Intel processors could borrow key feature for Apple and iPhone
Intel could exploit one of the best features of Mac and iPhone processors in its future processors, according to a new rumor. Youtube channel Moore’s Law is Dead (MLID) uploaded a video focused on a vision processing unit (VPU) in upcoming Intel processors, which could handle machine learning tasks such as scaling video, recovering overexposed photos and improved speech synthesis.
If this is the first time you’ve heard of VPU, you’re not alone. They come from a company called Movidius, acquired by Intel in 2016. It is similar to the Neural Engine that Apple uses on its M1 desktop and tablet chips, as well as the one featured in the iPhone chips from the iPhone 8 to the recently released iPhone 13. Pro.
It’s basically an AI accelerator, similar to the Tensor cores that appear on Nvidia RTX graphics cards. Rumor has it that Intel is considering incorporating a VPU into its Meteor Lake processors, which are slated to launch in 2023. Intel has confirmed that Meteor Lake will feature a modular design, which the company has built upon.
Using its Foveros packaging technology, Intel plans to combine multiple manufacturing processes on a single chip. Basically they will look a lot like the Apple M1. In the renderings shown by Intel, the processors will feature a compute matrix where the processor resides, a system on a chip (SoC) that includes I / O, and a GPU matrix. The VPU will likely live in the compute matrix as Intel moves to a smaller manufacturing process with Meteor Lake.
It is no evasion to say that the applications of a VPU are almost endless. Currently, Apple uses its Neural Engine in applications such as Translate for real-time text-to-speech translation and to retrieve details from HDR photos. Third-party apps like Seek also use it, making it possible to identify plants and animals from a photo in seconds.
For the Intel chip, one of the MLID sources provided an example where you can highlight text and have it read. Dedicated AI accelerators also have applications in games, as proven by Nvidia’s Deep Learning Super Sampling (DLSS). Intel is also using something similar for its upcoming Arc Alchemist graphics cards.
Despite the marketing buzz around AI and machine learning, dedicated accelerators have dozens of applications in the data center and in consumer devices. Today’s big tech companies pose tough problems for machine learning, and the results speak for themselves. The scaling of AI on the Nvidia Shield is proof of that.
Meteor Lake is still a long way off, and it’s important to reiterate that VPU is just a rumor at the moment. Intel has yet to take the first step to prove that it can design chips with a hybrid architecture, which it is supposed to do with upcoming Alder Lake chips. For Meteor Lake, we will have to wait and see. A dedicated AI accelerator is plausible, if not likely.