Remote FX for VMs

Installing the RemoteFX role on a server allows an application running in a virtual machine (VM) to use a graphics processing unit (GPU) in the server for rendering graphic images.

In a server running RemoteFX (left), the server must be running Windows Server 2008 R2 SP1 or Hyper-V Server 2008 R2 SP1, and the parent partition on the server will manage the server’s physical hardware, which will include a graphics processing unit (GPU). Virtual Service Providers (not shown) and a VMBus manage sharing between the device drivers running in the parent partition and the child partitions that use the devices.

In the VM running in the child partition, the Hyper-V integration services for Windows 7 SP1 install a virtual GPU (vGPU), which displays in the VM’s Device Manager as a RemoteFX GPU and appears to applications running in the VM as a DirectX 9-compatible GPU.

The remote client (right) connects to the VM over the network using RDP 7.1. When a graphical application starts in the VM, the application detects the vGPU and sends any graphic images it needs to render to the vGPU in the same manner that it would if it were sending data to a physical GPU. Instead of processing the data, the vGPU passes the data across the VMBus to a Render, Capture, and Compress (RCC) component running in the parent partition. One instance of the RCC component runs for each VM using RemoteFX. The RCC passes the graphical data to the physical GPU for rendering in the same manner as an application running locally on the server would.