Hot on the heels of AMD announcing what it claims is the world’s most powerful workstation GPU, Nvidia has announced the “world’s fastest, most efficient” GPU in the form of the Nvidia Kepler Maximus.
Maximus, Nvidia says, will hand workstation users the opportunity to perform complex analysis and virtualisation on a single machine. Especially because it is supported by Kepler, the company promises the architecture will deliver “unparalleled” performance, and efficiency, in fields like manufacturing, visual effects, and oil exploration.
Nvidia has boasted that the Quadro K5000 GPU sports bindless textures, so users can reference 1 million and more textures in memory while reducing the CPU overhead, as well as FXAA/TXAA anti-aliasing for image quality. Then there is an increased frame buffer capacity of 4GB, and a next gen PCIe-3 bus interconnect which improves on data movement. Nvidia’s included a new display engine, which can drive four displays with one K5000, the company claims, as well as display port 1.2 support for resolutions up to 3840×2160 at 60GHz.
The Tesla K20 GPU, meanwhile, sports SMX streaming multiprocessor technology, which Nvidia says can bring about a 3x performance per watt advantage, and dynamic parallelism and Hyper-Q technologies with simplified parallel programming and faster performance in mind.
Nvidia, in a press statement, wheeled out spokespeople from a variety of industries including oil discovery. They all agreed: Maximus is rather fast. Powerful? They didn’t say. We can only speculate on what’s next for the title of world’s somethingorother, and from who – perhaps, world’s cheapest?