Learn about Meta's next generation AI hardware platform for training and inference use cases in this technical talk presented by Meta engineers Matt Bowman and Jeremy Baumgartner. Explore the modular system design of Grand Teton while discovering key design challenges and their innovative solutions. Gain insights into the architectural decisions and engineering approaches that shaped this cutting-edge AI infrastructure platform during this 14-minute presentation from the Open Compute Project.
Grand Teton - Meta's Next Generation AI Hardware Platform for Training and Inference
Open Compute Project via YouTube
Overview
Syllabus
Grand Teton Systems Overview
Taught by
Open Compute Project
Related Courses
-
Supporting Meta ML Accelerators on the Grand Teton Platform
-
PCIe Express Error Handling and RAS Solutions for AI/ML Training Clusters
-
Using Modularity to Enable Hardware Reuse Across AI Platforms in a Rapidly Evolving Ecosystem
-
Data Center-Wide Rack Management and Beyond: Exploring Scalable Solutions for Gen AI Infrastructure
-
Meta's Llama 3 - The Next Generation of Open-Source Language Models
-
Optical Interconnects for Large-Scale AI Clusters - A Meta Perspective
