Google has officially entered the silicon game. The Google Pixel 6 and Pixel 6 Pro, released by the tail end of 2021, forgo the Qualcomm chipset in the past and instead go for a new chip, Google Tensor.

Being Google’s first entry into the SoC ecosystem, it certainly has its rough edges. However, it’s an excellent first step into a battlefield with rough competition. However, by launching late last year, it needs to put up a much more constrained battle with the two main chips—the Snapdragon 8 Gen 1 and the Samsung Exynos 2200—launched in early 2022.

How do all three chips compare? And should you care about the difference?

specific rundown

Google Tensor was notable because of its curious core configuration. The chipset comes with two ARM Cortex-X1 cores, which are clocked at 2.85 GHz. These are accompanied by a central cluster of two Cortex-A76 cores running at 2.4 GHz and a power-efficient cluster of four Cortex-A55 cores clocked at 1.8 GHz.

In contrast, the Snapdragon 888—Qualcomm’s main offering in 2021, and a close comparison of this chip—differs in that it packed a single Cortex-X1 core, but actually three Cortex-X1 cores in the middle cluster. A78 Corps included. 2.4 GHz comparison.

The difference is quite a toss-up. On the one hand, the Cortex-A76 cores are older and weaker than the Cortex-A78 cores in the 888. But the Tensor comes with two Cortex-X1 cores instead of one, to offset that somewhat performance penalty and with a few machines. Considering learning applications—more on that later.

But in terms of raw specifications, the chip pales in comparison to the 2022 chips. The Snapdragon 8 Gen 1 has one Cortex-X2 core running at 3 GHz, three Cortex-A710 cores running at 2.5 GHz and four Cortex-A510 cores running at 1.8 GHz. The Exynos 2200 has a largely identical core configuration, and it’s clear from just looking at the specifications that both chips will give Google a run for its money.

But there’s more to it than meets the eye.

Google knows its tensor chip isn’t the strongest around, and it never meant to. The star of the show is the TPU which is included in the chip. Google is all about that machine-learning and artificial intelligence shenanigans, and that co-processor is capable of speeding up things like real-time language translation, text-to-speech and image processing, and more importantly, Do-it-all tools.

This TPU also enables the application of Google’s HDRNet algorithm to video, which Monica Gupta, senior director of Google’s Silicon team, told Ars Technica—”should bring the Pixel’s signature look to video.”

We also have Titan M2 security chip. It is a secure enclave that provides a much needed extra level of security by protecting your sensitive information, such as biometric cryptography, and key processes such as Secure Boot.

It’s a different kind of chip for a user who doesn’t really care much about performance. Still, it should stay grounded in the performance department and not disappoint — and at this stage, it still looks very promising. However, we’ll need to look at benchmarks to set the record straight here.

benchmark rundown

Now, we’re going to take a look at some of the benchmark scores of all three chips. Benchmarks are hardly scientific evidence and do not necessarily reflect real-world performance, but they are still a simple way to determine whether one chip performs better than another. The difference is actually much less than you might think based on the spec sheet.

We tested three different devices—a Google Pixel 6 Pro and a Snapdragon and an Exynos version of the Samsung Galaxy S22 Ultra—on Geekbench. And there are some important conclusions from these results.

Starting with a single-core score. It’s consistently low on the Exynos 2200, but Google Tensor really goes toe (mostly) with the Snapdragon 8 Gen 1. We see some difference in multi-core, but the difference is very small; The Snapdragon 8 Gen 1 scores just 28% higher than Google Tensor and barely higher than the Exynos 2200.

These numbers indicate a handful of things. Yes, both Exynos 2200 and Snapdragon 8 Gen 1 are more powerful than Google Tensor, and if one needs to emerge as the winner, it is probably Snapdragon. They are the most capable chipsets.

We should also note here that regardless of benchmarks, the real-world differences between the two chipsets can vary wildly – as found by XDA-Developers, you have the Snapdragon version of the S22 Ultra compared to the Exynos version. Could be a better experience.

Leave a Reply

Your email address will not be published. Required fields are marked *