close
close

Full Tensor G5 spec leak: Pixel 10 chip details

Full Tensor G5 spec leak: Pixel 10 chip details

Google Tensor Pixel 6

Robert Triggs/Android Authority

Google's Tensor G5, slated for next year's Pixel 10, could be the most anticipated chip in the series yet. For the first time ever, Google is developing the chip entirely in-house, without Samsung having to do the majority of the work. This is expected to make it more competitive as the previous and even current Tensor models lag behind other chipset manufacturers. But will it actually happen? So far we haven't heard anything about the specifications of the new chip, but that's finally changing today.

Thanks to a massive leak from Google's gChips division Android Authority looked at credible documents around the end of the Tensor G5 design that tell us everything about Google's next-generation chip.

You're reading one right now Insights into the authority Story. Discover Authority Insights for more exclusive reports, app teardowns, leaks, and in-depth tech coverage you won't find anywhere else.

More strange CPU decisions

The Pixel 9's Tensor G4 has an improved CPU cluster compared to its predecessor – at least on paper, because in practice the performance is only marginally faster – only 6% faster in our multi-core tests. This is because Google has forgone a mid-cluster core, probably to reduce power consumption.

The Tensor G5 upgrades the CPU cluster again, but not in the way you would probably expect. Google has decided to keep the same single-arm Cortex-X4 primary core, which is an interesting choice as the new Cortex-X925 promises some big improvements. It was also decided to move the core clusters again: the medium cluster now has five Cortex-A725 cores, instead of three Cortex-A720 in the Tensor G4, and the small cluster has been reduced accordingly to just two Cortex-A520 cores. I've put together the following specs:

Tensor G3 (“zuma”) Tensor G4 (“zumapro”) Tensor G5 (“Lagoon”)

Big cluster

Tensor G3 (“zuma”)

1x Arm Cortex-X3

Tensor G4 (“zumapro”)

1x Arm Cortex-X4

Tensor G5 (“Lagoon”)

1x Arm Cortex-X4

Medium cluster

Tensor G3 (“zuma”)

4x Arm Cortex-A715

Tensor G4 (“zumapro”)

3x Arm Cortex-A720

Tensor G5 (“Lagoon”)

5x Arm Cortex-A725

Small group

Tensor G3 (“zuma”)

4x Arm Cortex-A510

Tensor G4 (“zumapro”)

4x Arm Cortex-A520

Tensor G5 (“Lagoon”)

2x Arm Cortex-A520

Of course, it's difficult to predict performance based on specs alone, but a change like this should at least give the Tensor G5 a decent improvement in multi-core performance. However, it is disappointing to see the same Cortex-X4 being used again.

A brand new GPU

One area that particularly surprised me when looking through the documents on the Tensor G5 was the GPU. There aren't many vendors offering GPU IP, and Google has used Arm Mali for all of its previous Tensor chips. Why should that change now? Well, I don't know, but it was like that. The Tensor G5 is equipped with an Imagination Technologies (or IMG) GPU – the DXT-48-1536 at 1.1 GHz.

Unfortunately, we don't know much about the new GPU except for two interesting details: First, it will come with ray tracing support, which is new for Google chips, which usually skip such “gaming” features. The second and perhaps equally important detail is support for GPU virtualization, which allows the use of accelerated graphics in a virtual machine. Google has been working on various virtualization-based features for some time, so it makes sense that these have been integrated.

I have compiled the new GPU specifications in the following table:

Tensor G3 (“zuma”) Tensor G4 (“zumapro”) Tensor G5 (“Lagoon”)

GPU

Tensor G3 (“zuma”)

Arm Mali-G715 (7 cores)

Tensor G4 (“zumapro”)

Arm Mali-G715 (7 cores)

Tensor G5 (“Lagoon”)

IMG DXT (2 cores)

frequency

Tensor G3 (“zuma”)

890MHz

Tensor G4 (“zumapro”)

940MHz

Tensor G5 (“Lagoon”)

1100MHz

Ray tracing

Tensor G3 (“zuma”)

Not supported

Tensor G4 (“zumapro”)

Not supported

Tensor G5 (“Lagoon”)

Supports

GPU virtualization

Tensor G3 (“zuma”)

Not supported

Tensor G4 (“zumapro”)

Not supported

Tensor G5 (“Lagoon”)

Supports

Modest improvements to the AI

Google Pixel 9 with the Pixel Studio app open displays a prompt to download a new AI image model

C. Scott Brown / Android Authority

Google's phones have always had great AI capabilities, and AI is obviously one of the main reasons why Tensor chips even exist. By using specially designed TPUs, Google can achieve far more immersive experiences than would otherwise be possible with a commercially available chip.

The Tensor G5 has a slightly faster TPU compared to the Tensor G4 (which itself was exactly the same as the Tensor G3). The TOPS (Trillions of Operations Per Second) value is almost 40% higher, but that doesn't translate well to real-world performance. Google's internal benchmarks suggest the new TPU is only 14% faster. The improved TPU also has some new features for Google developers, such as small embedded RISC-V cores to enable the execution of operations that are not implemented in hardware, as well as support for on-device training. I've compiled the new TPU specs below:

Tensor G3 (“zuma”) Tensor G4 (“zumapro”) Tensor G5 (“Lagoon”)

GREAT

Tensor G3 (“zuma”)

13 / 6.5 TOPS (INT8/FP16)

Tensor G4 (“zumapro”)

13 / 6.5 TOPS (INT8/FP16)

Tensor G5 (“Lagoon”)

18 / 9 TOPS (INT8/FP16)

Performance improvement over the previous generation

Tensor G3 (“zuma”)

+65%

Tensor G4 (“zumapro”)

0

Tensor G5 (“Lagoon”)

+14%


As we previously revealed, the Tensor G5 is based on TSMC's 3nm class N3E process node, similar to Apple's A18 Pro, for example. Interestingly, the chip size is 121mm^2, while the Apple A18 Pro's chip is only 105mm^2, making the G5 a significantly larger chip.

Even though the specs of Google's upcoming Tensor G5 don't look particularly impressive on paper, it's worth remembering that it's still a chip that will be in a Pixel phone where the software is arguably more important than the hardware . It will be interesting to see how Google uses the new hardware to integrate new features into the Pixel 10 series.

What do you think of the leaked Tensor G5 specs?

683 votes

Leave a Reply

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