WCSE 2017
ISBN: 978-981-11-3671-9 DOI: 10.18178/wcse.2017.06.001

The Advantages and Challenges of Spatial Computation

Jian Weng, Yanan Lu, Leibo Liu, Shouyi Yin, Shaojun Wei

Abstract— Recent years witness great interest to extract more performance and energy efficiency from spatial computation which gains great reputation on various domains, ranging from signal processing to highperformance computation. Its success does not only rely on the performance boosting, but also energy efficiency which is a severe problem for the popular portable and wearable devices. Spatial computation has gradually become a part of the mainstream computing infrastructure. This paper offers an observation on the issues of processors, reviews some recent researches, discusses the advantages of spatial computation on flexibility, utilization, scalability and energy efficiency, as well as the challenges from programmability, compatibility, locality, etc.

Index Terms— processor, spatial computation, performance, energy efficiency, etc.

Jian Weng, Yanan Lu, Leibo Liu, Shouyi Yin, Shaojun Wei
Institute of Microelectronics, Tsinghua University, CHINA

[Download]


Cite: Jian Weng, Yanan Lu, Leibo Liu, Shouyi Yin, Shaojun Wei, "The Advantages and Challenges of Spatial Computation," Proceedings of 2017 the 7th International Workshop on Computer Science and Engineering, pp. 1-7, Beijing, 25-27 June, 2017.