A16 in iPhone 14 Pro is 17% Faster Than A15 in iPhone 13 Pro in New Benchmark

[calstanford]

THIS is what the M2 should have been based on, not one year old A15. Goes to show Apple is just a phone company.

 

[Darth Tulhu]

At this point, I don’t care how much faster an iPhone is.

It will be faster than anything I want to do on it.

Ecosystem, battery, unique (but useful) features.

That’s all I care about.

Cameras, processor, network connectivity, storage, etc. have all exceeded my needs several gens ago.

It’s like having 1,000HP in a daily driver. Pointless.

 

[bassexpander]

I’m interested in the GPU improvements over my 12PM this time around and the heat management for gaming with max graphic settings at 60 or 120 fps.

And the screen at 2000 nit outdoors in the sun.

Can we cook eggs on the ceramic shield?

 

[rileyes]

wow! Definitely worth dropping a grand for that! 17%!!!!! Of course, there is satellite sos too, which will help save me on one of those crazy off the grid adventures that I never do. But, never say never! We are talking about life or death here. So, worth it! And also, the improved camera will really bring the pictures of my 6 year old, that I only ever send to my mom, to the next level. So, I mean, I owe it to mom to give her the best possible pictures of my kid. Sold!! It’s really a gift for her, if you think about it. Monthly payments? Worth it, because it’s also an investment for my son’s future, because one day he’s going to really appreciate having so many childhood photos of himself where he’s in focus but the background is all blurry. He won’t be able to place where any of those places are, what with the backgrounds being all blurry in portrait mode, but boy do they look pro! You’re welcome, Son. And thank you, Tim.

 

[Velli]

17% faster in synthetic tests, but hardly noticeable in real world use. But some will upgrade regardless.

I think the reason Apple didn’t put up a graph for A15 vs A16 is that it’s main improvements is in other areas than outright performance, for instance performance per watt. But they know that the mainstream customer won’t understand a performance per watt graph.

But, I’m just guessing.

 

[Tony Shi]

Yep. The 4659 score mentioned in the original post of this thread is too low, and thus somewhat inflates the percentage speed boost when A16 is compared. A 4800+ score for A15 multi-core is rather common.

Nope! I can only get 4800+ MT if I put my phone in the freezer. My room temp MT is around 4600-4650. iPhone 13 Pro Max user in Canada here where there’s nothing but igloos.

 

[EntropyQ3]

I think the reason Apple didn’t put up a graph for A15 vs A16 is that it’s main improvements is in other areas than outright performance, for instance performance per watt. But they know that the mainstream customer won’t understand a performance per watt graph.

But, I’m just guessing.

There is no point in a performance per watt graph when you can directly report the two metrics that do matter for the user - performance and battery life. Interestingly, Apple did neither. No suspect graph of performance without scales on the axis, no xx hours of battery life. The one metric they did report was number of transistors (?!?) which looked suspiciously like the number of transistors in the A15 bionic, only rounded up.
Maybe because they needed to somehow reassure buyers of the "Pro" models that they were getting a new SoC.

Apple, like everyone else, is having a tough time advancing performance (and performance/Watt) by headline grabbing amounts, so they emphasize the new colours instead, along with a couple of new features.
No need to assume that mainstream users are idiots and that Apple is trying not to confuse them.

 

[scaramoosh]

I doubt most people would care if it was 100 percent (twice as fast) as their current phone, speed doesn't matter anymore. Like I have a Razor 5G foldable phone with an SoC that this thing would be 10 times faster than...... it doesn't matter, it works smoothly. No one is gonna care that their new phone is .17 times faster than their current one.

 

[EntropyQ3]

I doubt most people would care if it was 100 percent (twice as fast) as their current phone, speed doesn't matter anymore. Like I have a Razor 5G foldable phone with an SoC that this thing would be 10 times faster than...... it doesn't matter, it works smoothly. No one is gonna care that their new phone is .17 times faster than their current one.

While I overall tend to agree, I’d like to point out that the entire point of having fast enough processing is that users shouldn’t notice slowdowns or bad responsiveness. If a user doesn’t notice slowdowns it demonstrates that the system is fast enough for that user, not that it is needlessly fast. And then of course there is a range of users and use cases that Apple tries to satisfy with their phones.

Direct tactile responsiveness is the goal. Now if only ads on the internet didn’t….

 

[dogstar]

While I overall tend to agree, I’d like to point out that the entire point of having fast enough processing is that users shouldn’t notice slowdowns or bad responsiveness. If a user doesn’t notice slowdowns it demonstrates that the system is fast enough for that user, not that it is needlessly fast. And then of course there is a range of users and use cases that Apple tries to satisfy with their phones.

Direct tactile responsiveness is the goal. Now if only ads on the internet didn’t….

The new iPhones can support 4X the ads per internet page displayed! It’s like living in the future!

 

[Velli]

There is no point in a performance per watt graph when you can directly report the two metrics that do matter for the user - performance and battery life. Interestingly, Apple did neither. No suspect graph of performance without scales on the axis, no xx hours of battery life. The one metric they did report was number of transistors (?!?) which looked suspiciously like the number of transistors in the A15 bionic, only rounded up.
Maybe because they needed to somehow reassure buyers of the "Pro" models that they were getting a new SoC.

Apple, like everyone else, is having a tough time advancing performance (and performance/Watt) by headline grabbing amounts, so they emphasize the new colours instead, along with a couple of new features.
No need to assume that mainstream users are idiots and that Apple is trying not to confuse them.

That mainstream users are idiots is more like an established facts that most consumer faced companies are built upon, than an assumption… if mainstream users weren’t idiots, ads wouldn’t exist.

 

[Samdh90]

If it is true that the performance is around 20% faster, it was an interesting choice for Apple not to tell us. Perhaps they didn't want to make the A15 appear inferior to buyers.

Considering the fact that the iphone 14 and 14 plus models still use the A15, it would seem odd for Apple to make it sound inferior. What they did was promote the "Pro" features of getting a Pro model but didn't say, the other is crap. haha.

 

[Velli]

I doubt most people would care if it was 100 percent (twice as fast) as their current phone, speed doesn't matter anymore. Like I have a Razor 5G foldable phone with an SoC that this thing would be 10 times faster than...... it doesn't matter, it works smoothly. No one is gonna care that their new phone is .17 times faster than their current one.

Exactly. I think that Apple are trying hard to not give any specifications at all, and push people to focus on experience. Like in the old days when Rolls Royce just said “enough”, when you asked how much horsepower it has.

 

[Hkfan45]

Exactly. I think that Apple are trying hard to not give any specifications at all, and push people to focus on experience. Like in the old days when Rolls Royce just said “enough”, when you asked how much horsepower it has.

That strategy doesn’t work anymore.

 

[Velli]

That strategy doesn’t work anymore.

No, Apple is clearly a failing company 🙄

 

[Someyoungguy]

Moose’s law is about the number of transistors doubling every 2 years or so, not necessarily about the performance (although it tends to also increase with the transistor count albeit at a somewhat inconsistent rate). 2 years ago the a14’s transistor count was around 11 billion and the a16’s count is 16 billion. 2 years before that the a12 had 7 billion. The slower increase is not too surprising given Apple’s emphasis on efficiency over power with the 13 and 14.

 

[Baja63]

Exactly. I think that Apple are trying hard to not give any specifications at all, and push people to focus on experience. Like in the old days when Rolls Royce just said “enough”, when you asked how much horsepower it has.

Part of the problem is defining “performance“ and how it effects the user experience. It’s difficult for Apple to quickly, through marketing, educate a user on how memory bandwidth performance or an extra graphics core or AI performance improvements will improve their experience. Complicated issues are hard to relay in a graph.

 

[Matsamoto]

Not disappointing as such. But Geekbench gas always been biased in favour of Apple Devices.

Source?

 

[MauiPa]

There are significantly higher MT scores for A15 out there. When I looked this morning, 5455 was about 11% higher than the higher A15 MT scores.

Do you or do you not know how the geekbench browser works? A wide unexplained variability is pretty much there for all devices. Some super, ridiculously low scores are presented which are obviously incorrect, as well as some outlier high scores The low scores should probably be eliminated because they were done incorrectly under high loads. The high scores? I do not know, but the score presented is the average, hence the inappropriateness of retaining obviously flawed scores which lower averages (true for all devices, not just apple). So comparing one score, released early) to the highest score of another device is not valid

 

[MauiPa]

I wouldn't exactly call the higher A15 scores I'm talking about outliers but yes they are at high side of the range. However, the A15 MT score quoted in this thread is a bit on the low side.

Here is what geekbench says. Averages include ridiculously low scores that drag it down

 

[jamesrick80]

At this point, I don’t care how much faster an iPhone is.

It will be faster than anything I want to do on it.

Ecosystem, battery, unique (but useful) features.

That’s all I care about.

Cameras, processor, network connectivity, storage, etc. have all exceeded my needs several gens ago.

It’s like having 1,000HP in a daily driver. Pointless.

Someone here gets it......thank you

 

[dumastudetto]

17% faster in synthetic tests, but hardly noticeable in real world use. But some will upgrade regardless.

I don't ever upgrade for the faster performance anymore. I'm interested only in the new features such as Dynamic Island, the improved camera system, better battery, design tweaks, etc. I upgrade every year and I'm never disappointed.

 

[EugW]

Do you or do you not know how the geekbench browser works? A wide unexplained variability is pretty much there for all devices. Some super, ridiculously low scores are presented which are obviously incorrect, as well as some outlier high scores The low scores should probably be eliminated because they were done incorrectly under high loads. The high scores? I do not know, but the score presented is the average, hence the inappropriateness of retaining obviously flawed scores which lower averages (true for all devices, not just apple). So comparing one score, released early) to the highest score of another device is not valid

I guess you don't follow pre-release Geekbench leaks. Some leaks come from reviewers and are variable, but it's been a pattern for years that there is one early leak that is at the high side of the range, and I suspect they come from Apple itself, to generate press and hype. Geekbench averages for older chips are not very useful as a comparison because they are typically much lower in the curve.

That's besides the point anyway. The main point is that even if you think the single core speed increase is say 9%, 7% of that is attributable to the clock speed difference alone. That leaves all of a 2% difference in IPC. Either that is margin of error from an identical core, or else it's a very similar, tweaked core. Yes, Apple did say there is a potentially a 20% power utilization savings, but we have no idea whatsoever how that's measured and at what spot on the power curve, and it's also on a new process and after a year of improved yields.

 

[AgentAnonymous]

Wow. Apple is really trying to convince me not to upgrade. Well done, Apple. No usb c. Usb 2.0 speeds. 6GB RAM. It’s 2014 android!

USB-C isn't a benefit to anything and shouldn't be presented as a positive "just because; on the contrary, less competition there is worse long term. We don't need socialism for charging ports. Also 14 Pro runs circles around even 2022 phones that run Android.

 

[EntropyQ3]

That mainstream users are idiots is more like an established facts that most consumer faced companies are built upon, than an assumption… if mainstream users weren’t idiots, ads wouldn’t exist.

That ads look the way the do is due to the ad agencies - art directors and copy writers, and how they have done their craft over time. They are meant to capture your attention, first and foremost, not provide information. They need to try to cut through the noise, a noise created by themselves. At this point it basically comes down to shouting louder than all the others.

So the ads are, by and large, just ever louder noise clamoring for your attention. There is no particular assumption regarding the recipients level of intelligence other than the content should be brief and simple (by the standards of the ad producers who are typically not particularly informed regarding the specifics of what they are advertising).
Judging the cognitive level of the population by the intellectual level of ads is a very dubious methodology. 😀