Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
The whole world is going straight into IA, and you all think doubling the neural engine performance is bad? Apart from the redesigned GPU that probably will accellerate ML performance even more?

Just think about the possibilities of running a decently performance ligh ChatGPT in your pocket, without disclosing any of your personal data.

Now picture this scaling up to a M3 Ultra, if the current M2 Ultra performance with Lama model is already 12 tokes/sec…. Rivaling a Single A100 (nvidia best performance neural GPU), we could potentially have two A100 (or more) running in a small factor desktop machine.
 
I think M3 will be better than this 10%/7% boost because with A17 we are looking at the improvement over A16. M2 was likely based on A15, so I think a more reasonable assessment of M3 improvements would be to compare A17 to A15.
I've just had a look at Geekbench 6 results and the A15 is coming out at in the 2700-2900 region (single core) and 6800-7200 region (multicore), meaning that an M3 could show a 20-25% improvement over M2 (CPU).
 
  • Love
  • Like
Reactions: mech986 and Armada2
Always been like that, ipad pro 11 always scored better than the 12,9 counterpart because of lesser pixel resolution, so it taxes the cpu less to drive all those pixels, same logic with pc games where the common practice is to lower resolution to get higher fps or be able to use higher quality settings (Ex. 1080p60 high settings/1080p144 medium settings or run at 4k30 medium settings). I remember the days where the first iphone SE with the A9 was considered a monster for gamming running faster than the mighty iphone 6S because the SE screen was so low res that it improved performance quite a lot compared to the higher res 6s.,
if it s this way, then why the 14 pro and pro max have similar scores?
 
It's also about the cost. The new Apple A17 are manufactured on the 3nm TSMC process, which is much more expensive than the previous process for A16. So, it makes a lot of sense to keep the A16, as that chip still blows the competition away and delivers excellent performance. It can also be that TSMC didn't have the capacity to manufacture so many chips, that they could use them in all the phones, as Apple it the first customer on this process node, but I guess it's more of the cost issue.
If it’s true that they bought TSMC’s whole 3nm capacity, they probably couldn’t put the A17 in everything even if they wanted to. Even once the release surge dies down, a significant chunk is probably going towards M3 chips.
 
It looks like the A17 Pro will only be used in the iPhone 15 Pro, there will be a slightly different A17 (Bionic?) next year for the iPhone 16 and for use in iPads.


Maybe I'm being a bit cynical but I wonder whether next year the A17 Pro chip might get very slightly tweaked, or even not changed at all apart from a move to the N3E process, but in the process get renamed to an A18 so that it isn't quite so apparent that the SoC in the non-Pro models is essentially lagging a year behind what is used in the Pro models.

That wouldn't fool most Apple watchers of course but for the huge number of customers who don't follow sites such as this one having all the iPhone SoCs next year called A18-something (A18 and A18 Pro maybe?) would probably obscure the reusing-last-years-SoC thing for most people.

Let's face it, Apple has form on this. The s8 in the Series 8 Apple watch was, as I understand it, using exactly the same CPU and GPU cores as the previous year's s7 so apart from some changes to peripherals (gyroscope etc?) was identical to the year before in terms of the CPU & GPU cores.
 
Very disappointing gains given the move to 3 nm. ~9% increase in performance core clock speed giving ~10% single core boost and 7% multicore boost. Plus (as has been commented on above), 20% more GPU cores giving a 20% GPU speed boost... Taking into account the rumours from the last year about how the A16 was supposed to have ray tracing but it was dropped late in development due to too much power consumption this is starting to look to me like the A17 is just the original A16 envisioned with the power savings from the 3 nm process sacrificed to the power hungry hardware ray-tracing GPUs. I wonder if battery life will be better than stated unless using ray tracing...

Lots of talk in the Apple Silicon forum about whether M3 is based on A16 or A17 - could it be that A16 and A17 are the same (with just ML cores, ray tracing and USB 3.0 in A17)?

In terms of the iPhone 16 SoC - I've read that the current 3 nm process is a stepping stone before the more scalable second gen 3 nm process and that the differences mean you can't just switch a chip from one process to the other, meaning Apple would have to redesign the chip anyway for the second gen process. Whether they chose to redesign the A17 Pro or just design two versions of the A18 will probably come down to a battle between the financial interests of saving money on the production line (i.e. using binned chips for the regular/pro phones) vs "encouraging" people to buy the Pro models through having a higher SoC generation than the non-Pro.
I wonder if the different 3nm processes is how we’ll end up with an A17 non-pro next year for the iPhone 16.
 
I can see them going with A17 non-Pro next year that doesn't have some Pro features like ray tracing and perhaps omits a GPU core like a binned variant of the A17 Pro. Anything to separate the regular iPhone from the Pro. Not a fan of non-Pro variants coming with last year's chip and the implications that may have for software support. I think in the future we'll see Pro variants getting the new iOS the year the peasant edition gets dropped.
Or perhaps they could be looking into aligning the versions (say next year we get A18 and A18 Pro) where the A18 is a binned version of the A18 Pro that omits a GPU core and has less RAM etc.

Lol peasant version of the iPhone in my mind is the SE, but it works here Mr Flossy ;) lol.

I agree the Pro moniker on the SoC definitely is to further differentiate Pro models from regular iPhones. Im curious what this means for the SE, will it get 2 year older cpu or a further binned current cpu or stay getting the current regular cpu at launch?!

This will affect iOS features further down the road as well as future support of features:
What codecs will come about ProRAW & ProRES still only on the Pro iPhone’s,
- we already know about hardware variances (next year we may see 2+ GPU cores or 2 more efficiency cores for basic tasks at much less battery consumption).
Changes to how iOS works on the Pro’s first for 1-2 years before they trickle down to regular iPhones? Eventual advanced changes or an eventual UI overhaul may only come to Pro’s.
Alas doing more with your iPhone Pro will continue with more integration with macs n iPads n Pro macOS software.

More revenue per unit sold is the goal.
 
  • Love
Reactions: applepotato666
I've never understood why people get so hung up on Geekbench scores. Real world use is all that matters. I haven't owned a smart phone that lagged due to the gpu/processor in well over 7 years.
 
Just think about the possibilities of running a decently performance ligh ChatGPT in your pocket, without disclosing any of your personal data.
It’s not even close to doing that ChatGPT runs on thousands of NVIDIA’s most powerful GPUs.
 
Doesn't sound like M3 will be a huge leap over the M2 if this is any indication.
TSMC has said all along that the jump from 5mn to 3nm would only yield about a 10% performance increase at the same power consumption, so I think people had excessive expectations for the M3. Apple clearly wanted to focus on GPU performance, as despite their claims the M2 Max and Ultra chips offer nowhere near the performance of a dedicated GPU in the desktop space. And the Neural Engine boost is also nothing to sneeze at as a large number of third party apps (ie: Photoshop) use the NPU for certain tasks.

That being said, keep in mind the M2 is based on the A15, so if the M3 is based on the A17 it will be a two generation leap in performance, so closer to 20% CPU and 30-40% GPU when factoring in both the node shrink and the microarchitecture improvements.
 
  • Love
  • Like
Reactions: mech986 and Armada2
Ray tracing is the very last thing id expect from a phone tbh
A17 Pro is the foundation of the M3, so you should expect to see things (like hardware accelerated Ray Tracing) that will be more useful in MacBooks and desktops. Also, with Vision Pro around the corner, I’m sure Apple is working on continuity-type features between iPhone and the headset.
 
I was sort of expecting better battery life with all the talk of the 3nm being more power efficient
Official specs on battery life from Apple are just for music and video playback. We haven’t seen any actual battery life tests yet. Keep in mind that the 8GB of RAM in the A17 Pro will consume more battery than 6GB of RAM in last year‘s chip.
 
Official specs on battery life from Apple are just for music and video playback. We haven’t seen any actual battery life tests yet. Keep in mind that the 8GB of RAM in the A17 Pro will consume more battery than 6GB of RAM in last year‘s chip.
But don't forget that more Ram =less re-open apps, which consumes even more energy
 
  • Like
Reactions: mech986
Ray tracing in a phone lmao.
Only Apple can oversell something so useless
I don't think it's something we'll care about in the foreseeable future, it's a future-proofing thing and a glimpse at what the system UI could require at some point. Also A17 Pro is as powerful if not more powerful than M1 from what I've seen, so I suppose it's also about optimizing production costs later on. If they can stick a rebranded chip that was previously made for an iPhone in a baseline Mac and call it a day, that will save them a ton of cash with the iPhone's massive production runs. It's one of the big reasons they switched the Mac to ARM.
 
  • Like
Reactions: mech986
10 and 20 percent increase isn’t all that great. I’m only pleased because I don’t feel as bad not being able to upgrade this year haha.
 
  • Haha
Reactions: BorisDG
I don’t know, after all the talk all year about this chip, perhaps I was expecting more?
I didn't expect more, but I really did wonder about all the hype from people around here talking about it!
And can some explain ray tracing to me?
99% of the time it really doesn't do a thing unless they use it somehow in photo taking, which I don't see. It's basically enhancing the lighting in rendered scenes to make it look more natural, and with a GPU core to do it, it will have better gaming or any kind of rendering. I don't game or do rendering like that on a phone though, so it's about as interesting as a lump of coal to me.
 
  • Like
Reactions: abatabia
Ray tracing is completely 3D feature aka you need computer graphics/shaders aka gaming. Outside of that, ray tracing is useless for general consumer. Snapgradons/Exynos have RT since years so it's nothing groundbreaking. Still it's cool how far mobile chips has come.
 
I kind of agree as this can open new features and capabilities not possible before. There’s quite a few new software features being added these days that are driven by those parts of the chips. Things that might help sell more phones vs 10% more CPU performance on a phone that no one I’ve ever come across says is too slow.
If it makes Siri work better, I'm all for it, but I don't think I'll be buying a 15 Pro. Maybe a 15 no Pro...

In any case, I'll be waiting to see the real life battery life reports once it gets in people's hands.
 
I re-ran Geekbench 6 on my iMac. Why again is matching an octa-core desktop in performance, not considered to be impressive for a phone? 🤔 🖥️📲

M1: 2,384 and 8,642
A17P: 2,908 and 7,238
Most people just come here to whine about something.
They should just change the name of the site to MacWhiners......
 
  • Like
Reactions: smulji and Woohoo95
I didn't expect more, but I really did wonder about all the hype from people around here talking about it!

99% of the time it really doesn't do a thing unless they use it somehow in photo taking, which I don't see. It's basically enhancing the lighting in rendered scenes to make it look more natural, and with a GPU core to do it, it will have better gaming or any kind of rendering. I don't game or do rendering like that on a phone though, so it's about as interesting as a lump of coal to me.
Thanks so much, and it is as “interesting” to me too. I haven’t gamed since the Atari in 1981.
 
  • Like
Reactions: bobcomer
The whole world is going straight into IA, and you all think doubling the neural engine performance is bad? Apart from the redesigned GPU that probably will accellerate ML performance even more?

Just think about the possibilities of running a decently performance ligh ChatGPT in your pocket, without disclosing any of your personal data.

Now picture this scaling up to a M3 Ultra, if the current M2 Ultra performance with Lama model is already 12 tokes/sec…. Rivaling a Single A100 (nvidia best performance neural GPU), we could potentially have two A100 (or more) running in a small factor desktop machine.

The problem is that all of that is purely academical at this point. By the time applications that take advantage of this like you say are implemented, we will be way past the iPhone 15 Pro.

For most, neural engine performance will be a completely pointless update at this point in 2023, and I don't think it will really change in 2024.
 
  • Like
Reactions: Mac_The_Ripper
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.