About the talk: The landscape of AI is undergoing rapid transformation, marked by the escalating resource demands of new models. This requires specialized computer architectures that deliver robust performance within constrained power and cost footprints. Additionally, AI applications are expanding ubiquitously, spanning from edge devices to cloud infrastructure and high-performance computing systems. The AMD AI Engine addresses these demands by providing a scalable AI acceleration architecture focused on AI inference, targeting a wide spectrum of platforms from laptops to data centers. It achieves its goals through optimized compute and efficient data movement to maximize data reuse and latency tolerance. This presentation will encompass the diverse range of systems targeted by the AI Engine, providing an overview of its architecture, and highlighting key features that enable its high-performance and energy efficiency.

About the speaker: Juanjo Noguera currently serves as the Senior Director of AIE Architecture at AMD. He became part of AMD through the acquisition of Xilinx in 2022. Juanjo initially joined Xilinx in 2006, working in the CTO office, and later transitioned to the product engineering division, where he played a pivotal role in bringing the AIE architecture to market. Prior to his tenure at Xilinx, Juanjo held a position at Hewlett-Packard. He holds a Ph.D. in Computer Architecture from the Technical University of Catalunya (UPC, Barcelona, Spain). Juanjo has published numerous technical papers at international conferences and holds over twenty granted patent applications.