A second-level cache with the distance-aware replacement policy for NUMA systems

Sung Woo Jung, Jeong Heon Shin, Hyong Shik Kim, Chu Shik Jhon

Research output: Contribution to journalArticle

Abstract

The cache replacement policy is one of the most important factors that affect the cache performance. With the trend of increasing associativity in second-level caches, implementing an efficient replacement algorithm becomes more important than just eliminating conflict misses. The LRU cache replacement algorithm has been known to work well in a single processor system by reducing the cache miss rate, but it does not minimize the cache replacement cost on the interconnection for a multiprocessor system because it does not take the distance into account. In this paper, we suggest a distance-aware second level (L2) cache for scalable multiprocessors, which is composed of a traditional LRU cache and an additional SDF (Shortest Distance First) cache. The LRU cache selects a victim using age information, while the SDF cache does so using distance information. Both work together to minimize the overall replacement cost by keeping long-distance blocks as well as recently used blocks. The combined L2 cache reduces the cache miss rate compared to the original LRU cache in many cases. With 32 processors, a 512KB LRU/SDF L2 cache outperforms a 512KB LRU L2 cache. Moreover, the replacement traffic on an interconnection network such as the ring is suppressed by up to 69%, which is expected to bring more scalability to multiprocessor systems.

Original languageEnglish
Pages (from-to)803-813
Number of pages11
JournalJournal of Information Science and Engineering
Volume18
Issue number5
Publication statusPublished - 2002 Sep 1
Externally publishedYes

Fingerprint

interconnection
Scalability
Costs
costs
traffic
trend
performance

Keywords

  • Cache miss rate
  • Cache replacement policy
  • Memory hierarchy
  • Parallel architecture
  • Performance evaluation

ASJC Scopus subject areas

  • Information Systems

Cite this

A second-level cache with the distance-aware replacement policy for NUMA systems. / Jung, Sung Woo; Shin, Jeong Heon; Kim, Hyong Shik; Jhon, Chu Shik.

In: Journal of Information Science and Engineering, Vol. 18, No. 5, 01.09.2002, p. 803-813.

Research output: Contribution to journalArticle

Jung, Sung Woo ; Shin, Jeong Heon ; Kim, Hyong Shik ; Jhon, Chu Shik. / A second-level cache with the distance-aware replacement policy for NUMA systems. In: Journal of Information Science and Engineering. 2002 ; Vol. 18, No. 5. pp. 803-813.
@article{ba5e8ab42d994e0bae68175f4164666d,
title = "A second-level cache with the distance-aware replacement policy for NUMA systems",
abstract = "The cache replacement policy is one of the most important factors that affect the cache performance. With the trend of increasing associativity in second-level caches, implementing an efficient replacement algorithm becomes more important than just eliminating conflict misses. The LRU cache replacement algorithm has been known to work well in a single processor system by reducing the cache miss rate, but it does not minimize the cache replacement cost on the interconnection for a multiprocessor system because it does not take the distance into account. In this paper, we suggest a distance-aware second level (L2) cache for scalable multiprocessors, which is composed of a traditional LRU cache and an additional SDF (Shortest Distance First) cache. The LRU cache selects a victim using age information, while the SDF cache does so using distance information. Both work together to minimize the overall replacement cost by keeping long-distance blocks as well as recently used blocks. The combined L2 cache reduces the cache miss rate compared to the original LRU cache in many cases. With 32 processors, a 512KB LRU/SDF L2 cache outperforms a 512KB LRU L2 cache. Moreover, the replacement traffic on an interconnection network such as the ring is suppressed by up to 69{\%}, which is expected to bring more scalability to multiprocessor systems.",
keywords = "Cache miss rate, Cache replacement policy, Memory hierarchy, Parallel architecture, Performance evaluation",
author = "Jung, {Sung Woo} and Shin, {Jeong Heon} and Kim, {Hyong Shik} and Jhon, {Chu Shik}",
year = "2002",
month = "9",
day = "1",
language = "English",
volume = "18",
pages = "803--813",
journal = "Journal of Information Science and Engineering",
issn = "1016-2364",
publisher = "Institute of Information Science",
number = "5",

}

TY - JOUR

T1 - A second-level cache with the distance-aware replacement policy for NUMA systems

AU - Jung, Sung Woo

AU - Shin, Jeong Heon

AU - Kim, Hyong Shik

AU - Jhon, Chu Shik

PY - 2002/9/1

Y1 - 2002/9/1

N2 - The cache replacement policy is one of the most important factors that affect the cache performance. With the trend of increasing associativity in second-level caches, implementing an efficient replacement algorithm becomes more important than just eliminating conflict misses. The LRU cache replacement algorithm has been known to work well in a single processor system by reducing the cache miss rate, but it does not minimize the cache replacement cost on the interconnection for a multiprocessor system because it does not take the distance into account. In this paper, we suggest a distance-aware second level (L2) cache for scalable multiprocessors, which is composed of a traditional LRU cache and an additional SDF (Shortest Distance First) cache. The LRU cache selects a victim using age information, while the SDF cache does so using distance information. Both work together to minimize the overall replacement cost by keeping long-distance blocks as well as recently used blocks. The combined L2 cache reduces the cache miss rate compared to the original LRU cache in many cases. With 32 processors, a 512KB LRU/SDF L2 cache outperforms a 512KB LRU L2 cache. Moreover, the replacement traffic on an interconnection network such as the ring is suppressed by up to 69%, which is expected to bring more scalability to multiprocessor systems.

AB - The cache replacement policy is one of the most important factors that affect the cache performance. With the trend of increasing associativity in second-level caches, implementing an efficient replacement algorithm becomes more important than just eliminating conflict misses. The LRU cache replacement algorithm has been known to work well in a single processor system by reducing the cache miss rate, but it does not minimize the cache replacement cost on the interconnection for a multiprocessor system because it does not take the distance into account. In this paper, we suggest a distance-aware second level (L2) cache for scalable multiprocessors, which is composed of a traditional LRU cache and an additional SDF (Shortest Distance First) cache. The LRU cache selects a victim using age information, while the SDF cache does so using distance information. Both work together to minimize the overall replacement cost by keeping long-distance blocks as well as recently used blocks. The combined L2 cache reduces the cache miss rate compared to the original LRU cache in many cases. With 32 processors, a 512KB LRU/SDF L2 cache outperforms a 512KB LRU L2 cache. Moreover, the replacement traffic on an interconnection network such as the ring is suppressed by up to 69%, which is expected to bring more scalability to multiprocessor systems.

KW - Cache miss rate

KW - Cache replacement policy

KW - Memory hierarchy

KW - Parallel architecture

KW - Performance evaluation

UR - http://www.scopus.com/inward/record.url?scp=0036755882&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036755882&partnerID=8YFLogxK

M3 - Article

VL - 18

SP - 803

EP - 813

JO - Journal of Information Science and Engineering

JF - Journal of Information Science and Engineering

SN - 1016-2364

IS - 5

ER -