EMAXVR: A programmable accelerator employing near ALU utilization to DSA

EMAXVR: A programmable accelerator employing near ALU utilization to DSA
复制标题

DOI：
10.1109/coolchips.2018.8373078
复制
发表时间：
2018-04
期刊：
2018 IEEE Symposium in Low-Power and High-Speed Chips (COOL CHIPS)
影响因子：
--
通讯作者：
Takahiro Ichikura;Ryusuke Yamano;Yuma Kikutani;Renyuan Zhang;Y. Nakashima
Takahiro Ichikura;Ryusuke Yamano;Yuma Kikutani;Renyuan Zhang;Y. Nakashima
中科院分区：
其他
文献类型：
--
作者：
Takahiro Ichikura;Ryusuke Yamano;Yuma Kikutani;Renyuan Zhang;Y. Nakashima

文献摘要

相似文献

在许多真实的应用程序（例如深度学习）中，特定领域的特定加速器（DSA）似乎表现出色。但是，表演的好处是由于发展成本的增加和可编程性差而占据了一定的影响。在本文中，提出了一个可编程加速器，该加速器是通过将附近的Alu利用用于DSA提出的，DSA是我们先前报道的称为EMAXV的改进版本。结果，我们发现我们的可编程加速器可以计算Alexnet中的卷积操作，而与最新的灵活DSA相比，ALU的利用率仅降低17％，而数据重用速率仅14％。
The domain specific accelerators (DSAs) have appeared remarkable performances in many real-world applications such as deep learning. However, the benefit on performances is somehow eaten up by the increasing development cost and poor programmability. In this paper, a programmable accelerator is proposed by employing near ALU utilization to DSA, which is an improved version of our previously reported accelerator called EMAXV. As a result, we found our programmable accelerator can compute convolution operations in AlexNet with only 17% lower utilization of ALU and 14% upper rate of data reuse compared with the latest flexible DSA.