M2 MacBook Pro

[senttoschool]

Only the changes to accommodate new MMU and bigger cache, it basically performs like A14 with higher frequency, which means inner workings of the architecture remained pretty much unchanged.

Most certainly not true. There are big differences in the efficiency cores. Hence, the MT performance of the A15 went up higher than the ST. This also allows better battery life because iOS could then use the efficiency cores for things that it used to require performance cores.

The efficiency cores of the A15 have also seen massive gains, this time around with Apple mostly investing them back into performance, with the new cores showcasing +23-28% absolute performance improvements, something that isn’t easily identified by popular benchmarking.

AnandTech Forums: Technology, Hardware, Software, and Deals

Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

www.anandtech.com

 

[MayaUser]

And also the gpu core is better on the A15

 

[CWallace]

I'm hopeful that we're going to see M2 Pro/Max. I think a 2 year update cadence is too slow to compete with Intel and AMD in terms of raw performance.

I don't think Apple cares about matching Intel and AMD on synthetic benchmarks.

A two-year release cadence for the "pro" chips makes sense from a financial aspect in that it allows for a decent return on the R&D investment considering sales of MacBook Pros, iMac Pros and Mac Pros. Volumes are likely high enough for the MacBook Air, iMac and iPad Pro to justify an annual release cadence for the "baseline" chip.

 

[Kpjoslee]

Most certainly not true. There are big differences in the efficiency cores. Hence, the MT performance of the A15 went up higher than the ST. This also allows better battery life because iOS could then use the efficiency cores for things that it used to require performance cores.

I think I admitted there were changes on little cores but I still went by the statement that there weren't any significant changes on big cores other than the frequency increases.

M1 to M2 variants will probably focus on GPU improvements rather than CPU improvement which is going to be pretty minor especially for M2 pro/max since they only have 2 efficiency cores(unless they bring up the e-core count from 2 to 4).

 

[leman]

A 20-core GPU option for the base M2. That's very interesting. I wonder if that means we will see something like the 16" Macbook Air which will be equipped with the 20-core GPU.

The M2 Duo might be an option for the iMac and Mac Mini as Apple embraces the multi-chip package technology.

 

[leman]

I don't think Apple cares about matching Intel and AMD on synthetic benchmarks.

A two-year release cadence for the "pro" chips makes sense from a financial aspect in that it allows for a decent return on the R&D investment considering sales of MacBook Pros, iMac Pros and Mac Pros. Volumes are likely high enough for the MacBook Air, iMac and iPad Pro to justify an annual release cadence for the "baseline" chip.

It is possible that Apple has some tech which would allow them to build these pro-chip “cheaply”. If scaling up the consumer silicon can be largerly automated (and it does seem that Apple is going heavy on replicatable functional units), updating lower-volume devices might just end up economically feasible.

 

[thenewperson]

Would an M2 Duo be more efficient (or just better/easier in some way) than a hypothetical M2 Pro? Would they just switch to doing that instead of MX Pro chips down the line?

 

[leman]

Would an M2 Duo be more efficient (or just better/easier in some way) than a hypothetical M2 Pro? Would they just switch to doing that instead of MX Pro chips down the line?

Multi-chip packaging is probably cheaper in the long run, as one can reuse the chips across more configurations. I doubt it’s more efficient though, I’d expect a single die solution to use less power overall.

 

[senttoschool]

I don't think Apple cares about matching Intel and AMD on synthetic benchmarks.

I always hear this but this is most certainly not true. People choose between Windows/Macs all the time. There are more iPhone + Window users than iPhone + Mac users.

Apple has to beat AMD/Intel convincingly in order to take market share from Windows. Right now, Mac market share is tiny. Apple is the underdog and it needs to keep pushing.

While "perf/watt" is generally the best metric for laptops, raw performance is often easier to sell to regular people. It's easier to understand that this chip is 50% faster than that chip instead of this chip is 50% more efficient than that chip.

There's also the halo product factor. When the M1 first came out, it took the ST crown. On year later, Intel has regained it with Rocket Lake and will probably surpass it with Alder Lake.

 

[Kpjoslee]

The M2 Duo might be an option for the iMac and Mac Mini as Apple embraces the multi-chip package technology.

I really don't see the point of having M2 "duo" when Pro/Max already exists.

 

[leman]

I really don't see the point of having M2 "duo" when Pro/Max already exists.

Not today, no, but it might make sense for future products. Especially considering yields/production costs. A two-chip module for consumer desktops might be a more economically viable solution than a M Pro series chip. We still don't really know what Apple is planning.

 

[Kpjoslee]

Not today, no, but it might make sense for future products. Especially considering yields/production costs. A two-chip module for consumer desktops might be a more economically viable solution than a M Pro series chip. We still don't really know what Apple is planning.

I don't disagree that Apple will use multi-chip SoC on future products, but I just don't see M2 "duo" happening.
Since M2 is most likely be fabricated on N4 process which is just a refined version of their N5/N5P, I don't see why they would have to shift to using 2 M2 chips instead of Pro/max series considering they already getting good enough yield/volume of M1 Pro/Max to be used on Macbook Pro lineup.

 

[MrGunnyPT]

M2 and the "entry" line up will stay the efficient products with the best battery life without a doubt. I own the 14" with the M1 Pro full die and I gotta say this is already a battery champ and it's close to my Air base model.

 

[senttoschool]

I don't disagree that Apple will use multi-chip SoC on future products, but I just don't see M2 "duo" happening.
Since M2 is most likely be fabricated on N4 process which is just a refined version of their N5/N5P, I don't see why they would have to shift to using 2 M2 chips instead of Pro/max series considering they already getting good enough yield/volume of M1 Pro/Max to be used on Macbook Pro lineup.

A "M2 Duo" would be 8/8 CPU cores for a total of 16. That sounds more like the base M2 Pro to me.

I agree it wouldn't make sense to make a 16 CPU-core, 20 GPU core SoC that has two dies glued together just for the Macbook Air, Mac Mini, iMac, iPad Pro. That seems way overkill.

Hence, I belive "M2 Duo" is simply just an M2 Pro.

Now, it's entirely possible that Apple could go the tiles way. Basically, Apple would manufacture small dies, then glue them together. That would increase yields since M series have a lot more transistor than the average Intel/AMD chip.

 

[jdb8167]

Hence, I belive "M2 Duo" is simply just an M2 Pro.

Unless Apple has no intention of making an M2 Pro then they need a different name to keep the marketing clear. Right now the rumor mill is talking about M3 Pro and Max not M2.

 

[senttoschool]

Unless Apple has no intention of making an M2 Pro then they need a different name to keep the marketing clear. Right now the rumor mill is talking about M3 Pro and Max not M2.

Not convinced that Apple is skipping the M2 Pro/Max.

The A14 cores in the M1 Pro/Max will have to last 2.5 year before the next update? Seems too long.

 

[jdb8167]

Not convinced that Apple is skipping the M2 Pro/Max.

The A14 cores in the M1 Pro/Max will have to last 2.5 year before the next update? Seems too long.

Why do you say 2.5 years? A16 cores in Fall 2022. M3 Pro in Spring 2023. 18 months.

 

[CWallace]

I always hear this but this is most certainly not true. People choose between Windows/Macs all the time. There are more iPhone + Window users than iPhone + Mac users.

Sure, but that purchasing decision is driven by both the need/desire to run Windows instead of macOS (for work, for gaming, etc.) and the bulk of non-enterprise PC sales are sub-$1000 and that is not a market Apple participates in.

Apple has to beat AMD/Intel convincingly in order to take market share from Windows. Right now, Mac market share is tiny. Apple is the underdog and it needs to keep pushing.

Apple is not concerned about significantly growing PC market share just as they are not concerned about significantly growing smartphone market share.

The Mac division brings in enough revenue every year that if it was a separate company it would be in to Forbes 400. It also makes far more profit than the PC divisions of Dell, HPE, and Lenovo.

While "perf/watt" is generally the best metric for laptops, raw performance is often easier to sell to regular people. It's easier to understand that this chip is 50% faster than that chip instead of this chip is 50% more efficient than that chip.

"Regular people" tend to buy on price over everything else. They will buy a $400 Intel/AMD "slab" laptop over a $1000 MacBook Air no matter how much faster the Air is at the tasks they do. Time is not money to that market. Money is money to that market.

There's also the halo product factor. When the M1 first came out, it took the ST crown. On year later, Intel has regained it with Rocket Lake and will probably surpass it with Alder Lake.

And yet the M1 did not double Mac sales, much less triple them. They did drive them higher, to be sure, and did drive them higher than having the actually latest generation Intel CPU did (in part because the M1 was faster than the actual latest generation Intel CPU).

Alder Lake needs to be run at power and heat levels similar to a nuclear reactor core to get better-than-Apple Silicon performance out of it. If you think putting one of those in an iMac 5K case is going to drive sales more than an M1 Pro or M1 MAX would... I strongly disagree.

 

[dgdosen]

I don't disagree that Apple will use multi-chip SoC on future products, but I just don't see M2 "duo" happening.
Since M2 is most likely be fabricated on N4 process which is just a refined version of their N5/N5P, I don't see why they would have to shift to using 2 M2 chips instead of Pro/max series considering they already getting good enough yield/volume of M1 Pro/Max to be used on Macbook Pro lineup.

Prefacing this noting I'm a layman with respect to chip manufacturing:

If I were Apple, I'd be highly motivated to build SOCs using some fabric/chiplet design to combine multiples of silicon together as opposed to building bigger SOCs (like M1Pro -> M1Max). Think of the flexibility to be able to use the same silicon in an iPadPro and entry level mac; but then use multiples of that same silicon in more powerful laptops/desktops.

If TSMC was trying to prove their 3D-Fabric works, wouldn't their easiest path to success be to work with Apple on figuring out a way to combine two of Apples simpler designs (M1/M2 SOC) together? M2 Duo...

This all means that we can expect to see a fan-less, 2lb, 12" Macbook as powerful as a current M1Pro next summer.

 

[leman]

This all means that we can expect to see a fan-less, 2lb, 12" Macbook as powerful as a current M1Pro next summer.

Hardly… if M2 Duo is a thing at all, it will be reserved for larger machines.

 

[senttoschool]

Sure, but that purchasing decision is driven by both the need/desire to run Windows instead of macOS (for work, for gaming, etc.) and the bulk of non-enterprise PC sales are sub-$1000 and that is not a market Apple participates in.

For the vast majority of people, normal office workers, they can use Windows or MacOS. Only specialized workers need one or the other. Hence, the decision usually comes down to price, which you correctly asserted.

Which is exactly why I predict a Macbook SE:

Do you think Apple will make a Macbook SE? And at what price point?

Laying down some facts: Apple makes the iPhone SE ($400), Watch SE ($280), and iPad (SE in spirit for $330) The iPhone 8 launched at $699. The iPhone SE, based on the iPhone 8 but with a faster SoC, is $300 cheaper. Cheapest Macbook Air right now is $1000 Apple once sold an Intel Macbook Air...

forums.macrumors.com

Apple is not concerned about significantly growing PC market share just as they are not concerned about significantly growing smartphone market share.

The Mac division brings in enough revenue every year that if it was a separate company it would be in to Forbes 400. It also makes far more profit than the PC divisions of Dell, HPE, and Lenovo.

They're concerned. That's why they have the iPhone SE, Watch SE, $330 iPad, and I predict a future Macbook SE.

This isn't the old Apple. The new Apple is about selling hardware and services. In order to sell services, they need market share.

In the US, Apple owns more than 50% of the phone, tablet, wireless headset, and smart watch markets. For Macs, they only own 15% of so. See the problem? Now they have an opportunity to capture 50% of computer markets just like they did for other markets.

"Regular people" tend to buy on price over everything else. They will buy a $400 Intel/AMD "slab" laptop over a $1000 MacBook Air no matter how much faster the Air is at the tasks they do. Time is not money to that market. Money is money to that market.

Correct. Hence, why I think we will see a Macbook SE.

And yet the M1 did not double Mac sales, much less triple them. They did drive them higher, to be sure, and did drive them higher than having the actually latest generation Intel CPU did (in part because the M1 was faster than the actual latest generation Intel CPU).

Alder Lake needs to be run at power and heat levels similar to a nuclear reactor core to get better-than-Apple Silicon performance out of it. If you think putting one of those in an iMac 5K case is going to drive sales more than an M1 Pro or M1 MAX would... I strongly disagree.

Don't underestimate the effect of the halo product. Dumb people (non-geeks) won't know much about efficiency. They walk into a store thinking Intel machines are "faster" despite the obvious difference in perf/watt.

 

[CWallace]

I'd argue that the MacBook Air has served as the "MacBook SE" since the 2011 refresh and the price drop to $999. And we've heard that it is the most popular model of Mac due to that $999 price.

Current Average Sales Price of PC laptops is $699, so Apple would need to drop the price of the M1 MacBook Air by $300. That is a big ask for Apple because the current Air probably has a production cost of around $700 so they would have to significantly de-content the machine to hit that price and maintain a decent margin - something like the
2017 non-Retina model with an A10X...

 

[senttoschool]

I'd argue that the MacBook Air has served as the "MacBook SE" since the 2011 refresh and the price drop to $999. And we've heard that it is the most popular model of Mac due to that $999 price.

Current Average Sales Price of PC laptops is $699, so Apple would need to drop the price of the M1 MacBook Air by $300. That is a big ask for Apple because the current Air probably has a production cost of around $700 so they would have to significantly de-content the machine to hit that price and maintain a decent margin - something like the
2017 non-Retina model with an A10X...

Amazon just had the Macbook Air for $799.

Black Friday MacBook Air deals 2024: all the top offers still live

Black Friday MacBook Air deals are still live heading into the weekend

www.techradar.com

That suggests $750 - $800 price for a Macbook SE is still profitable for Apple. This laptop would be designed to take market share anyways and these future users will upgrade to a more expensive Mac. The key is to bring these users into the Mac world.

In 1-2 years, reusing the 13" Macbook Air chassis for an SE will be extremely economical as the Air moves to a new design and 14". All the assembly lines don't have to be wasted.

Put it this way, the iPad Air has better tech than the Macbook Air right now. It has a better display, touch, magnetic connector, better cameras, more sensors, and it's thinner. It costs less than the Macbook Air.

Lastly, Apple saves a few hundred dollars in BOM by manufacturing its own SoC instead of buying from Intel: https://www.trendforce.com/presscenter/news/20200707-10377.html

 

[Boil]

In 1-2 years, reusing the 13" Macbook Air chassis for an SE will be extremely economical as the Air moves to a new design and 14". All the assembly lines don't have to be wasted.

Apple needs all new chassis designs for all Apple silicon products...

"All the assembly lines don't have to be wasted."

That makes it sound like the 13" MacBook Air production line is going to just sit there, shut down forever; more like Apple would have that particular production line re-vamped to start producing the new 2022 M2-powered MacBooks (I believe the "Air" suffix needs to go)...

Put it this way, the iPad Air has better tech than the Macbook Air right now. It has a better display, touch, magnetic connector, better cameras, more sensors, and it's thinner. It costs less than the Macbook Air.

I would think Apple would rather folks seeking a lower cost laptop might consider an iPad; then down the road when they decide to move up to a real laptop, they would consider a slightly more expensive MacBook (since they are already in the Apple ecosystem)...?

Lastly, Apple saves a few hundred dollars in BOM by manufacturing its own SoC instead of buying from Intel: https://www.trendforce.com/presscenter/news/20200707-10377.html

Just because the BOM is lower does not mean Apple intends to pass those savings on to the consumer...!

 

[senttoschool]

Apple needs all new chassis designs for all Apple silicon products...

"All the assembly lines don't have to be wasted."

That makes it sound like the 13" MacBook Air production line is going to just sit there, shut down forever; more like Apple would have that particular production line re-vamped to start producing the new 2022 M2-powered MacBooks (I believe the "Air" suffix needs to go)...

A Macbook SE does not need a new design. The original iPhone SE used iPhone 5S design. The current SE uses the iPhone 8 design.

A future Macbook SE does not need a new design. It can/probably will use the current 13" Air design as the Air moves to 14".

All the assembly line, designs, materials, logistics are all there to make a Macbook SE today. Apple just needs to put an M2/M3 SoC in it. That's it.

I would think Apple would rather folks seeking a lower cost laptop might consider an iPad; then down the road when they decide to move up to a real laptop, they would consider a slightly more expensive MacBook (since they are already in the Apple ecosystem)...?

Yes, the Macbook SE would still be more expensive than the iPad Air. I predict a price of $750 - $800 for a Macbook SE. That's more than an Air but less than an iPad Pro, which makes sense.

Just because the BOM is lower does not mean Apple intends to pass those savings on to the consumer...!

No, Apple does not. But it does allow them to comfortably create a Macbook SE to win marketshare from Windows.