Tensor SoC in Google Pixel 6 approaches Snapdragon 888 and Exynos 2100 performance level in Geekbench 5 run
Google Tensor can hold up well against the Qualcomm Snapdragon 888 and Samsung Exynos 2100 in single-core performance.

An apparent Geekbench 5 run for the Google Pixel 6 smartphone has been discovered, revealing the performance of the custom Tensor chip. Multi-core performance was distant from rivals but the Tensor’s single-core performance has it rubbing shoulders with smartphone chip luminaries such as the Snapdragon 888 and Exynos 2100.
Daniel R Deakin, 10/13/2021 �� �� ...
Google Pixel ARM Charts Smartphone Android
A Geekbench 5 record for the Google Pixel 6 with Tensor chip has been shared by MySmartPrice. The listing seems to have disappeared from the Geekbench database now, so have a pinch of salt to hand. The scores produced are actually very similar to those managed by a tested Google Pixel 6 Pro we reported about last month. The Pixel 6 in this particular listing had 8 GB RAM, features a Mali-G78 GPU, and it has a Tensor SoC with a 4+2+2 cluster arrangement consisting of 4x cores @1.80 GHz, 2x cores @2.25 GHz, and 2x cores @2.80 GHz.

The multi-core score came out as 2,706 points, which would not put the Google Tensor chip at the head of either Geekbench’s chart or in our comparison with the Snapdragon 888 and Exynos 2100. However, the single-core score was once again impressive, with 1,027 points being sufficient enough to head the Geekbench chart (although that is based on the averages of many tests) and would position the Tensor SoC strongly against the Snapdragon 888 and Exynos 2100 in our numerous tests (see table below; based on median results).

This appears to be a recent Geekbench listing for the Tensor SoC, so future Google Pixel 6 owners may have to put up with considerably weaker multi-core performance when compared to Snapdragon 888 and Exynos 2100 smartphones (-25% behind the Qualcomm chip). But single-core performance is only -9% behind the SD888 and -7% from the Samsung processor, so for Google’s initial attempt at a custom smartphone system-on-a-chip, the Pixel 6’s Tensor unit is actually looking like a solid foundation to build upon.



Geekbench 5.3
64 Bit Multi-Core Score
Median Qualcomm Snapdragon 888 5G (3379 - 3810, n=21)
3629 Points ∼101%
Median Samsung Exynos 2100 5G (3228 - 3583, n=3)
3364 Points ∼93%
Tensor (Google Pixel 6)
2706 Points ∼75%
64 Bit Single-Core Score
Median Qualcomm Snapdragon 888 5G (989 - 1156, n=21)
1129 Points ∼101%
Median Samsung Exynos 2100 5G (1003 - 1106, n=3)
1100 Points ∼98%
Tensor (Google Pixel 6)
1027 Points ∼92%

https://www.notebookcheck.net/Tensor-SoC-in-Google-Pixel-6-approaches-Snapdragon-888-and-Exynos-2100-performance-level-in-Geekbench-5-run.572196.0.html

Google Tensor can hold up well against the Qualcomm Snapdragon 888 and Samsung Exynos 2100 in single-core performance. (Image source: Google/Qualcomm/Samsung - edited)

So it can't even beat current gen chips which are almost a year old and it's surely going to get smoked by upcoming chips from Qualcomm and Samsung which are just a few months away.

Overlord999:
So it can't even beat current gen chips which are almost a year old and it's surely going to get smoked by upcoming chips from Qualcomm and Samsung which are just a few months away.

E choke

After all the noise..."approaches"... that's lame

Overlord999:
So it can't even beat current gen chips which are almost a year old and it's surely going to get smoked by upcoming chips from Qualcomm and Samsung which are just a few months away.

How can I help you bros

These are results of upper midrange chipsets so I'm confused as to how it supposedly matches that of flagships like the SD 888

Meister:
These are results of upper midrange chipsets so I'm confused as to how it supposedly matches that of flagships like the SD 888

I taya

Google and too much hype. Tensor fit flop las las.

atheistandproud:


I taya

Google and too much hype. Tensor fit flop las las.

Bros I dey wait your review o

atheistandproud:


I taya

Google and too much hype. Tensor fit flop las las.

Just watched a hands on review from techspurt and it's performance seems pretty sleek

Benjaniblinks:
Bros I dey wait your review o

No problem

Meister:


Just watched a hands on review from techspurt and it's performance seems pretty sleek

If na optimization, then it shouldn't be a problem but I think it's too early for Google to want to chase Exynos, Mediatek and Bionic.

Na step by step.