Blog

Accelerating Private Clouds with NVIDIA GPUs on Nutanix AHV

Announcing advanced GPU technologies to boost your VDI, HPC and AI applications

By Luke Congdon

 

Nutanix and NVIDIA are breaking the boundaries for End User Computing (EUC) and virtualized computing with GPU-accelerated performance. The Nutanix Cloud platform delivers a powerful, flexible, and reliable platform for the full spectrum of desktop virtualization and GPU-accelerated server workloads. 

“From knowledge workers to AI researchers and data scientists, professionals today are choosing GPU-accelerated virtualized computing to meet the increasing demand of their graphics and compute workloads,” said Anne Hecht, senior director of product marketing for virtual GPU software at NVIDIA. “The Nutanix AHV platform makes it easy for IT teams to manage and configure GPUs, ensuring users can get the optimal experience of vGPU software, along with the enhanced performance and flexibility they need to get their important work done.”

Over the years, the GPU use case has evolved from high density VDI workloads to high performance GPU-intensive applications, such as machine learning (ML), artificial intelligence (AI), deep learning AI, high-end complex design and visualization, and advanced rendering and ray tracing. High density workloads need support for a range of users--from hundreds of general knowledge workers and task workers, to power users. In contrast, high performance workloads demand support for fewer specialized users per compute and expansion modules. 

GPU-accelerated clouds allow elastic environments to co-host GPU-heavy applications (such as visualization and rendering) and compute workloads (AI, ML) at the same time.

The Nutanix AHV platform simplifies GPU configuration and allows users to build general purpose infrastructure to improve performance and agility, while reducing TCO. Nutanix AHV fully supports NVIDIA’s industry-leading virtual GPU software editions for visual graphics, computing, and GPU acceleration.

New GPU Feature Announcement on Nutanix AHV

We are thrilled to announce the general availability of the following capabilities in Nutanix AHV, enabling seamless support for modern GPU applications. 

Multi-GPU per VM

Nutanix AHV has extended its GPU operational configuration range to support multiple vGPU instances for a single VM, benefitting from the technology provided by NVIDIA cards (this functionality was part of the AOS 5.18 release in August 2020). This offers flexibility during VM creation to assign multiple vGPU instances to a VM, based on workload needs, in a seamless and scalable manner. This also allows users to bring up multiple apps that rely on compute and graphics in the VM.

Figure 1: GPU Modes in Nutanix AHV

There are many possible use cases for this new capability, including 

  1. Accelerating computation performance of real-time simulations for product design.
  2. Deep learning for ground-breaking research and HPC applications.  
  3. Running multiple GPU-intensive applications, such as rendering workflow and ANSYS Discovery Live to run real-time interactive simulations. This also helps in potentially sharing GPU resources.

The Nutanix Prism management interface extends the existing UI to allow the configuring of multiple vGPU instances for a VM.

Figure 2: Multiple vGPU UI (PE) I

vGPU Live Migration 

Starting with the AOS 5.18.1 release (September 2020), Nutanix AHV has enabled the ability to live-migrate GPU-accelerated VMs from one node to another within a Nutanix cluster, without incurring workload downtime. This eliminates the need for significant resource planning, testing, and scheduling prior to performing lifecycle maintenance activities. With live migration, users can manage their hardware infrastructure easily and perform upgrades and maintenance seamlessly without disrupting running VMs. 

The vGPU VM live migration capability on Nutanix AHV is available with:

  • NVIDIA vGPU software version 10.1 (440.53) or later, and
  • On AOS 5.18.1 or later

vGPU Live migration of VMs provides the following benefits: 

  • Minimal Business Disruptions: Allows server maintenance activities, such as hardware replacement, upgrades, or software updates, without scheduled downtime.
  • Improved Server Density: Provides the ability to manually load balance users and consolidate common profile types for better density.
  • Optimized Infrastructure Utilization: Optimize VM placement around GPU location and time of day to maximize GPU availability based on the workload-- for example, graphics by day and computation by night.
  • Graphics by Day/Compute by Night: Historically, in use cases such as VDI, when desktop users leave at the end of the workday, a large amount of invested CapEx for GPU would potentially remain underutilized. According to NVIDIA, GPUs can be 20X or more faster than CPUs and hence are extremely useful in carrying out complex calculations and advanced simulations. The above features can thus provide flexibility in running VDI workloads by day and compute workloads by night when knowledge workers log off at the end of the day. This results in better utilization of GPU resources, resulting in increased ROI for your CapEx investments.

Figure 3: Graphics by Day/Compute by Night

NVIDIA® Quadro RTX™ 6000/8000 

Starting with AOS 5.19, the AHV platform enables support for NVIDIA’s advanced ray tracing capable hardware cards: RTX 6000 and RTX 8000. The NVIDIA® Quadro RTX™ 6000, powered by the NVIDIA Turing™ architecture and the NVIDIA RTX™ platform, brings the most significant advancement in computer graphics in over a decade to professional workflows. Designers and artists can now wield the power of hardware accelerated ray tracing, deep learning, and advanced shading to dramatically boost productivity and create amazing content faster than ever before. Refer to our documentation to find out which platforms support the NVIDIA® Quadro RTX™ 6000 and RTX 8000. 

NVIDIA Host Driver Download for AHV

AHV-compatible NVIDIA host drivers are now downloadable on the Nutanix Portal customer support site. Customers that are using vGPU licenses can now easily find and download the correct AHV host drivers required per version of Nutanix AOS.

Boost the Mileage on your GPU Investment with Nutanix AHV

The availability of features such as multiple vGPU per VM, vGPU live migration, and support for advanced GPU hardware cards (NVIDIA® Quadro RTX™ 6000 and 8000) makes Nutanix AHV a truly robust platform with seamless infrastructure lifecycle management, to run some of the most advanced VDI and GPU-accelerated workloads with minimal disruptions and maximum agility. It blurs the lines between graphics and compute further by breaking down infrastructure silos. Use our portal to download the latest GPU driver that is best suited for your business so that you can run GPU-accelerated applications on Nutanix AHV. Reap the benefits of high performance, improved density, business agility, and CapEx and OpEx cost savings with Nutanix AHV and NVIDIA vGPU technology today! Learn more in this Solution Brief and  from this NVIDIA Blog to get started. 

© 2021 Nutanix, Inc. All rights reserved. Nutanix, the Nutanix logo and all Nutanix product, feature and service names mentioned herein are registered trademarks or trademarks of Nutanix, Inc. in the United States and other countries. Other brand names mentioned herein are for identification purposes only and may be the trademarks of their respective holder(s). This post may contain links to external websites that are not part of Nutanix.com. Nutanix does not control these sites and disclaims all responsibility for the content or accuracy of any external site. Our decision to link to an external site should not be considered an endorsement of any content on such a site. This post may contain express and implied forward-looking statements, which are not historical facts and are instead based on our current expectations, estimates and beliefs. The accuracy of such statements involves risks and uncertainties and depends upon future events, including those that may be beyond our control, and actual results may differ materially and adversely from those anticipated or implied by such statements. Any forward-looking statements included herein speak only as of the date hereof and, except as required by law, we assume no obligation to update or otherwise revise any of such forward-looking statements to reflect subsequent events or circumstances.