Genaro Costa
Affiliation: Phd at EVIDEN
Country: Brazil
Title: About the exascale HPCs and its AI workloads
Abstract
The exascale barrier has been surpassed. The Top500 now lists two machines with this level of performance, and more are expected soon. Most of these machines are designed to target AI workloads. On the positive side, AI workloads have driven the recognition of the value of larger machines, increasing the demand for enhanced network connectivity and storage capacity. However, significant challenges remain, not only in silicon design but also in cooling efficiency and overall system organization. In this talk, we will present the latest technologies in HPC components for AI and discuss their potential future developments.
Bio
He works as an HPC Distinguished Expert at EVIDEN, coordinating the R&D Labs at Eviden in SENAI CIMATEC. As an Adjunct Professor at UFBA, he was the coordinator of the Interdisciplinary bachelor’s in science and technology and the vice-director of the Professor Milton Santos Institute of Humanities, Arts, and Sciences. He holds a degree in Computer Science from UCSal, and a master’s and doctorate in Informatics from the Universitat Autónoma de Barcelona (UAB). His research interests include High-Performance Computing, Artificial Intelligence, Big Data, and Quantum Computing.