NVIDIA vGPU software enables every virtual desktop infrastructure (VDI) user to harness the power of an NVIDIA GPU.
NVIDIA VIRTUAL GPUs MULTIPLIED
Supercharge Virtual GPU Performance
Multiple NVIDIA virtual GPUs (vGPUs) can now be deployed in a single virtual machine (VM) to scale application performance and dramatically speed up production workflows.
NVIDIA vGPU software enables every virtual desktop infrastructure (VDI) user to harness the power of an NVIDIA GPU.
For lighter workloads, multiple virtual machines (VMs) can share GPU resources with NVIDIA virtual GPU software.
For larger models and increasingly demanding workflows, NVIDIA Quadro® Virtual Desktop Workstation (vDWS) and NVIDIA Virtual Compute Server (vCS) software allow multiple GPUs to be assigned to a single VM.
NVIDIA® NVLink™ enables a high-speed, direct GPU-to-GPU interconnect that provides higher bandwidth for multi-GPU system configurations than traditional PCIe-based solutions.
A deep learning training workload running TensorFlow ResNet-50 with mixed precision can run up to 50 times faster with multiple NVIDIA V100 GPUs and vCS software than a server with only CPUs. Additionally, running this workload in a hypervisor-based virtual environment using vCS performs almost as well as running the same workload in a bare-metal environment.
Server Config: 2x Intel Xeon Gold (6140 3.2GHz) [VMware ESXI 6.7 U3, vCS 9.1 RC, NVIDIA V100 (32C profile), Driver 430.18] TensorFlow Resnet-50 V1, NGC 19.01, FP16 BS: 256
Engineering simulations run almost 7 times faster, more smoothly and securely when powered by multiple NVIDIA GPUs with Quadro vDWS, versus CPUs only. In some cases, they can be run for much lower cost than a vCPU only solution. Multiple vGPUs enable exponentially faster processing with higher fidelity models in a VDI environment.
Tests were run on a server with 2X Intel Xeon Skylake (6148 2.4 GHz Turbo - 3.6 GHz), NVIDIA Quadro vDWS software, Tesla V100 GPUs with 32Q profile, Driver - 410.53, 256 GB RAM, Cent OS 7.4 64-bit. Benchmark Model: Water Jacket Model, Unsteady RANS, Internal Flow, Fluid- Water, size 4, 20 time steps
Tests were run on a server with 2X Intel Xeon Skylake CPUs (Xeon 6148 2.4 GHz 32-core), NVIDIA Quadro vDWS software, Tesla V100 GPUs with 32Q profile, Driver - 410.53, 256 GB vRAM, Cent OS 7.4 64-bit. Benchmark Model: ~450-550 TFLOPs, 5.9M DOF, Highly Nonlinear Static, Axisymmetric model with non-axisymmetric loading and twist, Direct Spare Solver (Model courtesy: SIMULIA)
Render photorealistic scenes up to 4X faster, from just about anywhere, enabling designers to run multiple iterations in less time.
Tests were run on a server with 2X Intel Xeon Gold (6154 3.0 GHz) CPUs, 512 GB RAM, RHEL 7.5, NVIDIA Quadro vDWS software, Tesla V100-32Q, Driver - 410.39, 256 GB RAM, Windows 10 x64 RS3
Tests were run on a server with 2X Intel Xeon Gold (6154 3.0 GHz) CPUs, 512 GB RAM, RHEL 7.5, NVIDIA Quadro vDWS software, Tesla V100-32Q, Driver - 410.39, 256 GB RAM, Windows 10 x64 RS3
View product release notes and requirements.
Simulation and training