Intel’s 14th Gen Core processor will take AI a step further

Within its future generation of 14th generation Core chips called “Meteor Lake”, Intel would integrate a logical element resulting from its acquisition of Movidius. A chip much better at AI tasks – inference, deep learning – than GPUs.

It is an understatement to say that the 14e generation of Intel “Core” processors could be a milestone. Engraved in Intel 4 (4 nm equivalent) and scheduled to arrive in the course of 2023, “Meteor Lake” by its nickname introduces a lot of things, including an AI part that could make a difference in many applications, in particular graphics and imagery.

Intel had already communicated on an AI “tile”. The “tiles” are the first major novelty of Meteor Lake: like some of its competitors like AMD, Intel is going to play Lego with its 14e generation. A batch that brings with it new cores both high performance (Redwood Cove) and low energy consumption (Crestmont) and even a new integrated GPU. But it is on the side of AI that we know more today.


Simply stamped “Integrated AI Acceleration” on official Intel slides, very specific mentions within Linux drivers refer to a “Versatile Processing Unit” or VPU. A chip – or rather a piece of a chip – which is defined as “an inference accelerator integrated into the CPU to accelerate the Computer Vision (computer vision, ndr) as well as Deep Learning applications (deep learning, ndr)”. A “VPU” who would be a champion of inference, does that remind you of anything? Yes: the Myriad chips from Movidius that Intel bought six years ago and which were already VPUs but which then meant “Visual Processing Unit”. Chips champion of image analysis that the ESA has even sent into space to sort through its images directly in orbit and thus free up bandwidth for its satellites.

See also  Windows 10 Lifetime License is $14 and Windows 11 is $21
Intel’s Myriad 2 Vision Processing Unit brings onboard artificial intelligence to satellites on a system built by Ubotica Technologies and paired with a hyperspectral-thermal camera from cosine measurement systems. (Credit: Tim Herman/Intel Corporation)

Within a consumer processor, such an accelerator could greatly improve not only performance, but also and above all the performance/watt ratio of image-related operations: automated sorting, effects applications (Photoshop), recognition face in a photo database (Lightroom), etc. If Intel and its Core i9 still has its hands on very high performance, it is at the cost of high energy consumption. Such an accelerator would be welcome to limit consumption on these specific tasks.

Quite similar in idea to the Tensor Cores of Nvidia graphics cards – an inference accelerator – Intel’s VPU could bring to the 14e generation of Intel’s Core the boost to make the difference against competing chips. Because whether it’s Apple and its excellent multimedia compression/decompression engine, AMD and its latest generation of RDNA2 integrated GPUs, etc. all the semiconductor actors are accelerating on the elements around the CPU itself. Because if improvements are always possible, it is above all the finesse of the engraving that drives the performance upwards in this area. While adding a third-party accelerator (graphing, inference, etc.) promises much more gains on given tasks.

Before the arrival of Meteor Lake 14e gen, we will already have to see the arrival of the 13e generation of Core chips (Raptor Lake) stamped Raptor Lake. A less “innovative” generation of chips since it focuses on increasing cores and seeking performance, but which will have the advantage of being compatible with the 12e current generation. In all likelihood, Intel is expected to unveil its 13e back-to-school generation. And start distilling information about Meteor Lake in early 2023.

See also  This is what Adele and Rich Paul's new love nest looks like

Source :




Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Social Media

Most Popular

On Key

Related Posts