Ron Amadeo / Intel
It looks like this custom Google SoC pixel is really going to happen. Echoing reports from about a year ago, 9to5Google is reporting that Pixel 6 should be launched with Google’s custom “Whitechapel” SoC instead of a Qualcomm Snapdragon chip.
The report says that “Google refers to this chip as” GS101 “, with” GS ” potentially being the abbreviation for “Google Silicon”. He also notes that the chip will be shared between the two Google phones that are currently in development, the Pixel 6 and something like a “Pixel 5a 5G”. 9to5 claims to have seen the documentation that points to the involvement of Samsung’s SLSI division (Team Exynos), which aligns with Axios’ previous report saying the chip is “designed in cooperation with Samsung” and must be built on the lines of Samsung’s 5 nm foundry. 9to5Google says the chip “will have some points in common with Samsung Exynos, including software components”.
XDA Developers says it can corroborate the report, saying “According to our source, it looks like SoC will feature a 3-cluster configuration with a TPU (Tensor Processing Unit). Google also refers to its next Pixel devices as ‘fearlessly equipped phones’ which we believe refers to them having an integrated Titan M security chip (code-named “Citadel).” A “3-cluster configuration” would be something like the Snapdragon 888 works, which has three CPU core sizes : a single large ARM Core X1 for large single-thread workloads, three medium Cortex A78 cores for multicore work and four Cortex A55 cores for background work.
The Pixel 6 is expected to launch sometime in the fourth quarter of 2021, and Pixel smartphones always leak a lot, long before they launch. So, I’m sure we’ll see more of that soon.
Whitechapel’s reasonable expectations
It’s easy to get over the top of Google’s first internal SoC for smartphones – “Google is ready to take on Apple!” the headlines will no doubt scream. The fact is that Apple is a $ 2 trillion hardware company and the iPhone is its biggest product, while Google is an advertising company with a hardware division as a small side project. Whitechapel will give Google more control over its smartphone hardware, but Google’s custom chips in the past haven’t exactly set the world on fire, so it’s reasonable to moderate expectations for the company’s first-generation SoC.
Google’s consumer hardware team has shipped several custom chips, and I don’t know if you could call them world champions:
- O Pixel Visual Core on Pixel 2 and 3 there was a custom camera coprocessor created with the help of Intel. Visual Core helped with HDR + processing, but Google managed to achieve the same image quality on the Pixel 3a, which did not have the chip.
- O Pixel Neural Core on Pixel 4 it was derived from the efforts of the company’s Tensor Processing Unit (TPU) AI accelerator and had a similar job doing camera work and AI voice recognition. It wasn’t important enough to just cut the Pixel 5 entirely.
- There was the aerial gesture detection chip, Soli Project, on Pixel 4. This was a concept of radar on a chip that Google originally proposed as capable of detecting “submillimetric movements of your fingers”, but when it was commercialized, it could only detect large, waving arms gestures. The feature still exists today in the new Nest Hub, for sleep tracking, but it was not good enough to make the leap to the Pixel 5.
- The company Titan M security chip works as the secure element on some Pixel phones. Google says this makes Pixel phones more secure, although a roughly equivalent secure element also comes with a Qualcomm chip, or at least the company has never shown a tangible difference.
I think the biggest benefit that we’ll see in a Google SoC is an expanded update schedule. Android updates are much smoother when you get support from the SoC manufacturer, but Qualcomm drops all of its chips after the three-year mark for major updates. This lack of support makes updates significantly more difficult than they need to be, and today is where Google is limited to updates. With Qualcomm out of the way, there is no excuse for Google not to comply with Apple’s five-year update policy on Apple’s iPhone. With a custom SoC, Google will have complete control over how long it can update devices.
Google is currently in the embarrassing position of offering less support for its devices than Samsung, which now offers up to three years of major updates (at most from Qualcomm) and four years of security updates, while Google offers only one year less security updates. It is a strange position for Google, which previously led the ecosystem in support of hardware. Perhaps Google did not immediately match Samsung because it is awaiting the launch of the Pixel 6, where it will announce dramatically longer support schedules thanks to its own chip?
In fact, competing in the SoC business is difficult
Aside from easier updates, I don’t know if we can expect much from Whitechapel. Many Android manufacturers have made their own chips now, with varying degrees of success. Samsung has the Exynos line. Huawei has its HiSilicon chips. Xiaomi made the Surge S1 SoC in 2017, recently launched the Surge C1 camera chip in the Xiaomi Mi Mix Fold and has an investment in a silicon designer. Oppo is also working on the development of internal chips. None of the existing efforts has been able to significantly beat Qualcomm, and most of these companies (except Huawei) still choose Qualcomm over their own chips for major devices. Everyone, even Qualcomm, relies on the same company, ARM, for their CPU designs, so there is not much room for differences between them. When everyone is using out-of-the-box ARM CPU designs, the main areas of differentiation remaining are the GPU and the modem, two areas where Qualcomm excels, so it is chosen for most major devices.
Companies that take hardware seriously do their best to separate themselves from the basic ARM CPU designs, choosing to design their own cores based on the ARM instruction set. Apple dominates the performance of the mobile CPU thanks to the acquisition of an entire semiconductor company, PA Semi, in 2008. Qualcomm is doing its best to achieve it by buying Nuvia, a chip design company founded by some of these former Apple chip designers, and plans to launch its internally designed CPUs in 2022. Google has made some chip design hires, but these are split between separate hardware and server teams and pale compared to buying one entire company. When even Qualcomm isn’t shipping custom chips, I don’t see how Google uses anything other than commercially available ARM CPU designs.
Google’s modem and GPU solutions will be an area of great interest. There are not many GPU designs available. Qualcomm has its own Adreno division, which it bought years ago from ATI. Samsung has an agreement with AMD for its future GPUs, but I doubt that this is possible in its partnership with Google. If this chip is actually adjacent to Exynos, Samsung and many other SoC vendors also operate with ARM Mali GPUs, which are generally not competitive with what Qualcomm offers. Samsung signed this partnership with AMD for a reason!
Imagining Google’s SoC with an integrated modem is a challenge. Generally, you are unable to integrate a modem into your SoC, unless you are the owner of the modem project, and Google does not have a modem IP. Samsung produced chips with integrated 5G modems, but they generally don’t come to the U.S., so a Samsung modem would require sharing the design with Google and bringing it to the U.S. for the first time. Qualcomm is, of course, the queen of strong-framed companies with its IP modem and keeping competitors out of the United States, and is also generally a leader in modem technologies like 5G. Apple has succeeded so far with separate cellular modems – today the iPhone 12 comes with a low-profile Qualcomm modem for 5G, which is probably the most likely option for Google. Apple also bought Intel’s modems division for a billion dollars, indicating that it is working toward onboard modem technology.
Along with the usual CPU / GPU / modem options, Google could also include some camera and special AI sauce in the form of some kind of coprocessor (hopefully we will also get the first Pixel camera sensor upgrade in four years) . Google is also likely to include a Titan security chip. Even if it did, I can’t imagine it making a big difference compared to something like sending with a low-quality GPU or modem. Google has never shown a great benefit to the end user of its custom silicon in the past, just a lot of hype.
It’s hard to be optimistic about the future of Google’s SoC when the company doesn’t seem to be making the big money acquisitions and licensing deals that Apple, Qualcomm and Samsung are making. But at least it’s a start.