Abstract

Deep learning processors (DLPs) have become the de facto processing architecture in today's AI computing systems. Many DLPs have been proposed and used in the past few years. In fact, a DLP is simply an integrated circuit specially designed for deep learning algorithms, while leveraging their unique features for high performance and efficiency. To clearly reveal the DLP inside, this chapter walks through the entire design process of a DLP from the initial algorithm analysis to detailed component design, pipeline arrangement, and the final algorithm mapping. After this chapter, readers should be able to and are suggested to design their own DLPs using the same design philosophy.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call