Cache Memory Energy Efficiency Exploration for the HEVC Motion Estimation

Martins, Anderson; Penny, Wagner; Weber, Matheus; Palomino, Daniel; Mattos, Júlio C. B.; Porto, Marcelo; Agostini, Luciano

doi:10.1109/sbesc.2017.11

Cited by 3 publications

(2 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, modern information systems, which are data-intensive in terms of their business model, tend to use cache memories alongside the persistent storage. In a similar fashion, processing in memory (PIM) has been used with in-memory computing for processing large data-intensive applications such as machine learning, graph processing, social network analysis, and image processing (Ahn et al, 2015;Chi et al, 2016;Fattahi et al, 2019;Martins et al, 2017;Nai et al, 2017). In addition to providing faster access to the data, cache memory also plays a critical role in power reduction (Panda et al, 2016).…”

Section: Introductionmentioning

confidence: 99%

Java Ön Bellek Gerçekleştirmelerinin Bir Performans Karşılaştırması

Kabakuş¹

2020

Gümüşhane Üniversitesi Fen Bilimleri Enstitüsü Dergisi

View full text Add to dashboard Cite

Nowadays the information systems are substantially data-intensive and the data is going to be more critical than before. For these systems, which are intolerant in terms of time latency, the way of accessing data becomes more critical. In these situations, an additional data layer named cache memory is used. There are various both open-source and commercial Java cache memory implementations based on the specifications defined by Java Community Process. In this study, the most widely used Java cache memory implementations are evaluated in order to compare their performances in terms of elapsed time and memory consumption. The experimental results imply that the architectural design of cache memory has a great effect on performance and there is no winner that provides the best performance for all data operations.

show abstract

Section: Introductionmentioning

confidence: 99%

Java Ön Bellek Gerçekleştirmelerinin Bir Performans Karşılaştırması

Kabakuş¹

2020

Gümüşhane Üniversitesi Fen Bilimleri Enstitüsü Dergisi

View full text Add to dashboard Cite

show abstract

“…Along the years, many optimizations were proposed to video coding standards, and such optimizations vary in a myriad of ways. Among such ways, there are proposals to reduce the complexity and/or energy consumption of video encoding [7][8], explore parallel systems more properly [9][10], optimize memory usage [11], among others.…”

Section: Introductionmentioning

confidence: 99%

Impacts of Equirectangular 360-degrees Videos in the Intra-Frame Prediction of HEVC

Storch

Cruz

Agostini

et al. 2019

JICS

Self Cite

View full text Add to dashboard Cite

Recent technological advancements allowed videos to come from a simple sequence of 2D images to be displayed in a flat screen display into spherical representations of one’s surroundings, capable of creating a realistic immersive experience when allied to head-mounted displays. In order to explore the existing infrastructure for video coding, 360-degrees videos are pre-processed and then encoded by conventional video coding standards. However, the flattened version of 360-degrees videos present some peculiarities which are not present in conventional videos, and therefore, may not be properly exploited by conventional video coders. Aiming to find evidence that conventional video encoders can be adapted to perform better over 360-degrees videos, this work performs an evaluation on the intra-frame prediction performed by the High Efficiency Video Coding over 360-degrees videos in the equirectangular projection. Experimental results point that 360-degrees videos present spatial properties that make some regions of the frame likely to be encoded using a reduced set of prediction modes and block sizes. This behavior could be used in the development of fast decision and energy saving algorithms by evaluating a reduced set of prediction modes and block sizes depending on the regions of the frame being encoded.

show abstract