DCV – High-End Remote Desktop for VFX

On Windows NICE DCV offers high performant remote 3D desktop access for:

• NICE DCV on physical Windows workstation or server with or without GPU
• NICE DCV in virtualized Windows VM without GPU or with vGPU or physical GPU attached

vGPU or virtual GPU are subsets of a physical nVidia Grid GPU or AMD GPU where e.g. 1GB, 2GB, … frame buffer can be allocated for the respective vGPU. The GPU rendering performance is typically distributed to the different vGPUs on a first come first serve basis.

The vGPU is made available to the Windows OS in the Virtual Machine and NICE DCV via GPU-passthrough enabled by the hypervisor which could be e.g. Redhat-KVM, VMWare or XenServer. Our analysis of price performance of different nVidia GPUs provides additional background.

Installation of NICE DCV for Windows is done in 2 minutes by downloading and executing the installer – just try it yourself.

More information about the features of DCV for Linux: NICE DCV Product Codes and Features.

NICE DCV for Linux works in console mode for high-end remote desktop streaming of the console with or without GPU or many concurrent virtual desktops for different users without GPU.

The very efficient NICE DCV GPU sharing use case supports GPU sharing for e.g. 10-15 full-performant 3D Linux virtual desktops on physical Linux servers for different users sharing one or multiple GPUs, CPU cores and memory. This is a very cost- and resource efficient and straight-forward solution on physical Linux widely used offering full performance without the need for virtual machines or Grid licenses for every VM.

NICE DCV can launch different sessions for each 3D desktop sharing the GPU. Todays GPUs computation power is sufficient to support many concurrent users. GPU memory of 8GB to 16GB or even up to 48 GB supports many concurrent 3D Linux applications as often the consumption of GPU memory by the 3D applications is less than expected also for large models.

The GPU consumption of each 3D application can be analyzed with nvidia-smi which helps to understand overall GPU memory consumption (please find more background on nvidia-smi and price performance of different GPU types in our article Overview of selected nVidia GPU cards). Using virtual machines wastes resources compared to shared servers – we explain why in The hidden cost of VMs: Why They Waste Resources and How To Avoid It.

More information about the features of DCV for Linux: NICE DCV Product Codes and Features.