Systematic Embedded Development and Implementation Techniques on Intel Myriad VPUs
2022 IFIP/IEEE 30th International Conference on Very Large Scale Integration (VLSI-SoC)(2022)
摘要
The worldwide demand for speed in applications challenges the deployment of compute-intensive algorithms at the power-constrained edge. Novel embedded devices such as the heterogeneous Vision Processing Units (VPUs) emerge as a promising solution for low-power embedded imaging/vision applications, as they accelerate computer vision algorithms and convolutional neural networks with only 1-2W. In this brief, we propose a development methodology for exploiting the full potential of the VPU heterogeneity and providing sufficient acceleration within their restricted power envelope. Based on this methodology, we demonstrate the development paradigm on the Myriad VPUs and report experimental results from the implementation of demanding image processing kernels.
更多查看译文
关键词
systematic embedded development,intel Myriad VPUs,worldwide demand,compute-intensive algorithms,power-constrained edge,computer vision algorithms,convolutional neural networks,VPU heterogeneity,sufficient acceleration,image processing kernels,heterogeneous vision processing units,low-power embedded imaging
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要