Google uses self-developed Tensor SoC on the Pixel 6 and Pixel 6 Pro released this year, replacing the Qualcomm solution that has been used in the past. At present, the evaluation of Pixel 6 and Pixel 6 Pro is not bad, especially the camera performance and battery life, which have received a lot of praise.
Tensor SoC uses a three-cluster design in the CPU part, including two Cortex-X1 super cores, equipped with Mali G78 GPU, at the same time, it is equipped with a TPU for AI computing and integrates a Titan M security chip, which is manufactured using Samsung’s 5nm process. Although it is Google’s first self-developed SoC, its performance in terms of camera and battery life has been beyond the expectations of many users. According to HotHardware reports, the first custom kernel for Pixel 6/Pro improves battery life.
After careful research, it is found that the core of Tensor SoC has many features in design, such as in-depth optimization for Android 12. In actual operation, when the screen is off, the frequency of the small core is limited to 1.4 GHz, the frequency of the large core is limited to 1.49 GHz, and the frequency of the super large core is limited to 1.58 GHz to reduce power consumption. This means that when playing music or downloading updates, compared to ordinary Android phones, the power consumption may be lower, without affecting the user experience, in exchange for longer battery life.
Previously, a developer found a device code-named “Whitefin” in the system program code, with parameters similar to those of Pixel 5. It is reported that Google may release a mid-range mobile phone Pixel 6a based on self-developed chips, using self-developed chips similar to Tensor SoC. However, it was also pointed out that this may only be a test model of Google’s early Tensor SoC on Pixel 5. Perhaps the original plan was to use self-developed chips in Pixel 5, but was forced to postpone it for various reasons.