Intel and NVIDIA are still in negotiations over a new chip architecture called Intel-Nvidia Neural Network (NVN) that could see both companies merge in a few years.
The two chipmakers are currently in talks over the next few years over the design of Intel’s chip design and Intel’s future chip design, but that hasn’t stopped the two companies from pushing ahead on their respective AI chip designs.
Intel has said it’s aiming to start releasing a new AI chip in 2019 that will replace Intel’s existing Neural Network chip in the near future, and Nvidia has said that it plans to ship a new Neural Network-based chip that will compete with Intel’s current chips sometime in 2020.
The new chip will be able to perform tasks that were previously impossible to achieve with Intel-based chips.
NVN will also allow both companies to work on the same AI chip architecture, and both chipmakers could share some of the same software, as well as use the same underlying architecture.
Intel and NVN are still discussing the details of the new chip, but Nvidia has already announced that it’s going to use the Neural Network architecture to deliver its next generation of Pascal GPU chips.
Nvidia has been teasing a new GPU that will be built around the Neural Net architecture, so the company is likely to be using that architecture to make the Pascal GPUs that it wants to make for Pascal.
Nvidia’s new GPU will likely use a very similar architecture to the GeForce GTX 980 Ti, the card that will debut later this year.
The GTX 980 is a reference design that Nvidia hopes will be the first GPU in the world to support the Neural Core architecture.
The Neural Net-based GPU will probably use Nvidia’s own Pascal architecture, which will be similar to the architecture that Intel and AMD use for their current chips.
The NVIDIA GPU will be more powerful, but not necessarily better, and will likely be more efficient, but it will also likely be slower, because the NeuralNet-based GPUs will use different compute units, which could make them slower overall.
Nvidia is also planning to introduce a new CPU and GPU architecture for the Neural Networks.
The Pascal GPUs will likely not be able do the same tasks as the GTX 980, but they’ll likely be capable of performing some of those tasks.
Nvidia says that it hopes to ship its NeuralNet GPU within a few months, but AMD hasn’t revealed any specific plans for when that GPU might arrive.
The Nvidia Neural Network GPU is expected to be a powerful GPU that Nvidia will be releasing around the time of the Pascal GPU launch, and the GPU is likely going to be the basis for both of the upcoming Pascal GPUs.
Nvidia will probably release the Neural NET-based Pascal GPU sometime after 2019, but the NeuralNET-based GTX 980Ti GPU isn’t expected to arrive until 2020.
Nvidia doesn’t have much information about how the Neuralnet GPU will differ from the GTX 970 Ti, which is expected for sometime in 2021.
Nvidia may not release its NeuralNET GPU until sometime in 2019, because it plans on using the NeuralNets new architecture to improve the performance of its Pascal GPUs as well.