Naming smartphone chips Tensor is Google’s way of asserting its own style

Google has launched Pixel 6 and one of the impressive points is that it uses Google’s own SoC called “Tensor”. The Pixel 6 could flop or become one of the best Android phones out there. We won’t know about that until the device officially goes on sale, but the Tensor chip will probably have a bigger impact than the Pixel 6, because Google has high hopes on it. If it does all the things a smartphone requires, Tensor will be a huge success that very few companies can match.

But let’s skip the technical side and talk about… the name Tensor because it’s the most “Google” name for phone hardware ever. To someone unfamiliar with Google and all it does, calling a chip “Tensor” may seem rather silly.

Tensor can be a form of muscle that stretches to cause another body part to stretch. We all have them in places like in our ears and mouth. Interestingly, tensor is also a company that manufactures skateboard parts. In mathematics, a tensor is an algebraic object that describes a polylinear relationship between sets of algebraic objects relative to a vector space. It doesn’t take a guess to know what Google uses the name “Tensor” by definition.

That’s because these particularly complex math operations can also be used for artificial intelligence, and Google loves all things AI. Google even named the first chip they invented Tensor Processing Unit. That’s right, the Pixel 6 isn’t the first device with a chip called Tensor.

Most companies do AI computing and machine learning through what are known as neural networks. Neural networks inside our bodies basically exist and lie in between the “input” and the “output” of the action. For example, if you want your hand to move and grasp something, it will move and grasp that thing.

Computer neural networks do the same thing: they take one or more inputs, drive them through a set of in-between and hidden processes (meaning only input computations and output paths). can access them), then output the results.

Google Cloud AI

To perform the entire neural network operation efficiently, processors have several AI accelerators if you want predictable and precise results. NVIDIA, Intel and AMD have spent a lot of time and money making GPUs (Graphics Processing Units) that act as perfect AI accelerators and we can see that with the leap in graphics games over the past few years.

The TPU has one function: to act as an AI accelerator. You’ll find a bunch of them inside Google’s cloud servers and inside the Pixel 6 as well.

It’s important that we all understand that for Google, the word Tensor means a piece of hardware used for AI and machine learning, and all other definitions don’t matter. That’s why the company named its smartphone chip Tensor. If you were Google and you made a smartphone chip designed primarily for AI, you would obviously name it Tensor.

Reference: AndroidCentral