of small performance degradation. The evaluation results show that our mechanism improves energy efficiency by 23.9% and 7.0% on average over the baseline and the cache-level bypassing mechanisms, respectively. In addition, even if the LLC resource contention occurs, the proposed mechanism is still effective for improving energy efficiency.
ER -