文章基本信息

标题：Artificial Neural Network Stream Processing Core (ANnSP) Implementation in Embedded Systems
本地全文：下载
作者：JNANA RANJAN TRIPATHY ; HRUDAYA KUMAR TRIPATHY
期刊名称：Oriental Journal of Computer Science and Technology
印刷版ISSN：0974-6471
出版年度：2013
卷号：6
期号：2
页码：9-14
出版社：Oriental Scientific Publishing Company
摘要：ANnSP is a stream-based programmable and code-level statically recon.gurableprocessor for realization of neural networks in embedded systems. ANnSP is provided with aneural-network-to-stream compiler and a hardware core builder. The ANnSP stream compilermakes it possible to realize various neural networks using ANnSP. On the other hand, the ANnSPbuilder makes the ANnSP processor an IP core that can be restructured to satisfy differentdemands and constraints. This paper presents the architecture of the ANnSP processor, thestreaming mechanism, and the builder facilities. Also, synthesis results of a 64-PE ANnSP on a0.18μm standard-cell library are presented. The obtained results show that a 64-PE ANnSP canperform computations of 25.6 giga connections in a second, while its throughput is upto 51.2giga 32-bit .xed point operations per second. Comparing with high performance parallelarchitectures locates 64-PE ANnSP among the best state of the art parallel processors
关键词：ANN; Embedded Systems; Stream Processing; LIFO; FIFO