Apple Event Rumors: iPad Pro With M4 Chip and New Apple Pencil With Haptic Feedback

[ric22]

Yes 2x of A16. Then you can do some research A16NPU -> A15 NPU -> M2 NPU -> M3 NPU. I also read about it and I couldn't find any evidence that the M3 got the NPU of the A17 pro. So according data it looks like it uses the NPU of A15/M2. A16 got the same, is a bit faster because it got higher clock than A15.

https://venturebeat.com/ai/as-ai-chips-improve-is-tops-the-best-way-to-measure-their-power/

TOPS is all a bit airy fairy pixie dust make believe, "customers don't need concrete facts" mumbo jumbo.

Give it a year or two and I'm sure someone will create some benchmarking tools to give us something with a bit more substance than a "theoretical" measure like this.

 

[ric22]

Yes 2x of A16. Then you can do some research A16NPU -> A15 NPU -> M2 NPU -> M3 NPU. I also read about it and I couldn't find any evidence that the M3 got the NPU of the A17 pro. So according data it looks like it uses the NPU of A15/M2. A16 got the same, is a bit faster because it got higher clock than A15.

The article that I read, which I can't for the life me of find now (I did try for you), showed that all the M series chips will out perform all the A series chips in AI tasks, regardless of how many gazillion TOPS each is supposed to have, and also had a look at how the physical size of the NE in the M series chips is physically bigger too.

 

[ric22]

@Torty an interesting page for you from these boards on the subject:

M2 vs. M3 Buyer's Guide: How Much Better Really Is M3?

Why is the M3' Neural engine so much worse than the A17 pro? Source: Wikipedia

forums.macrumors.com

Edit: Torty, I just saw that you were actively participating in that thread, and said thanks to the guy who provided all the info on how there's a LOT LOT LOT more to it than just comparing these fantasy land TOPS numbers. 🧐 👀

 

[Velli]

If the May '24 iPads Pro indeed get M4 then there has to be some weird beating around the AI bush to avoid killing most of the WWDC '24 and iPhone 16 hype.

There has been several WWDC’s that were software only. Totally possible that WWDC24 is going to be 2 hours of “AI this, AI that”, with all hardware pre-announced.

I don’t see how it is relevant to iPhone 16 hype?

 

[Torty]

@Torty an interesting page for you from these boards on the subject:

M2 vs. M3 Buyer's Guide: How Much Better Really Is M3?

Why is the M3' Neural engine so much worse than the A17 pro? Source: Wikipedia

forums.macrumors.com

Edit: Torty, I just saw that you were actively participating in that thread, and said thanks to the guy who provided all the info on how there's a LOT LOT LOT more to it than just comparing these fantasy land TOPS numbers. 🧐 👀

Yes but I Didn’t found any evidence that M3 uses the same NPU as A17 pro. And I found out that apple not only published the TOPS numbers but also said it’s up to 2x faster. That’s a big difference. They cant say it’s faster if it’s only some playing with numbers or they would lie.

 

[Torty]

https://venturebeat.com/ai/as-ai-chips-improve-is-tops-the-best-way-to-measure-their-power/

TOPS is all a bit airy fairy pixie dust make believe, "customers don't need concrete facts" mumbo jumbo.

Give it a year or two and I'm sure someone will create some benchmarking tools to give us something with a bit more substance than a "theoretical" measure like this.

As I wrote above it‘s not only about TOPS numbers but also the statement that it’s “up to 2x faster”.

also interesting: https://www.notebookcheck.net/Apple...-the-A17-Pro-A16-Bionic-designs.764280.0.html

 

[ric22]

Yes but I Didn’t found any evidence that M3 uses the same NPU as A17 pro. And I found out that apple not only published the TOPS numbers but also said it’s up to 2x faster. That’s a big difference. They cant say it’s faster if it’s only some playing with numbers or they would lie.

It can logically only be 2x faster if the M3 shares the A16 NE. Where's the evidence for that?!? Also, as someone else pointed out to you, the M series chips have 3x or more bandwidth than the A series chips.

 

[jonnysods]

I think the M3 transition is complete. Rumors are the Mac Studio, Mac mini, and Mac Pro will skip the M3 family. Since those are the only computers not running the M3 family, then that means there aren’t any more M3 announcements left, except maybe the iPad Air. But that’ll happen simultaneously with the iPad Pros.

M3 only came out in October, seems quite for M4 to come out though don't you think?

 

[Frantisekj]

Skipping a gen for the chip while going OLED + redesign? Too many upgrades in one update. Does not sound like Apple.

But I think Apple needs dethrone Qualcomm X chip so it woudl make sence with other reasons like cost of manufacturing.

 

[Realityck]

M3 only came out in October, seems quite for M4 to come out though don't you think?

Yes too quiet.

I’d like to see someone point us to any documentation or spoken communication that even names a M4 From Apple?

macworld november 2023 Article

We could see an M4 chip towards the end of 2024. But even if we do, we’ve entered a maturation phase of the M-series, similar to what we experienced with Intel chips.

April 11, 2024

Apple is getting ready to refresh its entire lineup of Macs with M4 processors that have an emphasis on AI, according to a report from Bloomberg’s Mark Gurman. The new M4-equipped Macs will reportedly start arriving later this year, with the rest coming in early 2025

April 28, 2024

Apple is preparing for its big AI coming out party in this year’s Worldwide Developer Conference; that much you can count on. But apparently, the company is going to start that party a little early with the OLED iPad Pro that it’s expected to unveil on May 7th. According to Bloomberg’s Mark Gurman, there’s “a strong possibility” the tablet will launch with an M4 chip and its accompanying neural engine, making it Apple’s “first truly AI-powered device.”

This speculation has no other parties involved. Just replicated. The M3 was announced Oct 30th 2023, slightly more then 6 months ago.

 

[DaveCummings]

Who spread the rumors that was the case? Mark Gurman right? Apple hasn’t implied that was what they were doing.

I think it's funny that Gurman said M4 on the iPad next week and people are acting like it's confirmed....but this is the guy who said the iPads were going to come out in March, then April and also said iPads weren't going to have an event

 

[Realityck]

It makes no sense to give the iPad Pro an M4 chip before the Mac

Apple's confusing tablet range is about to start creating problems for other products.

www.macworld.com

Giving the iPad Pro an M4, then, would be a bizarre decision for two reasons. It wouldn’t really benefit the iPad Pro, because it doesn’t need an M4. And it would hurt the Mac range, whose M2 and M3 machines would instantly appear underpowered, and create confusion for Mac buyers.

The only beneficiary of the decision would be Apple’s tablet marketing department, who need an eye-catching way to upsell iPad buyers from the Air to the Pro model. Don’t buy that stinky old M2-powered Air! Buy this M4-powered Pro instead. “It does AI.”™️

Mark Gurman, the author of that surprising report, points out the M4’s advanced neural engine as a main reason for Apple to add the M4 to “lay out its AI chip strategy without distraction.” If that’s the case, perhaps Apple will brand the iPad Pro’s new chip as a unique edition of the M3 (Bionic? Plus?) as it did with the A12X in 2018 and A12Z in 2020. Then Apple can focus on the neural engine and better AI capabilities, eliminate any potential confusion with the Mac, and still give the iPad Pro its own identity.

The last comment is an interesting point, what happens if Apple indeeds includes a special M3 SoC for the iPads.

 

[tobybrut]

M3 only came out in October, seems quite for M4 to come out though don't you think?

Don’t know. It depends on the cost difference between the N3B and N3E processes. N3B is supposedly a lot harder to produce and a lot more expensive. The M3 is N3B based while M4 will almost certainly be N3E based. Apple’s incentive is to get off the more expensive process faster. The A18 for the upcoming iPhone 16 will certainly be N3E, specializing in neural engine performance. It makes sense for Apple to get their iPad Pro to that new neural engine as fast as possible. Apple’s not going to release another iPad Pro for at least a year, if not 18 months after this one, so they can’t really afford to have an inferior neural engine for that long.

While I would agree it seems like a very short time since the M3, N3B was always a stopgap process. I would also point out that the M3 MBP’s came out merely 9 months after the M2 MBP’s. It seems Apple’s release schedule is accelerating, possibly due to outside competition. Perhaps Apple is worried about the upcoming Snapdragon X Elite.

 

[jonnysods]

Wow you’re right I didn’t know M3 was only 9 months after M2.

 

[Realityck]

I would also point out that the M3 MBP’s came out merely 9 months after the M2 MBP’s. It seems Apple’s release schedule is accelerating,

June 6, 2022 M2 announced. 13" MBA/13" MBP
Jan 17th 2023 M2 Pro/Max MBP announced also Mac mini using M2 and M2 Pro
Oct 30, 2023 M3/M3 Pro/M3 Max MBPs announced

So how does the difference in M2 base to M3 Base denote that a M4 base will be only 9 months later? When it took more than a year June WWDC 2022 to almost Nov 2023 for it to happen? 🧐

 

[Neil Harrison]

Skipping a gen for the chip while going OLED + redesign? Too many upgrades in one update. Does not sound like Apple.

Thought the same especially with landscape camera rumours

OLED 11 inch redesigned iPad Pro with landscape camera and sporting M4 chips is just too much of an upgrade

 

[tobybrut]

June 6, 2022 M2 announced. 13" MBA/13" MBP
Jan 17th 2023 M2 Pro/Max MBP announced also Mac mini using M2 and M2 Pro
Oct 30, 2023 M3/M3 Pro/M3 Max MBPs announced

So how does the difference in M2 base to M3 Base denote that a M4 base will be only 9 months later? When it took more than a year June WWDC 2022 to almost Nov 2023 for it to happen? 🧐

There was only a 9 month gap between the introduction of the M3 Pro and M3 Max from the introduction of the M2 Pro and M2 Max. When going from M1 family to M2, it took about 18 months. It just looks like Apple’s product introductions are accelerating. I was guessing that there were two reasons for that. The M3’s, which are based on the N3B process, were only meant to be stop gap and competition is getting fiercer. I think Apple wanted to get out of the N3B into N3E as fast as they could while at the same time keeping an eye on that Snapdragon X Elite that does beat the M3’s in some respects.

 

[Torty]

There was only a 9 month gap between the introduction of the M3 Pro and M3 Max from the introduction of the M2 Pro and M2 Max. When going from M1 family to M2, it took about 18 months. It just looks like Apple’s product introductions are accelerating. I was guessing that there were two reasons for that. The M3’s, which are based on the N3B process, were only meant to be stop gap and competition is getting fiercer. I think Apple wanted to get out of the N3B into N3E as fast as they could while at the same time keeping an eye on that Snapdragon X Elite that does beat the M3’s in some respects.

Why did they release the MBA with the M3 instead of M4?

 

[Chuckeee]

Why did they release the MBA with the M3 instead of M4?

Because Apple does not want customers selecting an MBA over an MBP because it has a better processor

 

[LRB55]

Skipping a gen for the chip while going OLED + redesign? Too many upgrades in one update. Does not sound like Apple.

Agree, too strong a competitor for the MacBook lines.

 

[jazz1]

Agree, too strong a competitor for the MacBook lines.

As a consumer and fan I’m so confused about Apple’s upgrade hierarchy for its devices! Maybe there is really not a hierarchy at all and just too many factors in play. Apple, I guess, has to consider factors in and out of their control while upgrading anything. As well as $$$$ profit.

 

[Mimiron]

with these beast M chips ipads are like ferraris limited by their os, make ipados on par with mac and it would make sense to include m chips in there.

 

[Torty]

As a consumer and fan I’m so confused about Apple’s upgrade hierarchy for its devices! Maybe there is really not a hierarchy at all and just too many factors in play. Apple, I guess, has to consider factors in and out of their control while upgrading anything. As well as $$$$ profit.

This are the other factors:

 

[DaveCummings]

It feels way too early for M4s. If the iPhones came out and the A18 chips were out, then I’d lend more validity to this M4 rumor, because the A series chip is always the blueprint for the M series chips. I mean, the A17 laid the groundwork for M3

 

[ric22]

Hi @Torty I finally stumbled across the scientific explanation again of why the 35TOPS advertised on the iPhone 15 Pro was in no way more powerful than the 18TOPS on the M3- it's all to do with how it's calculated, specifically the level of precision achieved by the chip. They began basing the numbers on INT8 precision.

"Apple's M3 was rated for 18 TOPS at FP16 precision, but the M4 is rated for 38 TOPS with INT8. That means, if equalized to INT8 precision, we're looking at a 5% improvement in TOPS for the M4 over the M3."

Alleged Apple M4 Geekbench scores show incremental improvement in machine learning over last gen

Incremental improvement over last gen

www.tomshardware.com

😊 Glad I finally found it! It bugged me that I wasn't able to find the explanation again, until now. It shows how unreliable TOPS numbers are when compared to each other. Apple were, for example, bragging about the huge leap in the M4, whereas it was just a baby step forwards. Clearly a huge leap sounds better, as does the doubling of numbers they know people are confused by.