arrow
返回

Low-energy motion estimation memory system with dynamic management

delete2021-06-11
delete1
delete
OA
AI
D
Dieison Silveira *
L
Lívia Amaral
G
Guilherme Povala
Z
Zatt, Bruno
L
Luciano Agostini
M
Marcelo Porto
S
Sérgio Bampi
DOI:10.1007/s11554-021-01138-3delete
delete原文链接
delete分享
delete收藏
查看原文
摘要

摘要

En 中文
The digital video coding process imposes severe pressure on memory traffic, leading to considerable power consumption related to frequent DRAM accesses. External off-chip memory demand needs to be minimized by clever architecture/algorithm co-design, thus saving energy and extending battery lifetime during video encoding. To exploit temporal redundancies among neighboring frames, the motion estimation (ME) algorithm searches for good matching between the current block and blocks within reference frames stored in external memory. To save energy during ME, this work performs memory accesses distribution analysis of the test zone search (TZS) ME algorithm and, based on this analysis, proposes both a multi-sector scratchpad memory design and dynamic management for the TZS memory access. Our dynamic memory management, called neighbor management, reduces both static consumption-by employing sector-level power gating-and dynamic consumption-by reducing the number of accesses for ME execution. Additionally, our dynamic management was integrated with two previously proposed solutions: a hardware reference frame compressor and the Level C data reuse scheme (using a scratchpad memory). This system achieves a memory energy consumption savings of 99.8% and, when compared to the baseline solution composed of a reference frame compressor and data reuse scheme, the memory energy consumption was reduced by 44.1% at a cost of just 0.35% loss in coding efficiency, on average. When compared with related works, our system presents better memory bandwidth/energy savings and coding efficiency results.
Keyword:
Video coding
Motion estimation
Test zone search
Dynamic memory management
Energy optimization
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

Journal of Real-Time Image Processing 封面图
Journal of Real-Time Image Processing
IF:
3
论文数:
391
被引数:
2.0K

机构

U
Universidade Federal do Rio Grande do Sul
学者数:
2.6W
论文数: 1.7W
被引数: 1.6W
Universidade Federal de Pelotas 封面图
Universidade Federal de Pelotas
学者数:
6.7K
论文数: 4.2K
被引数: 3.9K
引用论文

引用论文

err分享
err收藏
err分享
err收藏
Pro-rich Antimicrobial Peptides from Animals: Structure, Biological Functions and Mechanism of Action
err2002-04-01
err0
PREAI
errRenato Gennaro; Margherita Zanetti; Monica Benincasa; Elena Podda; Monica Miani
err分享
err收藏
Common Textbook and Teaching Misrepresentations of Lewis Structures
err1995-07-01
err0
PREAI
errLaila Suidan; Jay K. Badenhoop; Eric D. Glendening; Frank Weinhold
err分享
err收藏
Persistence of Human Parvovirus B19 in Multipotent Mesenchymal Stromal Cells Expressing the Erythrocyte P Antigen: Implications for Transplantation
err2008-10-01
err0
errOAAI
errMikael Sundin; Anna Lindblom; Claes Örvell; A. John Barrett; Berit Sundberg; Emma Watz; Agneta Wikman; Kristina Broliden; Katarina Le Blanc
err分享
err收藏
学者 查看更多内容