Management and Monitoring

Insights for the GPU Virtualization


NVIDIA virtual GPUs (vGPUs) give IT a complete lifecycle approach for managing the entire user experience—from infrastructure right-sizing to proactive management and issue remediation. Whether you’re managing virtual desktops or GPU clusters for compute-intensive server workloads, IT can use the same management tools from hypervisor and leading monitoring software vendors for deep insights into each virtual machine. Coupled with live migration capabilities, IT can now deliver reliable performance and high availability.


Virtual gpu solutions

Live Migration
Ensure continuous uptime by migrating live, GPU-accelerated VMs, with no impact to the user and no loss of data.

Simplified Deployment
Easily understand GPU utilization during the assessment and proof-of-concept phase and ensure an optimized performance from day one.

Proactive Management
Create thresholds that expose consumption trends impacting performance before they become an issue.

Support for Every GPU-Accelerated Workload
Support AI, Deep Learning, and Data Science workloads with standard virtualization tools and increase infrastructure utilization with GPU sharing and aggregation.

3 Ways To Integrate Your It Department With NVIDIA vGPU Monitoring

  • Leverage familiar tools, including perfmon and NVIDIA SMI (CLI).
  • Choose from an ecosystem of partner solutions that offer single-pane-of-glass management of the virtualized infrastructure—from the client to the server to the GPU and beyond. 
  • Use the open NVIDIA SDK, ideal for organizations who want to build their own infrastructure monitoring and management tools leveraging host, guest, and application graphics insights.

Quickly Enable Live Migration

Enable Live Migration In As Little As 2 Clicks!

Only NVIDIA supports live migration for GPUs. In GPU-enabled hypervisor-based virtualization environments, the IT administrator can move a running virtual machine from one physical server to another with minimal downtime and no loss of data. During the migration, users continue to have access to the applications they need and even compute-intensive workloads run uninterrupted.

Learn more about Live Migration >

A More Flexible And Agile It Department

Dramatically Lower Operational Costs

“Thanks to NVIDIA virtual GPU monitoring, we have a team of 3 people who easily manage 1,500 users daily between the Linux and Windows VDI environments. That’s a 500 to 1 management philosophy. You don’t get that anywhere. It’s a huge cost savings in administrative time.”

– Mike Bantz, Engineer and Technical Lead, Digital Globe

Download the Story > (PDF 1024KB)

Better Optimize Resources

“Honda’s IT departments can now allocate the right amount of vGPU necessary for the right users through a unified view of all physical GPUs. With NVIDIA virtual GPU monitoring, Honda can now proactively meet the requirements of both power users and knowledge workers based on more efficient operation.”

-Masashi Okubo, Assistant Chief Engineer, Honda R&D Co., Ltd.

Download the Story > (PDF 530KB)

Improve Time to Resolution

“NVIDIA virtual GPU monitoring has enabled us to respond to the needs of our users on the fly, with the ability to bump them up to something bigger. It’s a matter of two clicks and they’re on a machine sized perfectly for them.”

-Andrew Schilling, Chief Infrastructure Officer, CannonDesign

Download the Story > (PDF 1124KB)

Partner Integrations For GPU Management And Monitoring