Pre-requisites for Monitoring NVIDIA GPUs
VMware vSphere employs the NVIDIA GRID Virtual GPU (vGPU) technology to provide exceptional graphics performance for virtual desktops. NVIDIA GRID vGPU enables multiple Virtual Machines (VM) to have simultaneous, direct access to a single physical GPU, using the same NVIDIA graphics drivers that are deployed on non-virtualized Operating Systems. Under the control of NVIDIA's GRID Virtual GPU Manager, GRID physical GPUs are capable of supporting multiple virtual GPU devices (vGPUs) that can be assigned directly to VMs.
To determine whether/not the vGPUs assigned to a VM are adequate for the graphic processing requirements of the that VM, administrators must understand whether/not memory-intensive I/O operations are performed on that VM and if so, how they impact vGPU usage. The tests run by the eG remote agent provide these much-needed usage insights!
To run these tests, the following pre-requisites should be fulfilled:
-
NVWMI should be installed on the vSphere host and on each of the VMs. To know how, refer to
-
Only Shared GPU or Virtual GPU (vGPU) Technology should be used to deliver GPU resources to virtual desktops/VMs.
-
SSH should be enabled on the vSphere host, so that the eG agent can connect to the host via SSH and pull desired NVIDIA GPU metrics. To know how to enable SSH on the virtual host, follow the steps detailed in
