Google's Tensor inside of Pixel 6, Pixel 6 Pro: A Look into Performance & Efficiency
by Andrei Frumusanu on November 2, 2021 8:00 AM EST- Posted in
- Mobile
- Smartphones
- SoCs
- Pixel 6
- Pixel 6 Pro
- Google Tensor
Conclusion & End Remarks
Google’s newest Pixel 6 and 6 Pro are definitely most interesting devices, as in many ways they represent Google most competitive and value-rich phones the company has been able to make in years. While today’s article isn’t focusing on the device itself – more on that in a later review, including more in-depth camera coverage, what we did have a deeper look today was at the new chip powering the phones, the new Google Tensor.
The company notes that the primary reason they saw the need to go with a customized silicon approach, was that current merchant silicon solutions didn’t allow for the performance and efficiency for machine learning tasks that the company was aiming for in their devices. This performance and efficiency is used to enable new use-cases and experiences, such as the many ML features we see shipped and demonstrated in the Pixel 6 series, such live transcribing, live translation, and image processing tricks, all that run on the Tensor’s TPU.
While Google doesn’t appear to want to talk about it, the chip very clearly has provenance as a collaboration between Google and Samsung, and has a large amount of its roots in Samsung Exynos SoC architectures. While yes, it’s a customised design based on Google’s blueprints, the foundation means that some of the defining characteristics of Exynos chips is still found on the Tensor, particularly power efficiency is one area of the SoCs that are very much alike in, and that also means that the Tensor falls behind, much like the Exynos, against Qualcomm’s Snapdragon solutions when it comes to battery life or efficiency.
Google’s CPU setup is a bit different than other SoCs out there – a 2+2+4 setup with X1 cores, A76 cores and A55 cores is unusual. The two X1 cores are fine, and generally they end up where we expected them, even if there’s a few quirks. The A76 cores, ever since we heard those rumours months ago that the chip would feature them, made no sense to us, and even with the chip in our hands now, they still don’t make any sense, as they clearly fall behind the competition in both performance and efficiency. Who knows what the design process looked like, but it’s just one aspect of the chip that doesn’t work well.
GPU performance of the Tensor seems also lacklustre – while it’s hard to pinpoint wrong-doings to the actual SoC here, Google’s choice of going with a giant GPU doesn’t end up with practical advantages in gaming, as the phones themselves have quite bad thermal solutions for the chip, not able to properly dissipate the heat from the chip to the full body of the phones. Maybe Google makes more use of the GPU for burst compute workloads, but so far those were hard to identify.
So that leads us back to the core aspect of the Tensor, the TPU. It’s the one area where the SoC does shine, and very clearly has large performance, and likely also efficiency advantages over the competition. The metrics here are extremely hard to quantify, and one does pose the question if the use-cases and features the Pixel 6 comes with were really impossible to achieve, on say a Snapdragon chip. At least natural language processing seems to be Google’s and the Tensor’s forte, where it does have an inarguably large lead.
One further aspect that isn’t discussed as much is not related to the performance of the chip, but rather the supply chain side of things. We of course have no idea what Google’s deal with Samsung looks like, however both new Pixel 6 phones are devices that seemingly are priced much more aggressively than anything we’ve seen before from the company. If this is related to the SoC bill of materials is just pure speculation, but it is a possibility in my mind.
In general, I do think Google has achieved its goals with the Tensor SoC. The one thing it promises to do, it does indeed do quite well, and while the other aspects of the chip aren’t fantastic, they’re not outright deal-breakers either. I still think energy efficiency and battery life are goals of highest priority in a design, and there we just absolutely need to see better improvements in the next generation Tensor. We don’t know what path Google is taking for future designs, but it’ll be interesting to see.
We’ll be following up with a more in-depth review of the actual Pixel 6 phones, starting with a camera-focused article – stay tuned.
Facebook
Twitter
RSS
108 Comments
View All Comments
Alistair - Tuesday, November 2, 2021 - link
It's very irritating how slow Android SOCs are. I'll just keep on waiting. Won't give up my existing Android phone until actual performance improvements arrive. Hopefully Samsung x AMD will make a difference next year.Speedfriend - Thursday, November 4, 2021 - link
Looking at the excellent battery life of the iPhone 13 (which I am currently waiting for as my work phone) does iPhone till kill suspend background tasks. When I used to day trade, my iPhone would stop prices updating in the background, very annoying when I would flick to the app to check prices and unwittingly see prices hours old.ksec - Tuesday, November 2, 2021 - link
Av1 hardware decoder having design problem again?Where have I heard of this before?
Peskarik - Tuesday, November 2, 2021 - link
preplanned obsolescencetuxRoller - Tuesday, November 2, 2021 - link
I wonder if Google is using the panfrost open source driver for Mali? That might account for some of the performance issues.TheinsanegamerN - Tuesday, November 2, 2021 - link
Seems to me based on thermals that the pixel 6/pro suffer from thermal throttling, and thus have power power budgets, then they should have given the internal hardware, leading to poor results.Makes me wonder what one of these chips could do in a better designed chassis.
name99 - Tuesday, November 2, 2021 - link
I'd like to ask a question that's not rooted in any particular company, whether it's x86, Google, or Apple, namely: how different *really* are all these AI acceleration tools, and what sort of timelines can we expect for what?Here are the kinda use cases I'm aware of:
For vision we have
- various photo improvement stuff (deblur, bokeh, night vision etc). Works at a level people consider OK, getting better every year.
Presumably the next step is similar improvement applied to video.
- recognition. Objects, OCR. I'd say the Apple stuff is "acceptable". The OCR is genuinely useful (eg search for "covid" will bring up a scan of my covid card without me ever having tagged it or whatever), and the object recognition gets better every year. Basics like "cat" or person recognition work well, the newest stuff (like recognizing plant species) seems to be accurate, but the current UI is idiotic and needs to be fixed (irrelevant for our purposes).
On the one hand, you can say Google has had this for years. On the other hand my practical experience with Google Lens and recognition is that the app has been through so many rounds of "it's on iOS, no it isn't; it's available in the browser, no it isn't" that I've lost all interest in trying to figure out where it now lives when I want that sort of functionality. So I've no idea whether it's better than Apple along any important dimensions.
For audio we have
- speech recognition, and speech synth. Both of these have been moved over the years from Apple servers to Apple HW, and honestly both are now remarkably good. The only time speech recognition serves me poorly is when there is a mic issue (like my watch is covered by something, or I'm using the mic associated with my car head unit, not the iPhone mic).
You only realize how impressive this is when you hear voice synth from older platforms, like the last time I used Tesla maybe 3 yrs ago the voice synth was noticeably more grating and "synthetic" than Apple. I assume Google is at essentially Apple level -- less HW and worse mics to throw at the problem, but probably better models.
- maybe there's some AI now powering Shazam? Regardless it always worked well, but gets better and faster every year.
For misc we have
- various pose/motion recognition stuff. Apple does this for recognizing types of exercises, or handwashing, and it works fine. I don't know if Google does anything similar. It does need a watch. Not clear how much further this can go. You can fantasize about weird gesture UIs, but I'm not sure the world cares.
- AI-powered keyboards. In the case of Apple this seems an utter disaster. They've been at it for years, it seems no better now with 100x the HW than it was five years ago, and I think everyone hates it. Not sure what's going on here.
Maybe it's just a bad UI for indicating that the "recognition" is tentative and may be revised as you go further?
Maybe the model is (not quite, but almost entirely) single-word based rather than grammar and semantic based?
Maybe the model simply does not learn, ever, from how I write?
Maybe the model is too much trained by the actual writing of cretins and illiterates, and tries to force my language down to that level?
Regardless, it's just terrible.
What's this like in Google world? no "AI"-powered keyboards?, or they exist and are hated? or they exist and work really well?
Finally we have language.
Translation seems to have crossed into "good enough" territory. I just compared Chinese->English for both Apple and Google and while both were good enough, neither was yet at fluent level. (Honestly I was impressed at the Apple quality which I rate as notably better than Google -- not what I expected!)
I've not yet had occasion to test Apple in translating images; when I tried this with Google, last time maybe 4 yrs ago, it worked but pretty terribly. The translation itself kept changing, like there was no intelligence being applied to use the "persistence" fact that the image was always of the same sign or item in a shop or whatever; and the presentation of the image, trying to overlay the original text and match font/size/style was so hit or miss as to be distracting.
Beyond translation we have semantic tasks (most obviously in the form of asking Siri/Google "knowledge" questions). I'm not interested in "which is a more useful assistant" type comparisons, rather which does a better job of faking semantic knowledge. Anecdotally Google is far ahead here, Alexa somewhat behind, and Apple even worse than Alexa; but I'm not sure those "rate the assistant" tests really get at what I am after. I'm more interested in the sorts of tests where you feed the AI a little story then ask it "common sense" questions, or related tasks like smart text summarization. At this level of language sophistication, everybody seems to be hopeless apart from huge experimental models.
So to recalibrate:
Google (and Apple, and QC) are putting lots of AI compute onto their SoCs. Where is it used, and how does it help?
Vision and video are, I think clear answers and we know what's happening there.
Audio (recognition and synth) are less clear because it's not as clear what's done locally and what's shipped off to a server. But quality has clearly become a lot better, and at least some of that I think happens locally.
Translation I'm extremely unclear how much happens locally vs remotely.
And semantics/content/language (even at just the basic smart secretary level) seems hopeless, nothing like intelligent summaries of piles of text, or actually useful understanding of my interests. Recommendation systems, for example, seem utterly hopeless, no matter the field or the company.
So, eg, we have Tensor with the ability to run a small BERT-style model at higher performance than anyone else. Do we have ways today in which that is used? Ways in which it will be used in future that aren't gimmicks? (For example there was supposed to be that thing with Google answering the phone and taking orders or whatever it was doing, but that seems to have vanished without a trace.)
As I said, none of this is supposed to be confrontational. I just want a feel for various aspects of the landscape today -- who's good at what? are certain skills limited by lack of inference or by model size? what are surprising successes and failures?
dotjaz - Tuesday, November 2, 2021 - link
" but I do think it’s likely that at the time of design of the chip, Samsung didn’t have newer IP ready for integration"Come on. Even A77 was ready wayyyy before G78 and X1, how is it even remotely possible to have A76 not by choice?
Andrei Frumusanu - Wednesday, November 3, 2021 - link
Samsung never used A77.anonym - Sunday, November 7, 2021 - link
Exynos 980 uses Cortex-A77