Apple has been using Google-built chips instead of those made by industry leader Nvidia to create two key components of its AI software infrastructure for its upcoming suite of AI tools and features, a research paper issued by the company has revealed.
Apple's decision to rely on Google Cloud is significant as Nvidia produces the most sought-after AI processors and controls roughly 80 per cent of the market.
In the research paper, Apple did not mention explicitly it did not use Nvidia chips, however, there was no mention of Nvidia in the description of the hardware and software infrastructure of its AI tools and features.
Apple used two variants of Google's tensor processing unit (TPU) to train its AI models. It used 2,048 of the TPUv5p chips to build the AI model that will operate on iPhones and other devices.
For its server AI model, it deployed 8,192 TPUv4 processors.
Nvidia is known for making graphics processing units (GPUs) mostly used for AI behind-the-scenes crunching.
Google sells TPUs through its Google Cloud platform, unlike Nvidia, which sells its product as standalone item. To buy access from Google, customers have to build software through Google's cloud platform to use the chips.
The company's engineers said in the research paper that it would be possible to make even larger, more sophisticated models with Google’s chips than the two models discussed in the paper.
Image Source: Unsplash
Comentarios