NVIDIA has announced the VGX platform, which utilizes the GPU to allow enterprises to efficiently deliver virtualized workstation performance and capabilities–including rich multimedia and 3D graphics–to users on smartphones, tablets or PCs. This will be especially important as businesses move away from traditional servers and start moving towards the cloud as a resource to deposit, retrieve, use and secure files and programs.
The VGX HyperVisor also enables the use of User-Selectable Machines (USMs)—allowing system administrators to configure and provision the physical VGX hardware to take on the characteristics needed for specific types of users. This gives IT departments the freedom to support multiple pools of user types from a single server, whether they are using traditional workstation PCs or mobile devices.
The two Kepler-based boards are the VGX K1 and K2, which are ideal for data centers that need large amounts of processing power. The VGX K1 board features four entry-level Kepler GPUs and 16 GB of DDR3 RAM, and can support up to 100 users. The K1 is focused more towards those basic desktop users such as those in call centers, whose graphical demands are not very high.
The K2 has two high-end Kepler GPUs for a total 8 GB of GDDR5 RAM. The VGX K2 can support up to 64 users for light graphic use, but caters more to users who need more graphics intensive programs, such as design engineers and video editors. In that case, a user can take control of a single GPU for a maximum of two users who will each have 4 GB of GDDR5 RAM at their disposal. There is no word yet on when either will reach our shores, but we do know that the VGX K2 will be available in the first quarter of 2013.