This document provides insights into deploying NVIDIA Omniverse Enterprise for VMWare vSphere and serves as a technical resource for understanding system pre-requisites, installation, and configuration.
The chapters within this guide were written in the following installation order:
VMware vSphere ESXi Hypervisor
NVIDIA vGPU Manager (vib)
vGPU software license server
Once the installation of the above components are complete, the guide will cover vGPU configuration items such as GPU portioning and scheduling options. This document also covers detailed steps for creating and configuring your first compute VM as well as installing Docker and getting started with Deep Learning benchmarking. For more information regarding architectural options for running compute workflows within VMs, please refer to the Multi-Node Training Application Guide. Before you begin, download the following software from the NVIDIA Licensing Portal:
NVIDIA vGPU software for VMware vSphere 7.0
NVIDIA vGPU software license server