Watch this 13-minute keynote presentation from the 2025 OCP EMEA conference where Ram Velaga, SVP & GM of the Core Switching Group at Broadcom, discusses the critical balance between compute, memory, and networking infrastructure required for effective AI implementation. Learn about Broadcom's perspective on optimizing these interconnected components to address the unique challenges and requirements of modern AI workloads.
Balancing Compute Memory and Networking for AI - 2025 OCP EMEA Keynote
Open Compute Project via YouTube
Overview
Syllabus
2025 OCP EMEA Keynote - Balancing Compute Memory & Networking for AI (Broadcom)
Taught by
Open Compute Project
Related Courses
-
Data Center Evolution Powering the AI Revolution Sustainably
-
Accelerating AI with Next-Gen Networking: SONiC Innovations and Scalable Designs
-
CXL 2.0 Switch Enabling Composable Memory Architecture in AI and HPC Computing
-
Leveraging OCP for Sovereign AI Plans
-
Near Memory Compute for AI Inferencing - Optimizing Data Center Design and TCO
-
Empowering AI Networking with SONiC - Building Robust Network Infrastructure
