Abstract

The Path to GPU as a Service in Kubernetes

Kubernetes modern production patterns for Deep Learning applications and a deep dive into the Kubernetes GPU subsystem and its challenges (performance, scheduling, monitoring). Autonomous vehicles, face recognition, High Performance Computing, Virtual Reality, NVIDIA GPUs are enabling a new computer era with cloud computing at its center. ....


Session No: SIL8143
Speaker: Adam Henryk Grzywaczweski
Type:

Developer Tools and Profesional Visualization

Date: Thursday - October 18, 2018 05:00 PM - 05:45 PM
Location: Hall D
Topics: Deep Learning and AI, Developer Tools
Industry: Software

Kubernetes modern production patterns for Deep Learning applications and a deep dive into the Kubernetes GPU subsystem and its challenges (performance, scheduling, monitoring). Autonomous vehicles, face recognition, High Performance Computing, Virtual Reality, NVIDIA GPUs are enabling a new computer era with cloud computing at its center. With kubernetes being the next iteration in cloud technologies, the NVIDIA container team with the kubernetes community is driving the advances in GPU integration. During this talk we will review how to deploy a GPU enabled Kubernetes and the modern production patterns for deploying GPU enabled services and applications. We will also dive into the details of the Kubernetes device plugin (its GPU subsystem), the NVIDIA container stack and the limitations provided by the kubernetes infrastructure. We will finally be discussing the latest improvements in the device plugin subsystem of Kubernetes, and the challenges ahead of it such as NUMA, sharing and scheduling.