Previously, NVIDIA relied on the use of special expensive chips in the server area to speed up the work of artificial intelligence systems. At Computex 2023 you specifiedwhich works with partners to ensure that the Tensor cores of its discrete graphics solutions are used to accelerate client-side AI.
A collaboration is underway with Microsoft to implement hardware-accelerated AI on laptops with discrete NVIDIA graphics using fourth-generation Tensor cores. The corresponding tools are already available to developers in the Windows 11 environment and the results of their work will be used on more than 100 million computers with graphics cards of the RTX generation and workstations configured accordingly. So far, both companies report that more than 400 games and applications have already been released with artificial intelligence acceleration.
According to NVIDIA representatives, notebooks and mobile workstations that support Max-Q technology will master the efficient acceleration of artificial intelligence systems with the help of tensor cores in terms of power consumption. Once the need for performance scaling arises, the number of hardware resources involved increases proportionately. Notebook manufacturers, represented by Dell, HP Inc., Lenovo and ASUS, have already expressed their willingness to adopt such technologies. The necessary software tools are also already available to application developers in the special section of the NVIDIA website.