A Prediction Based CMP Cache Migration Policy

Song, Hao; Du, Z.; Bader, David A.; Wang, Man

doi:10.1109/hpcc.2008.83

Cited by 4 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A fundamental limitation in performance scaling is memory bandwidth, which depends on how programs share cache on a CMP. Hao et al proposed a novel processor design that used data migration in shared L2 cache to mitigate the NUMA effect [30]. Cache migration, although reduces latency, does not directly change the bandwidth demand of a group of tasks.…”

Section: Related Workmentioning

confidence: 99%

Cache Conscious Task Regrouping on Multicore Processors

Xiang

Bao

Ding

et al. 2012

2012 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (Ccgrid 2012)

View full text Add to dashboard Cite

Abstract-Because of the interference in the shared cache on multicore processors, the performance of a program can be severely affected by its co-running programs. If job scheduling does not consider how a group of tasks utilize cache, the performance may degrade significantly, and the degradation usually varies sizably and unpredictably from run to run.In this paper, we use trace-based program locality analysis and make it efficient enough for dynamic use. We show a complete on-line system for periodically measuring the parallel execution, predicting and ranking cache interference for all co-run choices, and reorganizing programs based on the prediction. We test our system on floating-point and mixed integer and floating-point workloads composed of SPEC 2006 benchmarks and compare with the default Linux job scheduler to show the benefit of the new system in improving performance and reducing performance variation.

show abstract

Section: Related Workmentioning

confidence: 99%

Cache Conscious Task Regrouping on Multicore Processors

Xiang

Bao

Ding

et al. 2012

2012 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (Ccgrid 2012)

View full text Add to dashboard Cite

show abstract

“…Song Hao [1] has defined a prediction based approach for the L2 cache management and to optimize the page access. The author has defined the work for multi processor cache system with shared cache.…”

Section: Review Of Literaturementioning

confidence: 99%

A Two Layered approach to perform Effective Page Replacement

Rohilla¹

2013

IOSRJEN

View full text Add to dashboard Cite

-In a multi processor system, when the request are performed more frequently, more effective management of cache is required. Cache is small size memory available in system that holds both data and instruction. In case of shared cache, the management of cache is more critical. Cache memory provides the static storage to the processor for most required data and instruction. As of limited size, the contents of cache are been replaced regularly. To improve the efficiency of the system, an effective cache replacement algorithm is required. The presented work is in same direction. In this work a two level cache is been defined. In first level the prioritization of the cache contents is been defined based on the type of contents. At the second level the replacement is been performed by analyzing the page utility and the frequency of the access. The analysis of the work is been performed with different cache size. The analysis of work is been performed in terms of hit ratio and miss ratio. The obtained results show that the presented work has improved the efficiency and reliability of the multiprocessor systems.

show abstract