An Energy-Aware Exploration Approach Based on Open Software Environment

Silva-Filho, Abel G.; Eskinazi, Renan Harmes; Nascimento, Paulo Sérgio B.; Lima, Maria Luíza Carvalho de

doi:10.1007/11523277_10

Cited by 3 publications

(3 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We consider some optimization mechanisms in our analysis, are they: TECH [5], TCaT [4], TECH-CYCLES [6] and TEMGA [7]. We exhaustively simulated the design space defined for two-level cache hierarchies according to the experimental environment defined in the previous subsection in order to better evaluate the obtained results.…”

Section: Analysis Of the Resultsmentioning

confidence: 99%

“…The parameters are tested according to the following order S2, S1, L2, L1, A2 and A1. Previous researches [5] have shown that different order exploration can conduce the memory hierarchy energy consumption for different levels [5]. The best size for the second level of the cache memory (S2) is obtained by decreasing its value from a maximum (MAX) to a minimum (MIN).…”

Section: Instruction Cachesmentioning

confidence: 99%

“…After that, Silva-Filho et al also proposed some optimization mechanisms for two-level caches: TECH [5], TECH-CYCLES [6] and TEMGA [7]. The former indicates that the order of exploration of the cache parameters can have a considerable impact in the energy saving.…”

Section: Introductionmentioning

confidence: 97%

See 2 more Smart Citations

A methodology for tuning two-level cache hierarchy considering energy and performance

Silva-Filho

Araujo

2009

Proceedings of the 22nd Annual Symposium on Integrated Circuits and System Design: Chip on the Dunes

Self Cite

View full text Add to dashboard Cite

In this paper we propose a hybrid methodology for tuning both instruction and data cache configurations in a two-level memory hierarchy. The method aims to minimize energy consumption without compromise the performance. It combines two optimization mechanisms in order improve energy and performance results without increasing the configuration space. Experiments based on simulations were performed for 12 applications from the Mibench suite benchmark and the proposed methodology achieved better efficiency in 60% of the evaluated cases compared with existing heuristics.

show abstract