Discussion Apple Silicon SoC thread

[Eug]

M1
5 nm
Unified memory architecture - LP-DDR4
16 billion transistors

8-core CPU

4 high-performance cores
192 KB instruction cache
128 KB data cache
Shared 12 MB L2 cache

4 high-efficiency cores
128 KB instruction cache
64 KB data cache
Shared 4 MB L2 cache
(Apple claims the 4 high-effiency cores alone perform like a dual-core Intel MacBook Air)

8-core iGPU (but there is a 7-core variant, likely with one inactive core)
128 execution units
Up to 24576 concurrent threads
2.6 Teraflops
82 Gigatexels/s
41 gigapixels/s

16-core neural engine
Secure Enclave
USB 4

Products:
$999 ($899 edu) 13" MacBook Air (fanless) - 18 hour video playback battery life
$699 Mac mini (with fan)
$1299 ($1199 edu) 13" MacBook Pro (with fan) - 20 hour video playback battery life

Memory options 8 GB and 16 GB. No 32 GB option (unless you go Intel).

It should be noted that the M1 chip in these three Macs is the same (aside from GPU core number). Basically, Apple is taking the same approach which these chips as they do the iPhones and iPads. Just one SKU (excluding the X variants), which is the same across all iDevices (aside from maybe slight clock speed differences occasionally).

EDIT:



M1 Pro 8-core CPU (6+2), 14-core GPU
M1 Pro 10-core CPU (8+2), 14-core GPU
M1 Pro 10-core CPU (8+2), 16-core GPU
M1 Max 10-core CPU (8+2), 24-core GPU
M1 Max 10-core CPU (8+2), 32-core GPU

M1 Pro and M1 Max discussion here:

Page 78 - Discussion - Apple Silicon SoC thread

Page 78 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M1 Ultra discussion here:

Page 109 - Discussion - Apple Silicon SoC thread

Page 109 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M2 discussion here:

Page 127 - Discussion - Apple Silicon SoC thread

Page 127 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

Second Generation 5 nm
Unified memory architecture - LPDDR5, up to 24 GB and 100 GB/s
20 billion transistors

8-core CPU

4 high-performance cores
192 KB instruction cache
128 KB data cache
Shared 16 MB L2 cache

4 high-efficiency cores
128 KB instruction cache
64 KB data cache
Shared 4 MB L2 cache

10-core iGPU (but there is an 8-core variant)
3.6 Teraflops

16-core neural engine
Secure Enclave
USB 4

Hardware acceleration for 8K h.264, h.264, ProRes

M3 Family discussion here:

Page 215 - Discussion - Apple Silicon SoC thread

Page 215 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

M4 Family discussion here:

Page 263 - Discussion - Apple Silicon SoC thread

Page 263 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

 

[Doug S]

Is a vapor chamber a slam dunk?
If we consider the generic concept of using a phase change to absorb heat, you obviously also have solid-liquid transitions, but also even solid-solid transitions.
While (usually, but not always) liquid to gas has higher latent heat than solid to liquid or solid to solid, it also comes with the hassles of (again usually) a substantial change in volume.

I know nothing about how this space is playing out in other phones or even on PCs, but I could imagine Apple prioritizing other issues than raw "amount of heat that can be absorbed" leading to their use of a phase change material but not specifically a vapor chamber.

It has been a persistent rumor I've seen reported for a while now, and other phones have already been using it. The device still has the same amount of heat to radiate to its environment so it isn't helping with actual cooling, just the perception of cooling since you can feel the hot spots in your hands. It might increase the amount of time it takes for a CPU/GPU to get hot enough it needs to throttle by moving heat from the SoC to cooler parts of the device, but it still will.

If Apple thinks it has value in the iPhone, you'd think they'd be considering it for Macbook Air as well. The MBA's case has far more surface area over which to spread the heat. It still wouldn't cool as well as the MBP with its fan, but it would close the gap between the two. Of course that assumes Apple WANTS to close the gap, rather than relying on "same SoC but less efficient cooling" as a way of segmenting the product lines vs how a PC OEM would do it by giving the less expensive model a slower binned CPU.

 

[Doug S]

That’s a lot of money!

I still remember before N3 came out and everyone was reporting it was going to be $25K per wafer, which turned out to be much higher than the actual price. I'm skeptical about that $30K number for N2 until we see quarterly results from TSMC that reflect 50%+ higher per wafer revenue for N2.

Most of the layers were already made in EUV, so EUV driven price increases are mostly played out. At least until they get into heavy multipatterning or high NA. Could the difference in the deposition/etch steps between FinFET and GAAFET really account for that much additional manufacturing cost?

Of course TSMC might be charging $30K per wafer not because their costs have gone up by 50% but because they know it is in heavy demand and they are exercising their monopoly pricing power.

 

[LightningZ71]

Remember, aside from the actual cost of the steps of production and materials consumed, they also have to amortize the development costs for the node. Each node is becoming more and more expensive to develop, and no foundry is going to voluntarily eat those costs unless they aren't competitive.

 

[Doug S]

Remember, aside from the actual cost of the steps of production and materials consumed, they also have to amortize the development costs for the node. Each node is becoming more and more expensive to develop, and no foundry is going to voluntarily eat those costs unless they aren't competitive.

Sure but that can't account for more than a couple thousand bucks per wafer. They're going to be running 100k wpm by the end of 2026. They said it will be their bigger node ever and I think N3 is over 150K wpm now. They probably ship 5 million N2 class wafers by the end of the decade. Even if it cost $10 billion to develop and they want to amortize that quickly that's $2K per wafer.

All I'm saying is I wouldn't take the $30K figure as gospel, since we have the previous experience of the $25K figure to consider.

 

[johnsonwax]

Sure but that can't account for more than a couple thousand bucks per wafer. They're going to be running 100k wpm by the end of 2026. They said it will be their bigger node ever and I think N3 is over 150K wpm now. They probably ship 5 million N2 class wafers by the end of the decade. Even if it cost $10 billion to develop and they want to amortize that quickly that's $2K per wafer.

All I'm saying is I wouldn't take the $30K figure as gospel, since we have the previous experience of the $25K figure to consider.

Probably safe to say that anyone guaranteeing purchase volume with these sizable prepayment agreements aren't paying full price. TSMC isn't going to charge less than the marginal cost of a wafer + margin, and they need to charge more than that to recoup their fixed costs which are substantial. Prepayments go a LONG way to securing that those fixed costs will be covered. I'd take the $30K price as what someone not guaranteeing what size check they are going to write needs to pay. That is, you negotiate down from $30K depending on how many wafers you guarantee you're going to buy. And then as the fixed costs are covered (plus operating, etc.) TSMC can lower that $30K toward the marginal cost, which itself is likely declining as you get productivity gains in wafer production, etc. It will be interesting to see how tariffs impact the AZ costs over Taiwan costs on slightly older nodes.

 

[Mopetar]

Wake me up when M4 Max 36GB laptop becomes available for $2500 and Windows on ARM achieves 90% Steam library compatibility.

You might want to change your name to Igor_van_winkle in that case.

I just don't think FLIR is desired by nearly enough people to justify making it a standard part of a mass market phone.

You're probably right in that it's not really a big deal, but there's a long history of technology purchases over far more silly gimmicks. If nothing else it makes for another stupid photo filter. Steve Jobs would be able to get some people to buy two.

The actual bigger deal as you also alluded to would be Apple managing to keep it a secret as well as they have. There may be a few people who would be legitimately excited if the new phones did have it, but will anyone be disappointed if they don't have it? Not really.

 

[name99]

It has been a persistent rumor I've seen reported for a while now, and other phones have already been using it. The device still has the same amount of heat to radiate to its environment so it isn't helping with actual cooling, just the perception of cooling since you can feel the hot spots in your hands. It might increase the amount of time it takes for a CPU/GPU to get hot enough it needs to throttle by moving heat from the SoC to cooler parts of the device, but it still will.

If Apple thinks it has value in the iPhone, you'd think they'd be considering it for Macbook Air as well. The MBA's case has far more surface area over which to spread the heat. It still wouldn't cool as well as the MBP with its fan, but it would close the gap between the two. Of course that assumes Apple WANTS to close the gap, rather than relying on "same SoC but less efficient cooling" as a way of segmenting the product lines vs how a PC OEM would do it by giving the less expensive model a slower binned CPU.

I think you missed my point.
I'm not discussing the existence of a way to boost thermal inertia, I'm discussing the technical specifics of the technology that might be used.

For example we have patents like this
https://patents.google.com/patent/US10714425B2
which imagine more or less filling unused volume within a phone with a phase change material. The scheme they have in mind looks like it only makes sense if the phase change is from one to another crustal structure (ie solid to solid).

The patent is almost ten years old. Which may mean it's been abandoned as unworkable? Or maybe it means that someone had the idea, and for a few years they then worked with chemical companies to put together something that had all the properties required (easily shapeable? insulating? non-toxic? etc etc)

Another way you could do this, if you could really achieve materials wizardry, would be to have the phase change material double as part of the battery... But that may be too much to hope for!

 

[johnsonwax]

I think you missed my point.
I'm not discussing the existence of a way to boost thermal inertia, I'm discussing the technical specifics of the technology that might be used.

For example we have patents like this
https://patents.google.com/patent/US10714425B2
which imagine more or less filling unused volume within a phone with a phase change material. The scheme they have in mind looks like it only makes sense if the phase change is from one to another crustal structure (ie solid to solid).

The patent is almost ten years old. Which may mean it's been abandoned as unworkable? Or maybe it means that someone had the idea, and for a few years they then worked with chemical companies to put together something that had all the properties required (easily shapeable? insulating? non-toxic? etc etc)

Another way you could do this, if you could really achieve materials wizardry, would be to have the phase change material double as part of the battery... But that may be too much to hope for!

Will note that the top F1 team is rumored to be using phase change to cool their brakes better than other teams. This is not a new idea - been around for years, but nobody did it until now - and they have effectively infinite budgets, at least the works teams do. So while Apple had a patent 10 years ago, the specific materials to do this are pretty new - with nanocomposite PCMs getting a lot of attention in the last 3 years. Basically carbon nanotube, etc. infused paraffin - paraffin is a good phase change material, quite stable during a phase cycle, but its thermal conductivity is garbage, but the nanocomposites change that, they can have excellent thermal (and/or electrical) conductivity.

Most research in nanocomposite PCMs has been in the last couple of years and that's the most likely candidate here. These materials also have a lot of latent heat - so in a bursty environment, they just soak up energy and likely never go above their phase change temp (50C is warm but not unpleasant). In a continuous environment, the benefits are less apparent because then the thermal conductivity of the material is much more important to continuously transfer heat to the sink, and none of these are going to be more conductive than just plain copper, etc. That's why they're used in F1 brakes because F1 brakes want to be at a specific temp - not above and not below, and they are very bursty on the scale of about 90s (every lap their brake pattern repeats so if you have one heavy braking corner you have 90s to do something with the excess heat). But in the case of F1 brakes, you don't want them to get colder, so the brake becomes the sink on the next straight keeping it at the right temp for the next braking section. Electronics don't work this way - you don't want the SOC to be the sink, you want to conduct that heat out of the device, and for things like laptops which have more continuous operation you're going to exhaust the latent heat capacity of the phase change material and need it to operate like a thermal conductor which is going to be worse than current options. So it may be well suited to phones and not to laptops.

 

[name99]

Will note that the top F1 team is rumored to be using phase change to cool their brakes better than other teams. This is not a new idea - been around for years, but nobody did it until now - and they have effectively infinite budgets, at least the works teams do. So while Apple had a patent 10 years ago, the specific materials to do this are pretty new - with nanocomposite PCMs getting a lot of attention in the last 3 years. Basically carbon nanotube, etc. infused paraffin - paraffin is a good phase change material, quite stable during a phase cycle, but its thermal conductivity is garbage, but the nanocomposites change that, they can have excellent thermal (and/or electrical) conductivity.

Most research in nanocomposite PCMs has been in the last couple of years and that's the most likely candidate here. These materials also have a lot of latent heat - so in a bursty environment, they just soak up energy and likely never go above their phase change temp (50C is warm but not unpleasant). In a continuous environment, the benefits are less apparent because then the thermal conductivity of the material is much more important to continuously transfer heat to the sink, and none of these are going to be more conductive than just plain copper, etc. That's why they're used in F1 brakes because F1 brakes want to be at a specific temp - not above and not below, and they are very bursty on the scale of about 90s (every lap their brake pattern repeats so if you have one heavy braking corner you have 90s to do something with the excess heat). But in the case of F1 brakes, you don't want them to get colder, so the brake becomes the sink on the next straight keeping it at the right temp for the next braking section. Electronics don't work this way - you don't want the SOC to be the sink, you want to conduct that heat out of the device, and for things like laptops which have more continuous operation you're going to exhaust the latent heat capacity of the phase change material and need it to operate like a thermal conductor which is going to be worse than current options. So it may be well suited to phones and not to laptops.

Thanks for the update.

You mean "These materials also have a large heat capacity".
Heat capacity essentially describes the heat absorbed while the temperature of the material changes.
Latent heat is the heat absorbed while the material changes phase (which is NOT accompanied by a change in temperature).

And yes, this stuff is absolutely only for burstiness
For pure conduction there is also interesting new tech. Did you see this:

https://www.servethehome.com/fabric8labs-ecam-enabled-thermal-solutions-at-hot-chips-2025/

 

[johnsonwax]

Thanks for the update.

You mean "These materials also have a large heat capacity".
Heat capacity essentially describes the heat absorbed while the temperature of the material changes.
Latent heat is the heat absorbed while the material changes phase (which is NOT accompanied by a change in temperature).

And yes, this stuff is absolutely only for burstiness
For pure conduction there is also interesting new tech. Did you see this:

https://www.servethehome.com/fabric8labs-ecam-enabled-thermal-solutions-at-hot-chips-2025/

Good to see some of that stuff getting out of the labs. We'll see if manufacturability works out. We had guys doing some work in that area in our labs and their results were fantastic, but making something work in the lab and making it work on a scale of tens of million of units, and in consumer environments is something else entirely. Not everything makes that journey.

This has always been one of Apple's bigger challenges - getting something new scaled up to iPhone scale is really hard and they often have to miss out on the first few tech rounds for that reason. See AVP relying on Sony micro-OLED that simply cannot scale up in volume to hit a consumer price point just yet. "Hey, that's amazing technology, can you make 100 million of them next year?" is kind of a big ask.

 

[mvprod123]

https://twitter.com/x/status/1964428927159382261

 

[igor_kavinski]

I wonder if it has a sensor chip to detect fart sounds.

 

[Mopetar]

I wonder if it has a sensor chip to detect fart sounds.

Of course it does. What are you using to activate Siri?

 

[poke01]

Couple more hrs.

 

[ashFTW]

Couple more hrs.

If your definition of couple is four

 

[mvprod123]

Couple more hrs.

The features of the A19 Pro will give us the first idea of the basis Apple has laid for the M5 chip family.

 

[Eug]

The features of the A19 Pro will give us the first idea of the basis Apple has laid for the M5 chip family.

I will be buying an A19 Pro soon. Truthfully though, I've been running an A14 Bionic now for 5 years, and it's still totally fine performance-wise for a phone.

Geekbench 6.5 for A14 Bionic is 2231 / 5375.

I suspect there will be a benchmark score increase of 75%-100% going from A14 Bionic to A19 Pro, considering that A18 Pro is already at 3593 / 9151. However, I'm only upgrading for the new camera.

 

[mikegg]

The features of the A19 Pro will give us the first idea of the basis Apple has laid for the M5 chip family.

Please matmul acceleration in GPU.

 

[mikegg]

Please matmul acceleration in GPU.

Got it. A19 Pro. Confirmed.

 

[Eug]

N1 chip - WiFi 7, Bluetooth 6, Thread
C1X - New 5G modem

 

[Eug]

A19 Pro - Apple indirectly admits it runs hot. iPhone 17 Pro has new vapour chamber heat shield and unibody aluminum shell.

 

[Eug]

iPhone Air uses binned A19 Pro SoC. iPhone 17 uses A19 non-Pro.



This could explain some of the competing rumours for iPhone Air. Some claimed it would use A19, and some claimed it would use A19 Pro. Well, like iPhone 17, the Air has an SoC with 6 CPU cores and 5 GPU cores, but the Air's SoC is a binned Pro whereas the 17's SoC is a separate SoC specifically designed that way.

 

[mvprod123]

A19 Pro:
Improved P-cores
50% more cache for E-cores
Massive GPU upgrade:
Next-gen dynamic caching
Doubled 16-bit floating point math rates
The new unified image compression
Neural core in each GPU core. (4x more compute than A18 Pro, MacBook Pro levels of compute)

 

[mvprod123]

"C1X is up to 2x faster than C1, and for the same cellular technologies, it is even faster than the modem in iPhone 16 Pro, while using 30 percent less energy overall. This makes C1X the most power-efficient modem in an iPhone."

Introducing iPhone Air, a powerful new iPhone with a breakthrough design

Apple today debuted the all-new iPhone Air, the thinnest iPhone ever made, with pro performance.

www.apple.com

 

[mikegg]

"C1X is up to 2x faster than C1, and for the same cellular technologies, it is even faster than the modem in iPhone 16 Pro, while using 30 percent less energy overall. This makes C1X the most power-efficient modem in an iPhone."

Introducing iPhone Air, a powerful new iPhone with a breakthrough design

Apple today debuted the all-new iPhone Air, the thinnest iPhone ever made, with pro performance.

www.apple.com

iPhone 17 Pro still uses Qualcomm?