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

 

[Markfw]

wrong people wrong.
Modern mobile chips do that still from Intel to AMD to Apple to Qualcomm...

https://twitter.com/x/status/1548771664430858241

The ex-nuvia in twitter thread also agrees with dylan

I NEVER said anything about laptops.... And again, his statement is wrong, as modern desktops will NOT throttle given a proper HSF (servers also). He said "No chip can ever not throttle." I was giving an example that proved his statement wrong.

 

[poke01]

I NEVER said anything about laptops.... And again, his statement is wrong, as modern desktops will NOT throttle given a proper HSF (servers also). He said "No chip can ever not throttle." I was giving an example that proved his statement wrong.

oh then my bad. The whole throttle drama was about the M2 Air for the past few pages so I thought you were comparing desktop and laptop.

 

[Eug]

The M2 MacBook Air has a very thin heatspreader, albeit wide. It does cover other chips though, like the SSD. No finned heatsink or anything like that. It reminds me of the heatspreaders I'd see on like circa 2000 DVD players, etc., not a computer.



In the first pic you can see the heatspreader flipped up, exposing the thermal paste and the SoC underneath.



Here is the M1 MacBook Air heatspreader:

M1 MacBook Pro and Air Teardowns | iFixit News

Past, meet future—it’s a mix of old and new in our dual teardown of the newest M1-powered MacBooks.

www.ifixit.com

 

[Markfw]

oh then my bad. The whole throttle drama was about the M2 Air for the past few pages so I thought you were comparing desktop and laptop.

Thank you. His blanket statement was just wrong. Yes, this is the Apple m-series thread, so it was a bity off-topic, I apologise, just responding to a WRONG statement.

 

[poke01]

The M2 MacBook Air has a very thin heatspreader, albeit wide. It does cover other chips though, like the SSD. No finned heatsink or anything like that. It reminds me of the heatspreaders I'd see on like circa 2000 DVD players, etc., not a computer.

The M2 iPad Pro won't even have that. The M1 iPad Pro does not even have a heat spreader. People are overreacting these chips are designed to throttle in fanless designs.

You have an expert from ex-Nuvia saying this...

 

[poke01]

Thank you. His blanket statement was just wrong. Yes, this is the Apple m-series thread, so it was a bity off-topic, I apologise, just responding to a WRONG statement.

I also apologise if it felt rude. I still cannot believe people in general expect a fanless device to not throttle when under full load. So I got a bit worked up.

 

[Eug]

The M2 iPad Pro won't even have that. The M1 iPad Pro does not even have a heat spreader.

Well, you can see in iFixit's teardown that the M1 SoC in the iPad Pro has thermal paste. Why would it be there then?

It wasn't spelled out in iFixit's teardown (which was a worse teardown than usual) but even if it doesn't have its own dedicated heatspreader, it seems like it would be using some other surface as the heatspreader. Otherwise the thermal paste would be pointless.

 

[repoman27]

AMD and Intel chips have what is called a base clock. That is the clock you should be looking at. Boost clocks just mean 'hey, I am capable of boosting much higher if there is extra headroom'.

The base clock of a 5950x is 3.4 ghz. The chip will always operate at that speed unless you do a poor job at cooling it. Note that a $20 air cooler can cool the chip.

So, I guess it depends on how you want to define throttling. No, it will not go below base clocks if provided the minimum rated CPU cooler that can dissipate 115W of heat. Yes, if you cheap out on the cooler, you'll see it drop below the base clocks.

The reason the MBA throttles is because Apple had two goals:

1. It must be thin and silent
2. It must be cheaper than the 13" MBP

Working with those two goals, they chose to differentiate the air from the pro by allowing the air to throttle (among other differences).

After all, if the air were every bit of awesome as the pro is, maybe minus some ports, why would you even buy the pro?

Apple chips don't have a turbo or TDP per se, the system will allow the SoC to pull as much power as it likes (short of tripping over-current protection and shutting down) until a temperature sensor tells it to back off. It's just a different design decision than x86 vendors opted for. Apple doesn't bin, advertise, or sell chips based on clock speeds; everyone pays the same amount and gets the same peak performance. Apple can set the maximum clock speed as close to the efficiency sweet spot as possible as long as they meet their performance targets. And seeing as Apple designs both the chips and the devices they go in, they know exactly what the cooling solutions are capable of, and where they need to be in terms of performance per watt to achieve those targets.

This isn't directed at anyone in particular, I just feel it's worth pointing out. Apple most likely spends more money engineering cooling solutions for their devices than all other OEMs combined. The notion that they're somehow cheaping out is laughable. Do you have a design that can provide better thermal performance while at the same time requiring no additional mass or volume, does not exceed any critical dimensions (especially z-height), and offers the same or better accoustic performance? If you do, go ahead and start filling out a patent application—you've got a fine career ahead of you.

Do you understand the difference between throttling and internally downgrading the clock ? Throttling is what the Old chips did when they got too hot. Internally downgrading the clock is a calculated action by the CPU to stay in spec on power limits, NOT HEAT. Now if you are too stupid to put the correct heatsink on there, then it may throttle, but I qualified my answer, saying with a heatsink rated at 150 watt or more. His statement was that a CPU would never NOT throttle, and thats totally wrong.

I do see what you're saying, but it's a bit of a semantic argument. The only reason why CPUs have power limits is because ultimately that power must be dissipated as heat. It's not just to save you money on your electric bill. I'm the type of person that reads datasheets in my spare time, though, so maybe I'm being overly analytical here?

 

[beginner99]

But great, everybody wants useless short term peak performance even in mobile form factors now, all companies fall in line and start caring more about peak performance than efficiency, and everybody starts spreading myths that any other approach would be impossible. I honestly hoped Apple would be more immune to this.

Put peak performance helps for most other like launching an application or loading a new website. It makes sense honestly for most consumers. Of course when you game a lot on your phone and FPS go down over time it is less great but not really a huge issue.

Only thing to argue is that they should allow you to choose a mode. Like you can turn turbo off in bios. Eco-mode with no boosting. But honestly. i doubt that will affect battery life much. The display is in general the biggest offender so having displays that are a lot more efficient would help more.

 

[poke01]

Well, you can see in iFixit's teardown that the M1 SoC in the iPad Pro has thermal paste. Why would it be there then?

It wasn't spelled out in iFixit's teardown (which was a worse teardown than usual) but even if it doesn't have its own dedicated heatspreader, it seems like it would be using some other surface as the heatspreader. Otherwise the thermal paste would be pointless.

The thermal paste helps but the M Series iPad Pros and Airs won't even get any type of heat spreader because Apple operates the SoC to not go above a certain limit and if it gets too hot the chips slow down.

Apple puts more effort in the Macbook Air because it runs a more powerful OS but the same limits apply to MBA as well.

 

[moinmoin]

Those two preferences are in conflict.

They are only in conflict because everybody in the industry suddenly decided it's better for CPUs to be impulsive sprinters that topple over after some short time instead being marathon runners and improving the latter.

But apparently that's what "nature" decided so the latter is "impossible", never mind all the human design and market decisions that brought us to that point.

 

[Nothingness]

They are only in conflict because everybody in the industry suddenly decided it's better for CPUs to be impulsive sprinters that topple over after some short time instead being marathon runners and improving the latter.

But apparently that's what "nature" decided so the latter is "impossible", never mind all the human design and market decisions that brought us to that point.

You could try an experiment: prevent your processor from using the frequency it can't sustain without having to scale down. Then use your laptop for surfing and other interactive tasks. I bet you won't find it's a pleasing experience.

 

[Eug]

The thermal paste helps but the M Series iPad Pros and Airs won't even get any type of heat spreader because Apple operates the SoC to not go above a certain limit and if it gets too hot the chips slow down.

Apple puts more effort in the Macbook Air because it runs a more powerful OS but the same limits apply to MBA as well.

You say this as if it's some sort of badge.

While I think the MacBook Air is quite nice overall, personally think they could still have put a little bit more effort into the cooling design. Dunno about the M2, but for the M1 model, there are videos of simple mods out there that involve installing thermal pads in strategic places above the heat spreader, which improves performance.

But then that would make the M1/M2 MacBook Pro even more pointless than it already is.

 

[Heartbreaker]

Dunno about the M2, but for the M1 model, there are videos of simple mods out there that involve installing thermal pads in strategic places above the heat spreader, which improves performance.

At the expense of making the bottom of the laptop too hot. Passive cooling in mobile is very limited.

 

[gdansk]

They are only in conflict because everybody in the industry suddenly decided it's better for CPUs to be impulsive sprinters that topple over after some short time instead being marathon runners and improving the latter.

But apparently that's what "nature" decided so the latter is "impossible", never mind all the human design and market decisions that brought us to that point.

It's not impossible but there's a reason every CPU designed for consumer devices decided on having high short term performance. Apple could ship these at 2.5GHz where they could run without throttling. But they didn't because it is a better experience to have a few minutes of 3.5GHz available. The silicon can do it, so why not?

Apple shouldn't because someone online will be upset it cannot be 3.5GHz all the time and clearly that's just for benchmarks. No one has bursty workloads.

 

[Mopetar]

Thank you. His blanket statement was just wrong. Yes, this is the Apple m-series thread, so it was a bity off-topic, I apologise, just responding to a WRONG statement.

Every chip eventually throttles to 0 Hz when the ambient temperature has become so great that the chip becomes physically damaged and incapable of operating. Apparently you could destroy older chips this way, but it hasn't been true since the 90s when CPUs couldn't be passively cooled and started needing a fan of some kind.

/evilgrin

Almost anything these days has some kind of circuitry to detect if that's close to happening and shutting down before physical damage can occur. That I suppose is also a technical throttle.

But I'm just being pedantic and mincing words. Why? Because it's the internet!

/evilgrin

 

[poke01]

They are only in conflict because everybody in the industry suddenly decided it's better for CPUs to be impulsive sprinters that topple over after some short time instead being marathon runners and improving the latter.

But apparently that's what "nature" decided so the latter is "impossible", never mind all the human design and market decisions that brought us to that point.

Try low power mode in macOS it limits the M1 and M2 to 2GHz

 

[poke01]

You say this as if it's some sort of badge.

While I think the MacBook Air is quite nice overall, personally think they could still have put a little bit more effort into the cooling design. Dunno about the M2, but for the M1 model, there are videos of simple mods out there that involve installing thermal pads in strategic places above the heat spreader, which improves performance.

But then that would make the M1/M2 MacBook Pro even more pointless than it already is.

Apple like you said already provides a M2 wiith active cooling. Apple never designed the M2 Air for longer workloads.

So tech youtubers want to push M1 Ultra mac studio workloads on a fanless M2 Macbook Air. Seems like that to me editing 8K Canon RAW and doing extreme cinebench tests is going in that direction. like come on who renders for long hrs on a macbook air.

 

[Eug]

Apple like you said already provides a M2 wiith active cooling. Apple never designed the M2 Air for longer workloads.

So tech youtubers want to push M1 Ultra mac studio workloads on a fanless M2 Macbook Air. Seems like that to me editing 8K Canon RAW and doing extreme cinebench tests is going in that direction. like come on who renders for long hrs on a macbook air.

Probably a lot of people have at least some moderate renders, considering the MacBook Air is Apple's #1 selling computer, and Apple is marketing the MacBook Air's rendering prowess after all.

5 minutes is not a long render, even for light video editing.

Like I said before, for the hardcore video types, they need to buy a 14" or 16" MacBook Pro. However, for the others, I don't see the point of the 13" MacBook Pro, since it's inferior in so many ways to the 13.6" MacBook Air. However, it would have been nice if the 13.6" MacBook Air at least got a bit of a better heatspreader. Either that, or else give the 13" MacBook Pro an actual proper update.





FWIW, for looped high load benches, my Core m3 MacBook lost all of 7% of its performance after 30 minutes. The M2 MacBook Air loses much more than that.

Yeah, I realize my Core m3 is much, much slower overall, but I'm just saying it'd be nice to have a MacBook Air throttle a little less for this stuff, esp. if they plan on either killing off the 13" MacBook Pro or else if they plan on not properly updating the 13" MacBook Pro.

 

[Doug S]

Probably a lot of people have at least some moderate renders, considering the MacBook Air is Apple's #1 selling computer, and Apple is marketing the MacBook Air's rendering prowess after all.

5 minutes is not a long render, even for light video editing.

Well you can get rendering software for iPhones. That doesn't make it the appropriate tool for the job.

If there was a fan people would be whining about how loud the fan gets when it is running flat out and say "Apple ought to have a laptop that works without a fan seeing as how the iPad Pro manages it".

There's no way Apple is going to make everyone happy here no matter what they do. They put in a fan "why can't we have a fanless Macbook?" They clock it slower so it never throttles "why do they clock it down, we know it is the same silicon they use in the Macbook Pro Apple is just screwing its customers by artificially limiting the M2's performance!"

 

[moinmoin]

The amount of "that tool is not appropriate for the job" is mind boggling. People won't ask for that, they'll use what they already have and only few will be comparing performance in specific jobs, and even fewer will be in the position of taking action afterward (be it at work, due to financial limit, lack of choices etc.).

You could try an experiment: prevent your processor from using the frequency it can't sustain without having to scale down. Then use your laptop for surfing and other interactive tasks. I bet you won't find it's a pleasing experience.

I'm experimenting plenty with my Renoir system. For my use case my 4500U laptop is actually quite nicely usable as low as 3.5W cTDP (when focused on ST heavy applications, which applies for most of my surfing). But that's not the topic. (Though does anybody know a way to set M1's frequency and/or cTDP? Would love to experiment with it now that I'm building a test case for Renoir.)

Yeah, surfing and other interactive tasks are better at peak, just sad that that peak only holds for a limited time making surfing and other interactive tasks worse thereafter.

What people seem to gloss over is that throttling does not exist to sustain performance longer, it does to make the peak higher and then accordingly has to have lower lows to make up for the heat caused, allowing the heat to recede again. Thus for sustained performance a lower level which doesn't cause throttling actually achieves both a more stable performance and higher performance than when throttling after boosting.

 

[Eug]

If there was a fan people would be whining about how loud the fan gets when it is running flat out and say "Apple ought to have a laptop that works without a fan seeing as how the iPad Pro manages it".

There's no way Apple is going to make everyone happy here no matter what they do. They put in a fan "why can't we have a fanless Macbook?" They clock it slower so it never throttles "why do they clock it down, we know it is the same silicon they use in the Macbook Pro Apple is just screwing its customers by artificially limiting the M2's performance!"

I actually addressed this right in my previous post.

Apple has a 13" MacBook Pro that is kind of a joke. If Apple discontinues the 13" MacBook Pro or else if Apple continues to refuse to properly update the 13" MacBook Pro, the de facto machine at this price point is the M2 MacBook Air.

If they're going to give us separate performance tiers (based on the presence or absence of a fan) in their 13.x" Mac laptops, they should probably actually make the so-called Pro better than the the Air, but right now, the Air is better than the Pro in almost every way aside from raw CPU power (and the presence of the Touch Bar that most of us hate).

 

[poke01]

I actually addressed this right in my previous post.

Apple has a 13" MacBook Pro that is kind of a joke. If Apple discontinues the 13" MacBook Pro or else if Apple continues to refuse to properly update the 13" MacBook Pro, the de facto machine at this price point is the M2 MacBook Air.

If they're going to give us separate performance tiers (based on the presence or absence of a fan) in their 13.x" Mac laptops, they should probably actually make the so-called Pro better than the the Air, but right now, the Air is better than the Pro in almost every way aside from raw CPU power (and the presence of the Touch Bar that most of us hate).

Apparently apple is making a 15" macbook Air next year. That may be the day when Apple kills thr macbook pro 13".

The 15" Air should have better battery life than all macbooks and should handle sustained loads better than the 13" Air.

 

[poke01]

The amount of "that tool is not appropriate for the job" is mind boggling. People won't ask for that, they'll use what they already have and only few will be comparing performance in specific jobs, and even fewer will be in the position of taking action afterward (be it at work, due to financial limit, lack of choices etc.).


I'm experimenting plenty with my Renoir system. For my use case my 4500U laptop is actually quite nicely usable as low as 3.5W cTDP (when focused on ST heavy applications, which applies for most of my surfing). But that's not the topic. (Though does anybody know a way to set M1's frequency and/or cTDP? Would love to experiment with it now that I'm building a test case for Renoir.)

Yeah, surfing and other interactive tasks are better at peak, just sad that that peak only holds for a limited time making surfing and other interactive tasks worse thereafter.

What people seem to gloss over is that throttling does not exist to sustain performance longer, it does to make the peak higher and then accordingly has to have lower lows to make up for the heat caused, allowing the heat to recede again. Thus for sustained performance a lower level which doesn't cause throttling actually achieves both a more stable performance and higher performance than when throttling after boosting.

I believe you can't set it manually but turning on low power mode in battery settings automatically sets the M1 clock speed to 2ghz and this should lower package power.

 

[poke01]

5 minutes is not a long render, even for light video editing

The M2 Air can handle 5-10 minutes of sustained light workloads but longer than that it starts to slow down.