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

 

[scineram]

Then again, maybe they will just abandon the tiny Mac Pro niche:

It might be time for Apple to throw in the towel on the Mac Pro

Opinion: Making the case against Apple's most expensive—and most ignored—Mac.

arstechnica.com

It refuses to say explicitly, but to me it sounds somewhat like the Threadripper killed the Mac Pro.

 

[Tigerick]

It might be time for Apple to throw in the towel on the Mac Pro

Opinion: Making the case against Apple's most expensive—and most ignored—Mac.

arstechnica.com

The Ars article makes sense. With each new revision of Mac Studio, Apple will keep adding features like more GPU cores, double the memory capacity, video encoders and so on. Expandability of machine is also against Apple's design philosophy and not really benefits Apple financially..

 

[Doug S]

It might be time for Apple to throw in the towel on the Mac Pro

Opinion: Making the case against Apple's most expensive—and most ignored—Mac.

arstechnica.com

The Ars article makes sense. With each new revision of Mac Studio, Apple will keep adding features like more GPU cores, double the memory capacity, video encoders and so on. Expandability of machine is also against Apple's design philosophy and not really benefits Apple financially..

That doesn't mean a bigger model with double the cores is not desired by some.

That's like saying you don't need faster PCs because tablets keep gaining in performance and capability.

 

[mikegg]

What software did you use to meadure power consumption? Also, you should use Cinebench rather than Geekbench. Geekbench doesn’t load the CPU as heavily as other software.

Cinebench does not load the CPU properly.

 

[eek2121]

You do understand that the individual SPEC benchmarks are actual programs that are chosen because the represent particular types of applications and workloads right? You can see what they cover here: https://www.spec.org/cpu2017/Docs/overview.html#benchmarks

No benchmarks are perfect indicators of general performance or performance in a specific application, but SPEC has a broad set of performance measures. You might not care about all of them, but some may be indicative of performance in an area you care about.

So does Geekbench, but more importantly, you do realize that was a joke, right?

 

[Heartbreaker]

That doesn't mean a bigger model with double the cores is not desired by some.

That's like saying you don't need faster PCs because tablets keep gaining in performance and capability.

"Desired by Some" is not necessarily a large enough niche to make money on.

If you read the comments, the big compute market has largely moved on from Macs,, to Linux render farms, or cloud rendering. No matter what they put in a Mac pro, it's not going to beat a Render Farm or Cloud compute.

 

[Eug]

There are a lot of conclusions being drawn based on Mark Gurman's comments.

Anyhow, it seems he's seeing the same dev box leaks we are: 24-core CPU plus 192 GB RAM plus PCIe slots. Previous comments from those like Gurman about Apple's original plans notwithstanding, I haven't seen leaks of a 48-core CPU / 384 GB RAM dev box either, and perhaps the lack of those dev box rumours after all this time leads him to believe it doesn't exist.

Perhaps that is true, but it should be noted that the 24-core CPU / 192 GB RAM dev box leaks shouldn't have happened in the first place. Maybe the devs with the 48-core CPU / 384 GB RAM boxes are better at keeping their mouths shut.

/wishful thinking

 

[Heartbreaker]

It refuses to say explicitly, but to me it sounds somewhat like the Threadripper killed the Mac Pro.

If you need more than what a Studio Ultra can deliver compute wise, then yeah, you probably want Threadripper or a render farm of Threadrippers, or several GPUs for GPU compute.

What the people that still want a Mac Pro, probably want is the expansion capability of a Mac Pro, more than the pure compute power.

That expansion role, seems to have been significantly curtailed by Apples SoC with Unified memory model.

I don't think Apple want's to support any third party GPU on ARM Macs, so that is one major blow.

RAM is currently soldered in, socketed RAM expansion would slow, so another major blow.

I guess they could have some PCIe based storage...

An ARM Mac Pro seems kind of problematic. Which is why it's the model I've been most interested in seeing.

 

[Doug S]

"Desired by Some" is not necessarily a large enough niche to make money on.

If you read the comments, the big compute market has largely moved on from Macs,, to Linux render farms, or cloud rendering. No matter what they put in a Mac pro, it's not going to beat a Render Farm or Cloud compute.

Sure, if you assume anything people want to do with a workstation is "rendering" I suppose.

The higher end you go the more niche everything is. I don't think any of us know how niche the Mac Pro is, obviously it is a small percentage of Apple's sales but 1.5% of sales would still add up a heck of a lot of units per year. If it is .015% then that's a different story, but I've never seen any analyst firm even attempt to break down Apple's Mac product mix so only Apple knows the real numbers.

 

[scineram]

Nowhere near 1%, they sell millions of Macbooks every quarter. The Mac Pro is also the most neglected and obsolete of the lineup. Has been for a decade. Most of the big costumers from long ago have defected by now. The new one would have to be something that actually entices buyers back.

 

[LightningZ71]

It makes FAR more sense to invest in a cloud service that works seamlessly with MacOS on client devices. Everything in the cloud can be abstracted and they can provide the service to any devices they want.

 

[Doug S]

Nowhere near 1%, they sell millions of Macbooks every quarter. The Mac Pro is also the most neglected and obsolete of the lineup. Has been for a decade. Most of the big costumers from long ago have defected by now. The new one would have to be something that actually entices buyers back.

Looking Statista's numbers they sell 6 or 7 million Macs a quarter since the ARM transition. They would only need to sell a quarter million ARM Mac Pros a year to reach 1%. Will they? I have no idea, but I don't think anyone can confidently claim they do not.

I don't think they'd need to sell a quarter million of them to make it worth the design effort. 100K is probably more than enough. 50K may be pushing it.

 

[scineram]

I believe current sales are already down there or lower. Not convinced such sales would justify the effort which is more than just chip design. But that itself bad enough apparently.

 

[mikegg]

It makes FAR more sense to invest in a cloud service that works seamlessly with MacOS on client devices. Everything in the cloud can be abstracted and they can provide the service to any devices they want.

Correct. Exactly what I predicted over a year ago.

An Apple Silicon Cloud would enable a Macbook Air user sitting in a coffee shop to tap into a 48-core CPU, 160-core GPU, 384GB of RAM/VRAM "Extreme" chip in the cloud somewhere for a few hours to finish heavy work.

Apple can do the integration work with macOS to make this experience seamless.

Apple can sell a physical Mac Pro with an "Extreme" chip or rent it out in the cloud. It could justify the cost of developing such an insane chip.

It'd open up a whole new business for Apple.

Mark Gurman is saying that Apple is working on a 40-core SoC for the Mac Pro for 2022.

You're Tim Cook, sitting in his nice office, looking at how much money you just spent to make this giant SoC for a relatively small market. In fact, you have to do this every year or every two years to keep the Mac Pro relevant. How do you recuperate some of this money spent?

You create "Apple Cloud". No, not iCloud. Apple Cloud. Like AWS. Where anyone can come and rent a 40-core M3 SoC running on macCloudOS. You get into the cloud hosting business. You file this under the "Services" strategy that you keep pushing to make Wall Street happy.

Soon, you'll be releasing 64-core SoCs with 128-core GPUs, then 128-core SoCs with 256-core GPUs, and so on. Somehow, you're actually beating anything AWS, Azure, Google Cloud can offer... without really trying.

Apple Silicon Cloud.

It wouldn't surprise me if Apple is already testing their own SoCs to power their iCloud service, which currently depend on AWS. Apple was reportedly spending $30m/month on AWS in 2019. It might be $100m+ per month by now given how fast services have grown.

You're Apple. You just built a 40-core SoC for Mac Pro. Now what?

Mark Gurman is saying that Apple is working on a 40-core SoC for the Mac Pro for 2022. You're Tim Cook, sitting in his nice office, looking at how much money you just spent to make this giant SoC for a relatively small market. In fact, you have to do this every year or every two years to keep...

forums.macrumors.com

 

[mikegg]

Looking Statista's numbers they sell 6 or 7 million Macs a quarter since the ARM transition. They would only need to sell a quarter million ARM Mac Pros a year to reach 1%. Will they? I have no idea, but I don't think anyone can confidently claim they do not.

I don't think they'd need to sell a quarter million of them to make it worth the design effort. 100K is probably more than enough. 50K may be pushing it.

They actually sold 10m in the last quarter.

Let's say that 0.01% of those are Mac Pros and the average revenue per Mac Pro was $25,000.

100,000 units * $25,000 = $2,500,000,000 billion.

Apple made $11b from Mac sales total last quarter.

So 2.5B / 11B seems too big. Mac Pros shouldn't be able to make up 22% of all Mac revenue. If so, Apple would pay far more attention to Mac Pros. But they don't.

So I'm guessing that Apple sells about 10,000 Mac Pros each quarter @ $25k on average. That would equate to $250m/quarter or $1b total per year in sales. I'm not so sure if $1b/year in sales would justify the insane engineering challenges of glueing 4x Mac dies together. Don't forget software optimization at the OS and apps level too.

 

[Eug]

I'm not so sure if $1b/year in sales would justify the insane engineering challenges of glueing 4x Mac dies together. Don't forget software optimization at the OS and apps level too.

Uh wut? One billion bux? Of course that would justify it. And it doesn't even have to be profitable on its own.

Research and development at the uber high end isn't just about catering to the uber high end.

BTW, I wonder how many sets of $400 Mac Pro wheels they sell...

 

[mikegg]

Uh wut? One billion bux? Of course that would justify it. And it doesn't even have to be profitable on its own.

Research and development at the uber high end isn't just about catering to the uber high end.

BTW, I wonder how many sets of $400 Mac Pro wheels they sell...

$1b in revenue. Not $1b in profit. The cost of materials, assembly, R&D, marketing, etc. are not included. Apple could be making a loss on the Mac Pro today, without Apple Silicon.

 

[Eug]

$1b in revenue. Not $1b in profit. The cost of materials, assembly, R&D, marketing, etc. are not included. Apple could be making a loss on the Mac Pro today, without Apple Silicon.

Yes, $1 billion in revenue would justify it. As mentioned, it doesn't even have to be profitable.

 

[mikegg]

Yes, $1 billion in revenue would justify it. As mentioned, it doesn't even have to be profitable.

Why not? As an Apple shareholder, why would Apple make a product that is intentionally unprofitable?

It doesn't make sense. Everything Apple makes is profitable or it gets killed.

The Intel Mac Pros are probably profitable because they simply use Intel and AMD chips. I'm not convinced that an Apple Silicon Mac Pro could be profitable.

 

[Doug S]

They actually sold 10m in the last quarter.

Let's say that 0.01% of those are Mac Pros and the average revenue per Mac Pro was $25,000.

100,000 units * $25,000 = $2,500,000,000 billion.

Apple made $11b from Mac sales total last quarter.

So 2.5B / 11B seems too big. Mac Pros shouldn't be able to make up 22% of all Mac revenue. If so, Apple would pay far more attention to Mac Pros. But they don't.

So I'm guessing that Apple sells about 10,000 Mac Pros each quarter @ $25k on average. That would equate to $250m/quarter or $1b total per year in sales. I'm not so sure if $1b/year in sales would justify the insane engineering challenges of glueing 4x Mac dies together. Don't forget software optimization at the OS and apps level too.

You mean 1% or 0.01 not 0.01%.

I think $25K is a massive overestimate of the average sales price. I doubt the ASP is over $10K, there aren't likely to be a lot of people loading them up with 1 TB of RAM or whatever - especially since the current Mac Pro can have DIMMs added later. If the new one is a fixed config that may increase the ASP somewhat, but unless they offer denser LPDDR stacks then $10K might end up the most you could possibly spend on it.

 

[MadRat]

Why would anyone that spends $10,000+ on a workstation for conceivably projects worth 10x-1000x the value of the workstation, want their data to touch Apple's cloud? That is ludicrous. Apple would be in position to spy on their information and possibly perform espionage. Anyone that believes Apple would not, I have a line of premium bridges to sell you.

 

[Nothingness]

Why would anyone that spends $10,000+ on a workstation for conceivably projects worth 10x-1000x the value of the workstation, want their data to touch Apple's cloud? That is ludicrous. Apple would be in position to spy on their information and possibly perform espionage. Anyone that believes Apple would not, I have a line of premium bridges to sell you.

Do you think other cloud providers are better? Or do you think Apple in particular is evil?

 

[lopri]

My M1 mini slowed down considerably since I started using. Not as snappy as it once was. I suppose this is an OS thing?

 

[Eug]

My M1 mini slowed down considerably since I started using. Not as snappy as it once was. I suppose this is an OS thing?

Probably. I have an M1 mini which I just got a few weeks ago, and it's still quite peppy running macOS 13.1 Ventura, but that was a clean install when I got it.

However, I've noticed with my Intel Macs at least that over time they sometimes have slowed down as you describe. Sometimes it's due to overly full SSDs due to unknown accumulations of "System" or "Other" files, and sometimes I just don't know why. If you clean out the junk it can often speed back up but occasionally I've done the nuclear option and just reformatted the drive. The latter always makes it snappy again.

And then there are some applications which are just slow in one way or another. For example, even though everything else is fast on my M1 Mac mini, I just installed Office 2016 on it and it takes a very long time to load (~10 seconds). One could blame the fact that it's running through Rosetta 2, but I don't think that's it, because it took even longer on my Intel Macs.

---

On other note: I am not getting HDCP 2.2 out of my M1 Mac mini. These Apple Silicon Macs are doing funny things with regards to HDCP on third party monitors.

On my 30" Apple Cinema HD Display with my Intel Macs, I could get HD video streaming through Safari from Disney+, but I can't with my M1 Mac. It's as if it just doesn't provide the required HDCP 1.x support.

On my new Huawei Mateview 28, I can get HD video streaming through Safari from Disney+ so it appears HDCP 1.4 works, but I cannot get 4K video streaming from Disney+. It's as if it just doesn't provide the required HDCP 2.2 support. (I tried USB-C, HDMI, and a third party USB-C to HDMI dongle.) However, I know HDCP 2.2 is supported by the monitor, because 4K Disney+ streaming works fine on the monitor when I plug in my Apple TV.

Here's hoping future iterations of Apple Silicon improve 3rd party monitor support.

 

[lopri]

My drive is definitely not full, but OS files may have been fragmented. What is the best way to "clean" it?