NVIDIA^® Virtual GPU Solutions

Supermicro has developed a line of fully configured GPU server solutions to address the rapidly growing Virtual Desktop Infrastructure (VDI) market. These solutions include a Supermicro 1U, 2U, FatTwin™, or 4U Tower SuperServer® configured with 2x Intel® Xeon® E5-2600 v4/v3 CPUs, up to 3TB ECC DDR4-2400MHz memory in 24 DIMM slots per node, and NVIDIA® Tesla® GPUs based on the Volta, Pascal, and Maxwell architectures which deliver best-in-class virtual desktop experience to users anytime and anywhere.
VDI has become increasingly popular with the broadened availability of mobile broadband, higher bandwidth, more compute power in modern mobile devices, higher display resolutions and a trend of BYOD within enterprises. NVIDIA virtual GPU software runs on NVIDIA® Tesla® GPUs and is based on NVIDIA Volta, NVIDIA Pascal™, and NVIDIA Maxwell™ GPU architectures.

Key Benefits

Best Windows 10 VDI user Experience.
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 user experience from day one.
Multiple Monitor. Support multiple HD displays and up to two 4K-resolution displays to expand your workspace, enhance multi-tasking, and increase productivity.
Native-PC Experience. Utilize NVIDIA GRID software with Tesla GPUs to deliver responsive, virtual workspaces that perform like native PCs to every user.
Linux OS Applications. Deliver the benefits of like-native performance of VDI environments to software developers and EDA designers that rely on a Linux-based development environment.

NVIDIA Virtual GPU Software Editions

NVIDIA virtual GPU has three different products designed to let you assign the right level of resources to each of your users based on their exact needs.

Feature	Grid vApps	Grid vPC	Quadro vDWS
Use Case	Use with Citrix XenApp or other RDSH solutions like VMware Horizon Apps.	For virtual desktops delivering standard PC applications, browser, and high definition video.	For professional graphics applications mainstream and high-end designers.
Desktop Virtualization	-	Yes	Yes
RDSH App Hosting	Yes	Yes	Yes
RDSH Desktop Hosting	Yes	Yes	Yes
Windows Guest OS	Yes	Yes	Yes
Linux Guest OS	-	Yes	Yes
Maximum Displays	1	Four HD, Two 4K	4
Maximum Resolution	1280x1024	4096x2160 (4K)	4096x2160 (4K)
CUDA & OpenCL Supported	No	-	Yes
GPU Pass-Through & Bare Metal Supported	Yes	-	Yes

Licensing Models

Annual License		Perpetual License
An annual subscription is active for a fixed period as defined by the terms of the subscription license. To be kept active, the license will need to be renewed at the end of the subscription period. Annual subscription includes software license and SUMS.		A perpetual license allows for use of the licensed software indefinitely. Users that opt to license using this model are required to subscribe to SUMS for the first year. The SUMS subscription can renewed on a yearly basis after the expiring of the initial subscription.

NVIDIA Virtual GPU Hardware

NVIDIA virtual GPU (vGPU) software runs on NVIDIA® Tesla® GPUs and is based on NVIDIA Volta, NVIDIA Pascal™, and NVIDIA Maxwell™ GPU architectures. Match your needs with the right GPU below.

GPU		Recommended NVIDIA GPUs for Different Use Cases
TESLA V100 / TESLA P100		High-end professional graphics users; includes use for double-precision compute workloads (e.g., running 3D models and design workflows, intensive computer-aided engineering [CAE] simulations)
TESLA P40		Mid-level to high-end professional graphics users; includes use for single-precision compute workloads (e.g., rendering and creating complex designs)
TESLA P4		Entry to mid-level professional graphics users, including deep learning inference workloads and Pascal features
TESLA M60		Entry to mid-level professional graphics users, including support for heterogeneous profiles.
TESLA M10		Best Density and Performance for the Knowledge Worker