Future-Proofing IT Infrastructure With Virtual GPUs

PNY has submitted this article.

Virtual GPUs have the potential to improve user and IT experience. (Source: PNY.)  

Industries such as manufacturing, architecture, engineering, construction and operations, along with media and entertainment face compressed design cycles and pressure to reduce costs to maintain competitiveness. Organizations typically have multiple field offices with geographically dispersed development teams. In addition, internal departments frequently collaborate with external vendors on all aspects of projects.

The size of 3D models required for work, network bandwidth, and workstation performance means loading times are excessive. This often results in lost productivity. Hardware required to run high-end applications makes mobility complex and difficult, and this is exacerbated by the requirements of remote work and spatially disbursed teams.

Data centers are undergoing a renaissance using GPU-accelerated solutions. Organizations use GPU-accelerated applications to help drive better business decisions, boost productivity, and design better products or entertainment offerings that get to market faster.

GPU virtualization is becoming an imperative due to changing work patterns, and virtualization software such as NVIDIA virtual GPU (vGPU) makes it possible for users to work from anywhere. IT groups that implement these solutions can easily handle resource utilization, live migrations, management, and security.

The Benefits of Virtual GPU Infrastructure

The move to remote work from anywhere is now part of the work environment driving the need for virtual GPUs. According to Gartner, 48 percent of employees will likely work remotely at least part time after the COVID-19 pandemic, compared to 30 percent pre-pandemic.

Over 90 percent of current desktop productivity applications utilize some degree of GPU acceleration, making GPU-accelerated remote desktops essential for maintaining productivity and performance. Implementing a virtual GPU solution can future-proof the IT environment and support remote work.

NVIDIA A40 GPU. (Source: PNY.)

Implementing virtual GPUs in a data center provides many IT benefits, such as:

  • Near bare metal performance: Deliver an experience virtually indistinguishable from a bare metal (local GPU installed) environment.
  • Management and monitoring: Leverage common data center management tools such as live migration.
  • Continuous uptime and proactive management: Provide management abilities without negatively impacting users or losing data.
  • Optimal resource utilization: Provision GPU resources with fractional or multi-GPU virtual machine (VM) instances for maximum efficiency and full use of all available data center GPU resources.
  • Resource optimization: Provide onboarding and offboarding of team members (internal or external) as a project progresses.
  • Business continuity: Respond to changing business requirements and the needs of remote teams.
  • Enhanced productivity with real-time performance: Deliver superior performance on virtual desktops from the data center with the same responsive experience users expect from a physical workstation.
  • Collaborate anywhere, on any device: Free users from physical workstations allowing use of thin clients—or the device of their choice—to access the applications and data they need, regardless of their location.
  • Protect intellectual property: Protect IP by centralizing data and moving mission-critical files into the data center.
  • Product Lifecycle Management (PLM): Centralize PLM solutions in the data center for greater consistency and consolidation of data, as well as control over design changes.

Virtual GPUs for Every Workload

Not all organizations have the same graphics requirements, and there are a variety of GPU virtualization solutions. For example, NVIDIA’s virtual GPU portfolio includes data center GPUs (and select NVIDIA RTX boards) with software for every workload. Its virtual GPU software comes in three editions:

  • NVIDIA RTX Virtual Workstation (vWS) leverages data center GPUs to deliver a graphics-accelerated data center, enabling IT to virtualize any application—such as ESRI ArcGIS Pro, Siemens NX, Dassault Systèmes SOLIDWORKS, and Autodesk AutoCAD—to virtual workspaces with a workstation-class user experience. IT can focus on managing the user instead of managing PCs and workstations while enabling more secure work-from-anywhere workstyles with reduced threat of data loss or leakage.
  • NVIDIA Virtual Applications (vApps) provides application streaming with Remote Desktop Session Host (RDSH) solutions. The software allows teams to use their apps virtually without losing performance or productivity. With GPU sharing, multiple VMs serving as virtual desktop access (VDAs), can be powered by a single data center GPU to maximize utilization and affordability. With support for all major hypervisor virtualization platforms, vApps enables data center administrators to use the same management tools for their GPU-accelerated servers as they do for the rest of their data center. (See the NVIDIA vApps sizing guide here.)
  • NVIDIA Virtual PC (vPC) is for knowledge workers who use office productivity applications and multimedia, essential in today’s rapidly changing digital workforce. Some companies struggle to offer a remote experience that matches what professionals get from PCs onsite, but virtualized GPUs increase data center user density and accelerate productivity apps by letting workers seamlessly access the tools they need from anywhere. (See the NVIDIA vPC sizing guide here.)

How Honda is Using Virtual GPUs

With remote workers and the need to work with external organizations, along with optimizing IT operations, many organizations are increasing the use of vGPUs and virtual desktop infrastructure (VDI) solutions to help IT provide virtual solutions for remote workers, as well as manage and protect their data.

Automaker Honda has deployed engineering VDI to enhance productivity and operational efficiency for its R&D production centers. Virtual GPUs empower Honda’s teams to use computer aided design and engineering applications on any device—even low-cost laptop computers. Additionally, Honda IT can allocate the right level of performance for power users and knowledge workers alike. Across all Honda group companies, more than 4,000 VDI systems are experiencing better application performance and user experiences, as well as faster access to data and enhanced security of IP.

Virtual GPUs enable powerful graphics performance for workloads ranging from graphics-rich virtual workstations to office applications. This enables IT to leverage the management and security benefits of virtualization as well as the performance of GPUs required for modern workloads.

For more information on choosing the right NVIDIA vGPU solution, visit PNY.com/vGPU.