IEICE TRANSACTIONS on Information
Adaptive Insertion and Promotion Policies Based on Least Recently Used Replacement
Summary :
To improve Last-Level Cache (LLC) management, numerous approaches have been proposed requiring additional hardware budget and increased overhead. A number of these approaches even change the organization of the existing cache design. In this letter, we propose Adaptive Insertion and Promotion (AIP) policies based on Least Recently Used (LRU) replacement. AIP dynamically inserts a missed line in the middle of the cache list and promotes a reused line several steps left, realizing the combination of LRU and LFU policies deliberately under a single unified scheme. As a result, it benefits workloads with high locality as well as with many frequently reused lines. Most importantly, AIP requires no additional hardware other than a typical LRU list, thus it can be easily integrated into the existing hardware with minimal changes. Other issues around LLC such as scans, thrashing and dead lines are all explored in our study. Experimental results on the gem5 simulator with SPEC CUP2006 benchmarks indicate that AIP outperforms LRU replacement policy by an average of 5.8% on the misses per thousand instructions metric.
- Publication
- IEICE TRANSACTIONS on Information Vol.E96-D No.1 pp.124-128
- Publication Date
- 2013/01/01
- Publicized
- Online ISSN
- 1745-1361
- DOI
- 10.1587/transinf.E96.D.124
- Type of Manuscript
- LETTER
- Category
- Computer System
Authors
Keyword
Latest Issue
Copyrights notice of machine-translated contents
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.
Cite this
Copy
Wenbing JIN, Xuanya LI, Yanyong YU, Yongzhi WANG, "Adaptive Insertion and Promotion Policies Based on Least Recently Used Replacement" in IEICE TRANSACTIONS on Information,
vol. E96-D, no. 1, pp. 124-128, January 2013, doi: 10.1587/transinf.E96.D.124.
Abstract: To improve Last-Level Cache (LLC) management, numerous approaches have been proposed requiring additional hardware budget and increased overhead. A number of these approaches even change the organization of the existing cache design. In this letter, we propose Adaptive Insertion and Promotion (AIP) policies based on Least Recently Used (LRU) replacement. AIP dynamically inserts a missed line in the middle of the cache list and promotes a reused line several steps left, realizing the combination of LRU and LFU policies deliberately under a single unified scheme. As a result, it benefits workloads with high locality as well as with many frequently reused lines. Most importantly, AIP requires no additional hardware other than a typical LRU list, thus it can be easily integrated into the existing hardware with minimal changes. Other issues around LLC such as scans, thrashing and dead lines are all explored in our study. Experimental results on the gem5 simulator with SPEC CUP2006 benchmarks indicate that AIP outperforms LRU replacement policy by an average of 5.8% on the misses per thousand instructions metric.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E96.D.124/_p
Copy
@ARTICLE{e96-d_1_124,
author={Wenbing JIN, Xuanya LI, Yanyong YU, Yongzhi WANG, },
journal={IEICE TRANSACTIONS on Information},
title={Adaptive Insertion and Promotion Policies Based on Least Recently Used Replacement},
year={2013},
volume={E96-D},
number={1},
pages={124-128},
abstract={To improve Last-Level Cache (LLC) management, numerous approaches have been proposed requiring additional hardware budget and increased overhead. A number of these approaches even change the organization of the existing cache design. In this letter, we propose Adaptive Insertion and Promotion (AIP) policies based on Least Recently Used (LRU) replacement. AIP dynamically inserts a missed line in the middle of the cache list and promotes a reused line several steps left, realizing the combination of LRU and LFU policies deliberately under a single unified scheme. As a result, it benefits workloads with high locality as well as with many frequently reused lines. Most importantly, AIP requires no additional hardware other than a typical LRU list, thus it can be easily integrated into the existing hardware with minimal changes. Other issues around LLC such as scans, thrashing and dead lines are all explored in our study. Experimental results on the gem5 simulator with SPEC CUP2006 benchmarks indicate that AIP outperforms LRU replacement policy by an average of 5.8% on the misses per thousand instructions metric.},
keywords={},
doi={10.1587/transinf.E96.D.124},
ISSN={1745-1361},
month={January},}
Copy
TY - JOUR
TI - Adaptive Insertion and Promotion Policies Based on Least Recently Used Replacement
T2 - IEICE TRANSACTIONS on Information
SP - 124
EP - 128
AU - Wenbing JIN
AU - Xuanya LI
AU - Yanyong YU
AU - Yongzhi WANG
PY - 2013
DO - 10.1587/transinf.E96.D.124
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E96-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2013
AB - To improve Last-Level Cache (LLC) management, numerous approaches have been proposed requiring additional hardware budget and increased overhead. A number of these approaches even change the organization of the existing cache design. In this letter, we propose Adaptive Insertion and Promotion (AIP) policies based on Least Recently Used (LRU) replacement. AIP dynamically inserts a missed line in the middle of the cache list and promotes a reused line several steps left, realizing the combination of LRU and LFU policies deliberately under a single unified scheme. As a result, it benefits workloads with high locality as well as with many frequently reused lines. Most importantly, AIP requires no additional hardware other than a typical LRU list, thus it can be easily integrated into the existing hardware with minimal changes. Other issues around LLC such as scans, thrashing and dead lines are all explored in our study. Experimental results on the gem5 simulator with SPEC CUP2006 benchmarks indicate that AIP outperforms LRU replacement policy by an average of 5.8% on the misses per thousand instructions metric.
ER -
English
