Google is Working on the Second Generation of Tensor


 Shortly after the Pixel 6 and 6 Pro were released, rumors emerged of a successor to the Tensor chip used in both phones.

This rumor comes from the Android source code on the Pixel 6, which mentions the second generation Tensor of the 'Cloudripper' code.


It is believed that this Cloudripper is not the name of the Tensor successor chip or the name of the Pixel 7. Most likely this Cloudripper is the name of the platform used to test the Tensor successor chip, which has the model number GS201, while Tensor is the model number GS101.



There is no more detailed information about this GS201, considering that Google has just released the Pixel 6, and the Pixel 7 which will use the new GS201 chip will be released next year.


Previously reported, Google does not use Qualcomm's system on a chip (SoC) in the Pixel 6 and 6 Pro, and uses a Tensor chip that they designed themselves.


After the Pixel 6 and 6 Pro were released, the specs were finally revealed as well. The tensor is made using a 5nm process and has eight cores, which are divided into large, medium, and small.


There are two large (or fast) cores, namely Cortex-X1 2.8GHz, more than Snapdragon 888 which only has one Cortex-X1. Then underneath there are two Cortex-A76 2.25GHz, and four Cortex-A55 1.8GHz.


Tensor was co-designed with Google Research, which means Google can embed the AI/ML platform into the SoC. According to Google, the presence of ML makes Tensor have new capabilities, such as Motion Mode, Face Unblur, Speech enhancement mode, and the application of HDRnet to video (up to 4K60).


This SoC is also claimed to have the most accurate word recognition system, and can be used in many applications. For example Recorder and Live Caption and without consuming too much battery power.


The Live Translate feature is also mastered by Tensor, which makes messages on WhatsApp for example, can be translated directly. Live Translate can also translate words from videos.

Previous Post Next Post

Contact Form