Paper Title: Accelerating AI and Hyperscale Computing with PCIe 5.0 Switches

Paper Abstract: Applications such as data analytics, autonomous driving and medical diagnostics are increasing demands for Machine Learning (ML) and hyperscale compute infrastructure. Technology, such as PCI Express must continue to evolve to meet these performance demands, as PCIe switches continue to be the leading interconnectivity solution to serve as the bridge between CPU, GPU, FPGA, networking and memory. In this discussion, we will show how PCIe Gen 5 switches are accelerating AI/ML performance in the hyperscale computing space by doubling the data rate of PCIe Gen 4, while still delivering ultra-low latency and advanced diagnostics for a seamless compute interface.

Paper Author: Tam Do, Manager, Microchip Technology

Author Bio: Tam Do is a member of the technical product marketing team within the Data Center Solutions business unit at Microchip, where he is responsible for PCIe interface products. Previously, he worked on video codec and high-speed interface products for the consumer and multimedia markets. Tam has a Bachelor of Science in Electrical Engineering from the University of Nevada-Reno and an MBA from the University of Phoenix.