Google’s Tensor Chip Is Slower Than Apple’s A12 Bionic, an SoC That Is Three Years Old Now

Before Google even officially announced its custom Tensor chip, there was a rumor doing the rounds that the company is not focused on making this SoC the fastest on the planet. Instead, the release would likely be more focused on efficiency than anything else, and according to this benchmark comparison, that rumor was true. Still, it is egregious to see a flagship chipset unable to compete with an A12 Bionic, a custom silicon Apple released three years ago.

Tensor Fails to Beat the A12 Bionic in Both Single-Core and Multi-Core Tests

The performance comparison provided by 9lekt on Twitter shows Geekbench 5 scores belonging to the Tensor and the A12 Bionic. Sadly, Apple’s chipset from three years ago beats the Tensor in both single-core and multi-core tests, and naturally, these figures would have potential customers confused. The first thing they will likely ask themselves is why Google refrained from using Qualcomm as its supplier for the Snapdragon 888.

In this way, the Pixel 6 and Pixel 6 Pro would at least remain competitive with the rest of the Android flagship smartphone camp. While we agree that these numbers are beyond disappointing, keep in mind that benchmark results only tell half the story, and what you are seeing does not necessarily translate into real-world performance. We also have to take into consideration the optimization bits.

Google Pixel 6 Pro (2021) vs iPhone XS Max (2018) Geekbench result
No comment pic.twitter.com/a11E4EqAhQ

— S Pain is busy irl (@9lekt) October 24, 2021

Qualcomm’s chipsets were no stranger to overheating, as evidenced when using the Pixel 5a, which does not even use a flagship SoC from the San Diego chip giant. With the Tensor, Google may have sacrificed the performance aspect of the Pixel 6 and Pixel 6 Pro, but with its added control on both hardware and software, it is possible that the operating system runs significantly smoother on the latest flagships than it does on competing handsets, with minimal dips and stutters.

The disappointing single-core and multi-core scores could also attribute to the energy-efficient side of the Tensor, and it is possible Google intentionally wanted this chip to underperform so it could provide the best battery life for Pixel 6 and Pixel 6 Pro owners. Oftentimes, you will sift through the hardware specifications of a smartphone and be thoroughly impressed with what is published on paper, only to be left bitterly disappointed and livid at the atrocious levels of software optimizations.

Well yes, Whitechapel is a 5nm chip with current performance on PVT units closer to SD870, they are not trying to match SD888. Google's focus is on ML & so the raw AI performance is matched to that of other leading mobile chips. Plus that Mali GPU is performing good under stress.

— Yogesh Brar (@heyitsyogesh) May 24, 2021

Google might have a different plan altogether with its Tensor, but we are not expected to see those results this early. When the first commercial Pixel 6 and Pixel 6 Pro reviews are out, we will update our readers accordingly, so stay tuned.

The post Google’s Tensor Chip Is Slower Than Apple’s A12 Bionic, an SoC That Is Three Years Old Now by Omar Sohail appeared first on Wccftech.