Google Tensor Did Not Perform as Expected Against Apple A15 ML Capabilities

Google Tensor without a doubt is one of the most talked-about aspects of the new Pixel 6 series. It is the first time that a Pixel device is equipped with a Google in-house chipset. What’s more, the chip is said to feature exceptional machine learning capabilities as well as other computational photography and transcription abilities.

Obviously, it is pleasing for Pixel lovers that the Tensor chipset’s ML capabilities will make for other things. It is worth noting that Snapdragon 888 and Exynos 2100 phones did outperform the Pixel 6. In fact, Apple A15 and even A14 Bionic did beat the processor. Moreover, in the later tests, the Apple A15 Bionic chipset performed better even in the machine learning department, according to Notebookcheck. Geekbench ML performed the test specifically to measure machine learning performance.

Also Read: DDR5 Pricing Might Be 50 To 60 Percent More Than DDR4.

Specifically, the Google Tensor scored 307 points in the Geekbench ML TensorFlow Lite CPU test. In comparison, the A15 Bionic managed to score 939 points. It is important to note that a score of 307 points is nowhere near, but is nowhere near Apple’s machine learning capabilities.

The test further performed comparisons on other aspects too. For instance, the A15 scored 2727 points in the neural accelerator test while the Tensor scored 1720. In the GPU test, the Tensor got 1428 points while the A15 responded with a score of 2727.

The total aggregate score of the A15 is 5,934 whereas the Tensor gets 3,455 points. This clearly suggests that the A15 is around 71% faster than the Google Tensor in machine learning capabilities, at least according to benchmark tests. It is worth noting that the real-world story could turn out different.