|SHC RISC-V Demo: Features Andes N22 and SiFive HiFive1 RISC-V connectivity with Amazon AWS IoT Service Platform|
|Adopted Chipyard Framework for Bayesian Neural Networks Training Accelerator Development|
|Please press here for pdf demo document.|
A powerful development environment is essential for co-development of dedicated hardware and new algorithms in small teams. Chipyard is a framework developed by UC Berkeley that makes it easy to develop RISC-V based custom SoCs. It contains an integrated SoC design, simulation, and implementation environment. We are developing a novel Bayesian neural network training accelerator using the Chipyard framework. In this session I give a brief introduction to our recent R&D project using the RISC-V ecosystem.