Masayuki Ito

Denso, Japan

Architecture Challenges and Platform Strategy for Heterogeneous Processors in Embedded SoCs

Abstract

Efficient execution of AI and other computing on edge devices has become an important issue in embedded systems due to strict heat and cost constraints. Special HW logic IPs are not suitable due to its specific dedication. Server/HPC processors sometimes do not fit to embedded feature. Generic and flexible processors with high efficiency are required for such embedded systems. There are a couple of important architecture challenges to efficiently accelerate the computation of wide range of applications. Firstly, such processors need to solve both high-load AI computation and legacy non-symmetric control and signal processing and so well-balanced heterogeneity and good parallelism should be considered. Secondly, such processors need to achieve both high power efficiency and high silicon efficiency comparable to dedicated HW logic IPs. Finally, such processors need to contribute to efficient hardware and software development and so strategic processor platform should be considered. This talk will review some architecture challenges to such new generation processors along with strategic processor IP platform and share some viewpoints on such solutions and techniques being explored to efficiently accelerate the computation of wide range of applications for the embedded systems.

Biography

Dr. Masayuki Ito is a Project Director in Semiconductor-Originated Key Technology R&D Division at DENSO, where he focuses on developments and promotions for its own RISC-V processors and AI accelerators targeting not only for automobile but also for other industries like Factory-Automation and IoT. He has more than 25 years of experience in the semiconductor industry. He started his career in the Semiconductor Development Center in Hitachi Ltd. in Tokyo, followed by Renesas Technology Corp. and Renesas Electronics Inc. Throughout his career he consistently worked for processor and SoC development for mobile and automobile. He holds a master’s degree in Computer Engineering from Kyoto University and a Ph.D. in Science and Engineering from Tokyo Institute of Technology.