TSMC starts production of A15 for iPhone 13. Macbook Pros will probably get SoCs based on A15.

[Serban55]

Apple is obviously going to milk the M1X name before going to M2 and M2X.

and why not? The M1 become a trademark by now in tech industry
And they will stick with the M mark
M2 will come in the fall with the mba being the first product with it..probably the next 24" imac will be next next year

 

[Serban55]

Now that we have all the various M1 Macs released together with the M1 iPad Pro, it got to me thinking that maybe the M1 is supposed to be the A14X but Apple was surprised by the performance they got out of it that they quickly bolted on the PCIe + whatever circuitry required to ship the base Macs with it. Craig Federighi did say that they were surprised by the M1's performance. That may explain why the I/Os (including external display support) are limited in the existing M1 Macs.

I suspect the next round of the Mx SoC are the actual SoC meant for the Macs.

Not a chance... M1 become a benchmark and a trademark for success in the tech industry. when you say M1, you say SoC quality and how great it is/it was. So they want that marketing to go with ipad pros as well.
And again, people are saying thunderbolt is locked for intel cpu...well....microsoft surface still dont have thunderbolt but ipad pro has it...

 

[Lemon Olive]

All these are conjectures. Maybe it’s common sense to you, but it doesn’t make it universally true. The simple fact is that we don’t know. One can argue one way or the other, there are plenty of good arguments for either side. My intuition tells me that the Pro Silicon will be very different from M1, because M1 microarchitecture simply doesn’t cut it for prosumer chips.

I'm tired of this.

"I want to be right" is not a position for argument. Use the available information to determine what is most likely, and go with it. Do not dismiss that information and say "Nah I'd rather think this, so I'm going to talk about as if it were more likely than what all available information says."

People have this idea that any position is as valid as any other and that simply is not true. ALL available information to this point indicates 1 direction here with extreme weight in its favor. If you want to argue that the possibility that has less than a 10% chance of being true is more likely, I can't take you seriously.

 

[Lemon Olive]

Apple is obviously going to milk the M1X name before going to M2 and M2X.

It's not a "name". They are going to use the M1 architecture as along as necessary, including using the higher performance variants that they had already developed a year ago and scheduled for products for 2021.

The people talking about the M2 being made right now for the MBP's in July are just very confused. They don't realize the chip for the MBP was reported on over a year ago, and was finalized at some point not long after the M1 release. They also don't realize that the A15 is still way too young. They also don't realize that no derivative of the A15 is going to ship before the A15 itself ships (the fab process takes time to ramp up for a new architecture and iPhone is priority 1...just like last year). They also don't realize that chip generations and microarchitectures is not how consumer and pro are segregated.

They are just pretending for fun.

 

[Lemon Olive]

Agreed. It seems like M1 is basically just a rebranded A14X which was going to be made for the iPad Pro anyways.

The SoCs going forward for Macs will actually be designed for Macs from the start.

No. That's not how any of this works. Just stop.

 

[CWallace]

Agreed. It seems like M1 is basically just a rebranded A14X which was going to be made for the iPad Pro anyways.

The SoCs going forward for Macs will actually be designed for Macs from the start.

The M1 was designed for Macs from the start. It was also very likely designed for the iPad Pro from the start.

Apple did not decide to move from Intel on a whim. So they clearly did not decide on a whim to use an iPhone SoC and slap some "Mac stuff" on it and call it good. The last thing Apple needed was for the first generation of Apple Silicon Macs to be little better - or worse - than their Intel models. They needed to "hit it out of the park" to get us all excited for the move.

And so they developed a chip that would do just that - and they called it "M1".

 

[jsnuff1]

No. That's not how any of this works. Just stop.

You dont know how it works either...unless you work in Apple's Mac Engineering department which I doubt it

 

[CWallace]

You dont know how it works either...unless you work in Apple's Mac Engineering department which I doubt it

Common Sense should be enough to make reasoned inferences, however.

And common sense tells me that Apple would have developed M1 from the start for the Mac (as well as the iPad Pro to be ready when iPad OS is as professional as the iPad Pro, itself).

 

[NotTooLate]

What accounts? There are no direct accounts making concrete claims, and no direct evidence. All we have are rumors and circumstantial evidence. Based in the later, I believe Pro Apple Silicon to be based on a different microarchitecture than M1.

I am not sure you and lemon are talking the same language , can you specify what you mean by M1 different uarch then the pros ? Lemon is talking about the cpu generation , are you also referring to it ? Or to something else ? Like how the interconnect works , memory subsystem and such ? I don’t think anyone that understand cpu design think you can just “slap more cpu and GPU” and call it a day , and it doesn’t seems to be lemon point , maybe if you can clarify here , you might even reach an agreement!

 

[leman]

I'm tired of this.

"I want to be right" is not a position for argument. Use the available information to determine what is most likely, and go with it. Do not dismiss that information and say "Nah I'd rather think this, so I'm going to talk about as if it were more likely than what all available information says."

I agree with this 100%. That's why I find it rather ironic that your argument so far boils down to "I am right and you are wrong". We both have equal rights to our conjectures, and we should be able to discuss our reasoning as rational people.

People have this idea that any position is as valid as any other and that simply is not true. ALL available information to this point indicates 1 direction here with extreme weight in its favor. If you want to argue that the possibility that has less than a 10% chance of being true is more likely, I can't take you seriously.

From my perspective (as an experienced software developer and someone who has been watching Apple's hardware progress closely for years), I would be surprised if the prosumer Apple Silicon is using the same microarchitecture as M1. My reasoning is:

- More than half a year has passed since M1 release, with no M1X-style chips in sight (the longer the delay, the less sense would such a chip make)
- Firestorm in M1 peaks at 3.2 ghz which is plenty for the entry-level hardware but not enough to take the performance crown for the prosumer hardware, it is questionable whether Firestorm can be stable at higher clocks
- Apple introduced ray tracing support in Metal last year, heavily hinting that hardware RT support is coming — and that it's coming this year. Apple G13 does not support hardware RT, ergo it is likely that the prosumer hardware will require a new GPU architecture
- Prosumer hardware is a likely candidate to support advanced vector extensions (SVE2). It would be odd if the first generation would not have that support, ergo it is likely that the prosumer hardware will require a new CPU architecture
- The leaks point to a 8+2 core configuration. All Firestorm/Icestorm chips so far feature 4 Icestorm cores. This again suggests that the prosumer hardware will use a different microarchitecture, with redesigned, more capable low-power cores

 

[leman]

The M1 was designed for Macs from the start. It was also very likely designed for the iPad Pro from the start.

This is true, but I would also argue that M1 is a direct application of the mobile technology. Basically, it is a relative low effort, conservative design. Going forward, I expect there to be a larger difference between the mobile and the desktop (Mac) chips.

 

[leman]

I am not sure you and lemon are talking the same language , can you specify what you mean by M1 different uarch then the pros ? Lemon is talking about the cpu generation , are you also referring to it ? Or to something else ? Like how the interconnect works , memory subsystem and such ? I don’t think anyone that understand cpu design think you can just “slap more cpu and GPU” and call it a day , and it doesn’t seems to be lemon point , maybe if you can clarify here , you might even reach an agreement!

The way I understand their claim is that prosumer hardware will be based on the same technology (Firestorm/Icestorm/G13), just scaled up to contain more cores, more RAM and more I/O.

I instead believe that the prosumer hardware will be based on a next-generation Apple Silicon that is specifically designed for high-performance prosumer desktop applications. See my post above for my reasoning.

What bugs me a bit is that Lemon Olive appears to be claiming that their account is supported by the recent leaks. It is most certainly not, as no leakers actually mentioned the microarchitecture these chips are going to based upon. Of course, it is entirely possible that they are right and I am wrong, but we will only know that after the product is actually officially announced. Right now, all we can do is speculate. Making authoritative claims and trying to shut down people that have different opinions with condescending remarks is not a way to carry out a conversation.

 

[NotTooLate]

The way I understand their claim is that prosumer hardware will be based on the same technology (Firestorm/Icestorm/G13), just scaled up to contain more cores, more RAM and more I/O.

I instead believe that the prosumer hardware will be based on a next-generation Apple Silicon that is specifically designed for high-performance prosumer desktop applications. See my post above for my reasoning.

What bugs me a bit is that Lemon Olive appears to be claiming that their account is supported by the recent leaks. It is most certainly not, as no leakers actually mentioned the microarchitecture these chips are going to based upon. Of course, it is entirely possible that they are right and I am wrong, but we will only know that after the product is actually officially announced. Right now, all we can do is speculate. Making authoritative claims and trying to shut down people that have different opinions with condescending remarks is not a way to carry out a conversation.

Can you clarify your statement a bit more (i think i am missing your reasoning still) , your premise is that Apple will use a new core i.e next generation (i.e A15) and it would be for the prosumer high performance applications , but i think you missed the part where this new core goes into an iPhone , how does this sit with your speculation of "specifically designed for high-performance prosumer desktop applications"

Do you believe Apple will design a new core (notice i mention core and not a SoC) lineup all together that will only be in the Pro machines ? this seems like an

Also , there is NOTHING simple to "adding more cores more RAM more IO more GPU" , we can discuss this topic in a different thread probably.

Lastly it is not a well thought out strategy to tapeout all of your new uarch silicon at the same time , this to address your comment about M1 introduced 6 months ago , it creates a bottleneck on your backend and frontend designers , it creates a bottleneck in post silicon validation , it creates an overhead of fixing bugs for the common IP`s , it is even worse when moving process nodes , it would be smarter when moving to a new arch to have a silicon you can validate for obvious bugs first and then go ahead with the rest of the lineup (this includes doing Analog characterization which is never easy for new nodes).

Also it seems this is how AMD and Intel does it , AMD went server first with Zen3 and then desktop and lastly mobile and APU`s - this usually depends on the market landscape I guess.

But of course you might be correct , until apple announces something we wont know , cant wait!!!

 

[CWallace]

This is true, but I would also argue that M1 is a direct application of the mobile technology. Basically, it is a relative low effort, conservative design.

It's based on a mobile SoC so of course it's a direct application of mobile technology.

And a pretty darn impressive one considering how it can keep up with - and in many cases, humble - non-mobile chips that draw significantly more power and generate significantly more heat. And yes, a not-insignificant part of that is Apple being able to optimize the M SoC to the specifics of how macOS operates where the Intel/AMD CPUs are more general-purpose to accommodate more OSes, but it still is a benefit.

Intel certainly was unable to do the opposite - successfully make a direct application of their non-mobile technology that worked with the mobile market (*cough* Atom *cough*).

Going forward, I expect there to be a larger difference between the mobile and the desktop (Mac) chips.

I expect A-series and M-series SoCs will continue to leverage the same basic foundation, with the differences being primarily the number of CPU and GPU cores and the physical interface they connect to the outside world with (Lightning vs. USB/TB).

 

[Lemon Olive]

I agree with this 100%. That's why I find it rather ironic that your argument so far boils down to "I am right and you are wrong". We both have equal rights to our conjectures, and we should be able to discuss our reasoning as rational people.

Except, no. This perspective that the internet has created where "everyone's opinion has the same weight" is nonsense. There is right, and there is wrong. As of right now, there is only 1 accurate take on what is most probably going to happen. It is not 50/50. That isn't reality.

- More than half a year has passed since M1 release, with no M1X-style chips in sight (the longer the delay, the less sense would such a chip make)

No products have been released yet that needed more than the M1. "no M1X style chips in sight" is patently false. They existed a year ago, and were reported on, and at that time said to be in products for late 2021. Long before the A15 development was even underway. So, false.

- Firestorm in M1 peaks at 3.2 ghz which is plenty for the entry-level hardware but not enough to take the performance crown for the prosumer hardware, it is questionable whether Firestorm can be stable at higher clocks

This is more nonsense speak. The performance of Firestorm in M1 speaks for itself. The M1X has 8 performance cores and 2 efficiency cores. You can do the math and understand that this chip will outperform nearly everything Apple has from Intel save for the highest end Mac Pros.

- Apple introduced ray tracing support in Metal last year, heavily hinting that hardware RT support is coming — and that it's coming this year. Apple G13 does not support hardware RT, ergo it is likely that the prosumer hardware will require a new GPU architecture

There are lot of "maybes" and assumptions in this thought process, too many to pick apart. Also, zero rumors or leaks to back up these maybes.

- Prosumer hardware is a likely candidate to support advanced vector extensions (SVE2). It would be odd if the first generation would not have that support, ergo it is likely that the prosumer hardware will require a new CPU architecture

No, this is simply guess work and is not expected for the upcoming MacBook Pro at all. Not supported by anything.

- The leaks point to a 8+2 core configuration. All Firestorm/Icestorm chips so far feature 4 Icestorm cores. This again suggests that the prosumer hardware will use a different microarchitecture, with redesigned, more capable low-power cores

This is the ONLY thing that has supported the possibility of new chip architecture, and yet is not guaranteed at all. Apple can (and more than likely is) simply adding fewer efficiency cores alongside twice as many performance cores in the newly designed Pro product that needs power more than it needs efficiency, and also likely has updated battery design to account for it.

Of all your points, you have only offered 1 that is based off any actual information that is available. You also haven't tackled any of the other points that overwhelmingly point away from your conclusion.

 

[Serban55]

Again, those who think Apple will offer the same arhitecture as the iphones for the mac...must be outside microarchitecture/ SoC knowledge
Apple will scale a lot faster the mac sillicon than the mobile platform...soon the iphone will get the lowest SoC from apple silicon family. Physics never lie
the next 14"/16" will have probably 4x8gb Ram on each side of the sillicon for a total of 64gb ram with a unified gpu cores that probably will be a second die
M1 just showed us how much more powerful and efficient the iphone/ipad SoC (after 10 years under Johny brain) is compared to intel/amd parallel segment

 

[Serban55]

Again someone here i wonder when he refers to the same "m1 architecture" if he refers that apple will still use arm architecture and not x86 ?! because otherwise, well....look my previous post.

 

[Serban55]

This is true, but I would also argue that M1 is a direct application of the mobile technology. Basically, it is a relative low effort, conservative design. Going forward, I expect there to be a larger difference between the mobile and the desktop (Mac) chips.

of course, some people are in denial...to think that Apple will offer the same SoC/architecture for iphones and ipads as they will for the bigger macbook pro or bigger imac, and not even talk about mac pro...is strange at least. While Apple showed us how far ahead an iphone/ipad SoC come along (compared to the intel, amd x86 chips)...at the same time, soon enough, i believe the iphone and ipad will start having the weakest SoC from Apple silicon family (excluding the wearables etc)

 

[leman]

Can you clarify your statement a bit more (i think i am missing your reasoning still) , your premise is that Apple will use a new core i.e next generation (i.e A15) and it would be for the prosumer high performance applications , but i think you missed the part where this new core goes into an iPhone , how does this sit with your speculation of "specifically designed for high-performance prosumer desktop applications"

I did not claim that they would use the A15, I just don't think they will use Firestorm/Icestorm/G13 combination. You see, I believe that people are getting a bit too much carried away in discussing names for the chips or insisting on strong chronology (M2 must come after A15 etc.) instead of actually looking on what is possible, what makes sense and what Apple needs right now.

My speculation is that Apple will use an updated microarchitecture for the prosumer silicon, not more and no less. I doubt that this microarchitecture will be identical to what M1 is using and I am not sure whether it is going to be the same as A15. It would not surprise me if it is a completely different core (albeit related to A15).

Do you believe Apple will design a new core (notice i mention core and not a SoC) lineup all together that will only be in the Pro machines ? this seems like an

I assume that you wanted to write something like "it seems like a cumbersome thing to do". Well, I agree but then again, Apple is a very opinionated company that wants to be the best. And it would not be unreasonable for them to have different designs for entry-level and prosumer level. ARM has a bunch of fundamentally different core designs for example, and they are much smaller than Apple.

Again, this is just my personal speculation, but I think we will see two different families of microarchitectures in Macs going forward. The one will basically scale up the mobile chip (just like M1 did) and be used in entry-level models. The other one will be a tweaked microarchitecture to be used in prosumer Macs.

Also , there is NOTHING simple to "adding more cores more RAM more IO more GPU" , we can discuss this topic in a different thread probably.

Very true, but that is just a question of time and resources.

Lastly it is not a well thought out strategy to tapeout all of your new uarch silicon at the same time , this to address your comment about M1 introduced 6 months ago , it creates a bottleneck on your backend and frontend designers , it creates a bottleneck in post silicon validation , it creates an overhead of fixing bugs for the common IP`s , it is even worse when moving process nodes , it would be smarter when moving to a new arch to have a silicon you can validate for obvious bugs first and then go ahead with the rest of the lineup (this includes doing Analog characterization which is never easy for new nodes).

Well, Apple was able to ship the A14 and the M1 within two months from each other, so I suppose they have the resources and the workflow to do this. Also, Jade Die has been reported to started production as of last month. I definitely think it is realistic for Apple to get out a new chip every half a year.

Except, no. This perspective that the internet has created where "everyone's opinion has the same weight" is nonsense. There is right, and there is wrong. As of right now, there is only 1 accurate take on what is most probably going to happen. It is not 50/50. That isn't reality.

We are speculating about a product that is not yet released. There are no 'rights' here. Anything we talk about only becomes right or wrong after Apple actually releases the thing.

No products have been released yet that needed more than the M1. "no M1X style chips in sight" is patently false. They existed a year ago, and were reported on, and at that time said to be in products for late 2021. Long before the A15 development was even underway. So, false.

It was reported that Apple is working on more powerful prosumer chips. There is not a single reliable report about which microarchitecture that chip will use. The only thing we know which are are more or less certain are the codenames — Jade Die and Lifuka (for the GPU) — and the core configurations.

This is more nonsense speak. The performance of Firestorm in M1 speaks for itself. The M1X has 8 performance cores and 2 efficiency cores. You can do the math and understand that this chip will outperform nearly everything Apple has from Intel save for the highest end Mac Pros.

You don't have to preach this to me. I spent as much time as anyone else here making a case for strong Firestorm performance and I was predicting that Apple will eventually overtake Intel more than two years ago, when most people though that very notion was ridiculous.

And yet, Firestorm alone is not enough to take the performance crown. Apple wants to be the best. They have hands-down the best low-power silicon, and they will want to have hands-down the best prosumer silicon. They need around 20-25% higher single-threaded performance to take a dominating, undisputed lead.

No, this is simply guess work and is not expected for the upcoming MacBook Pro at all. Not supported by anything.

Of course all I write here is guesswork. How are your claims any better supported though? Your argument basically boils down to "it will be an M1X because iPhone 13 is not out yet". At least I am trying to provide some technical reasoning, you don't provide anything except the claim that Apple has to do things in some very particular order that you have decided upon.

Again, there have been zero leaks about the microarchitecture of these upcoming chips or there branding. "M1X", "M2" — all of these are inventions by people who are reporting on these leaks.

 

[CWallace]

Again, this is just my personal speculation, but I think we will see two different families of microarchitectures in Macs going forward. The one will basically scale up the mobile chip (just like M1 did) and be used in entry-level models. The other one will be a tweaked microarchitecture to be used in prosumer Macs.

Anything is possible, but to me, this sounds like complication where the goal of Apple Silicon as a family appears to be one of simplification by using a common foundation and tweaking it for the final product it is in.

 

[Serban55]

Anything is possible, but to me, this sounds like complication where the goal of Apple Silicon as a family appears to be one of simplification by using a common foundation and tweaking it for the final product it is in.

For Apple is no complication...they are just derivations...you cant have the same architecture as you have in an 13" ipad pro/iphone as you do in an mac pro.
The nice things is that while on x86 architectures the xeon levels are on 10nm/14nm for the mac pro lets say...thanks to its own apple silicon the mac pro will have around the same level as the rest of the family...so expect next year the mac pro silicon to be already on 5nm/4nm. So, this is why is about simplification and common foundation

 

[leman]

Anything is possible, but to me, this sounds like complication where the goal of Apple Silicon as a family appears to be one of simplification by using a common foundation and tweaking it for the final product it is in.

There is no doubt that they will use a common foundation, my speculation is simply that "tweaking" will involve more than just adjusting the clocks and cache sizes. I think we are going to see some more significant internal differences between different families that would allow the prosumer silicon to achieve higher single-threaded performance. But again, it's just a conjecture. I will be happy as long as my new 16" can outperform a xeon-based desktop workstation

 

[leman]

For Apple is no complication...they are just derivations...you cant have the same architecture as you have in an 13" ipad pro/iphone as you do in an mac pro.

Why not though? I mean, if all Apple was after is multi-core performance, things would be easy. They already have the best performing core under 5 watts. All they need to do is slap several dozens of these together and all the Xeons and EPYCs can "nervously smoke in the corner" (as we used to say in my home country).

But Apple doesn't just want to have the best multi-core processors. They want to have the best processors, period. And that is much more complicated.

 

[Serban55]

Why not though? I mean, if all Apple was after is multi-core performance, things would be easy. They already have the best performing core under 5 watts. All they need to do is slap several dozens of these together and all the Xeons and EPYCs can "nervously smoke in the corner" (as we used to say in my home country).

But Apple doesn't just want to have the best multi-core processors. They want to have the best processors, period. And that is much more complicated.

you answeared to your own question already, because your last answer would have been mine as well

 

[Tagbert]

... there will be the MacBook (Air) to anchor "the low end" and they can add a fan to it to keep it cool ...

No fans! ?

One of the nicest features of the Air is that it doesn't have a fan. I would be fully willing to accept the slight limit in performance of heat related throttling - if I ever even hit that, which I haven't.