Google Pixel mobiles are the gold standard in Android, in many ways; They are the first to receive the latest news from the system and those who boast of advanced photography and artificial intelligence technologies.
[Pixel 7 Pro, análisis: todo gira en torno a la cámara]
Where pixels aren’t exactly pointers is in processor power. To put it mildly, none of the models in the range manage to compete with mobiles from other brands that cost the same price, let alone truly leading mobiles.
Why the Pixel isn’t as powerful
Since the Pixel 6, Google uses its own processor, the Tensor G, instead of taking it easy and opting for the market reference, the Snapdragon, or its rival, the Mediatek Dimensity. This means that the Pixels are highlighted when we use “benchmark” applications like Geekbench or AnTuTu compared to mobiles equipped with these processors. In 3D applications, the difference is particularly stark, although things have improved with the Pixel 7’s Tensor G2 and its new improved GPU.
Everyone knows that. What you may not have thought of is that this is not a mistake by Google, but a conscious decision; simply, power was not a priority for the company’s engineers when designing the chip.
That’s something Monika Gupta, Senior Director of the Google Silicon team behind Tensor G, confirmed in an interview on the official Made By Google podcast. When the conversation turned to benchmarks, Gupta was very clear that these types of apps served their purpose a long time ago, but that the industry has evolved since then. He thinks a benchmark score only tells part of the story, but not the whole story, and that’s why Google didn’t use them as a benchmark; instead, they focused on actual workloads.
He has all the sense in the world. Although many benchmark apps try to mimic real usage, the truth is that there is nothing quite like how you as a user use mobile; especially since each person uses the mobile in a different way.
That’s why Google feels “very comfortable” losing in all comparisons that have been posted on the internet, as it puts “the end-user experience first”. But perhaps more than that, the Tensor G isn’t such a powerful processor as it focuses on machine learning.
And not only in current technologies. Gupta says he doesn’t make decisions based on where machine learning is today, but what it will be like five years from now, so mobile will continue to be capable of new user experiences all this time. .