Graphics Processing Units (GPUs) are being used across the media and entertainment industry to accelerate complex and time consuming tasks in a range of applications from on-air graphics to real-time stereoscopic image generation. Whether they’re being used for the power they offer in graphics processing or in parallel processing, GPUs deliver the highest floating point performance of any processor you can put in an off the shelf workstation. This allows many previously slow and computationally complex problems to be solved in real-time. Historically, however, getting video into and out of the GPU from 3rd party hardware introduced unnecessary delay because there was no direct synchronization or communication between devices.
NVIDIA® GPUDirect for Video technology allows 3rd party hardware to communicate directly with NVIDIA GPUs. By taking advantage of this new API the historical issues of introducing too much delay or latency are gone. With NVIDIA GPUDirect for Video, devices are fully synchronized and the CPU doesn’t waste cycles copying data between device drivers. This flexible API gives full control to 3rd party hardware to stream video to and from the GPU at Sub-Frame transfer times. By allowing industry leading 3rd party hardware design companies to access the power of NVIDIA Quadro GPUs directly and openly for graphics or parallel processing, the limits of what’s possible using workstations in real-time video production is being redefined.
The diagrams below show the difference between a system with and without GPUDirect for Video technology.
GPUDirect for Video is available to application developers through the SDK's of the following I/O board vendors:
GPUDirect for Video is supported on Windows 7, 8 and Linux with the following GPUs:
For more information: