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

 

[Glo.]

Future Apple Silicon Macs Will Reportedly Use 3nm Chips With Up to 40 Cores

The Information's Wayne Ma today shared alleged details about future Apple silicon chips that will succeed the first-generation M1, M1 Pro, and...

www.macrumors.com

I wonder how many people remember this article by source who correctly "predicted" all of M2 series.

 

[A///]

Future Apple Silicon Macs Will Reportedly Use 3nm Chips With Up to 40 Cores

The Information's Wayne Ma today shared alleged details about future Apple silicon chips that will succeed the first-generation M1, M1 Pro, and...

www.macrumors.com

I wonder how many people remember this article by source who correctly "predicted" all of M2 series.

Ma has connections that Gurman and Kuo do not.

 

[Glo.]

Ma has connections that Gurman and Kuo do not.

Thats why I believe M3 Max will have 20 CPU cores, and 48 GPU cores, and what we see leaking are only base models.

 

[poke01]

Thats why I believe M3 Max will have 20 CPU cores, and 48 GPU cores, and what we see leaking are only base models.

A17 will have 6 GPU cores, so maybe.

OLED laptop displays are quite a bit inferior to mini LED used by Macbook Pros. The brightness difference is staggering. I doubt Apple would want OLED until the tech gets much better.

In addition, I haven't seen any OLED laptop screens that are above 4k resolution and has 120hz. Those are minimum requirements for a MBP 16".

Oh wow, you did manage to find some on sale. Good job! I admit defeat on that one.

I will now go back to my old challenge:

* The best entry-level laptop SoC on the market
* Fanless
* Best entry-level laptop speakers
* Best touchpad
* All metal enclosure
* Retina display that others go up to 4k to match
* Same performance plugged in and unplugged
* 18 hours of battery life
* Uses more expensive LPDDR RAM

Find me a Windows laptop that comes close for $899 and has 16/512. You can't.

Isn't Apple using OLED in the 2024/25 MacBook Pro?

 

[trivik12]

OLED laptop displays are quite a bit inferior to mini LED used by Macbook Pros. The brightness difference is staggering. I doubt Apple would want OLED until the tech gets much better.

There is no way mini led is better than OLED. You cannot compare backlit with self emissive displays. Plus the OLED laptop displays are quite bright. S9 tab ultra from Samsung has 930 nits bright display !!! How much brighter do you need a display that is about a feet from your eye.

That said main issues with OLED is image retention which could be an issue for laptops. I think Apple will move to OLED once PHOLED based screens are there, that is supposedly mitigated. Otherwise you have to keep refreshing the displays.

 

[faizan9872]

Certainly, having a configuration of 4 big and 4 small cores might seem relatively conservative, especially considering the entire Mac product range. However, it's important to note that Apple's design philosophy often focuses on optimizing the performance of their custom silicon and leveraging software enhancements to maximize efficiency. While the core count might appear lower compared to some other products, Apple's integration of hardware and software often delivers impressive real-world performance and power management. This approach helps balance performance, energy consumption, and thermal considerations for a more holistic user experience across their product lineup.

 

[TESKATLIPOKA]

Notebookcheck.net​

​

Apple MacBook Air M2 entry doesn't offer 18 hours runtime, but It's still very good as long you stay at 150 nits, at max brightness(~500 nits) you end up with only 6.5 hours.

I am not sure why we should find a comparable Windows laptop for $899 when that's a discounted price most likely.
At Apple webpage, I see $1099 for the base M2 model.
In my country the base model costs 1239 euro. From when is this a price of an entry level laptop? I always thought that's 400-600 euro.

 

[A///]

apple is good at being vague in their marketing. 16 hours of playback on h264 isn't bad, but the real test imo is h.265 now a days especially if you roll your own rips.

 

[Doug S]

apple is good at being vague in their marketing. 16 hours of playback on h264 isn't bad, but the real test imo is h.265 now a days especially if you roll your own rips.

I don't think that's them being vague but realizing that the majority of the market is still h.264. Look at how much people complained about iPhones saving photos in HEIC format by default a few years ago. I haven't heard as much about that lately so maybe software on other platforms has caught up.

Since they have dedicated h.265 hardware (for several generations now, including encoding) I expect the battery life is pretty much the same either way. Sure decoding h.265 takes a bit more power, but the above test showing the massive difference between 150 and 500 nit demonstrates quite aptly that the display is going to be the biggest power draw in any video playback benchmark, especially if you can keep the big cores idle.

 

[A///]

I don't think that's them being vague but realizing that the majority of the market is still h.264. Look at how much people complained about iPhones saving photos in HEIC format by default a few years ago. I haven't heard as much about that lately so maybe software on other platforms has caught up.

Since they have dedicated h.265 hardware (for several generations now, including encoding) I expect the battery life is pretty much the same either way. Sure decoding h.265 takes a bit more power, but the above test showing the massive difference between 150 and 500 nit demonstrates quite aptly that the display is going to be the biggest power draw in any video playback benchmark, especially if you can keep the big cores idle.

those were two different statements unrelated to each other.

 

[mikegg]

Arguably its not because the m1/m2 being so powerful they are held back by low system specs.

Why not? What low system specs are you referring to?

So a laptop that gets physically hotter and is prone to heatsoak?

Except that it doesn't get hot even without a fan.

My dude doesn't have a sound bar, or a bluetooth speaker, or headphones? You want laptop speakers?

Yes. Having good laptop speakers is very useful. Not everyone wants to wear phones or have Bluetooth speakers. It's a laptop. Why would you carry around bluetooth speakers?

Its barely a hair above 1440p in resolution, putting it on the level of 4k is about a 60% stretch.

It's not about the resolution. It's about PPI.

The only way that's possible is by power limiting your chip, AKA reducing potential performance. That's just physics.

Nope. All Apple Silicon Macs have the exact same performance on battery life as plugged in.

I'd be interested if this one were actually true but I'm sure its not. Maybe if you only use the device for safari.

Just look up Apple's definition.

I concede that you have a specific list of requirements that only the product you like can fulfill. For me I don't know why someone would want to buy a premium SOC like an m1 to only have it used for chromebook duties.

I used an M1 Macbook Air 8/256 for software development for a year. It was faster than any Windows laptop I used at that time.

An M1 Macbook Air can easily do heavy tasks like software development and video editing.

 

[mikegg]

Nah. you are comparing gaming vs thin and light laptop.

https://www.notebookcheck.net/uploads/tx_nbc2/AcerSwift3SF313-53__1_.JPG

Acer Swift 3 tends to be constantly on sale. I have bought 12th gen for $350 with 8/512. Its great for my daughter for sure.

I'm not comparing gaming vs thin and light. I'm simply referring to the Asus laptop that the person mentioned.

Ah yes. Cheap Acer laptops. They provide a suboptimal experience and it will break in about 2 years - creating e-waste.

 

[Glo.]

Why not? What low system specs are you referring to?

MacOS.

Glad that I could help.

 

[Eug]

https://twitter.com/x/status/1693133717449936936

Apparently these are IDs that are considered valid on Apple's Tatsu Signing Server, and would represent new Apple chips. The new ones are believed to correspond to A19 and M5 series chips, which presumably would come out in 2025 and 2026.

 

[Doug S]

How are people getting these? Why would this stuff be exposed to the outside world for chips that won't see the light of day for 2+ years?

Can someone explain or guess at what is going on here? They'd need CPIDs for software support on internal test builds of iOS/macOS I guess, right? The chips won't tape out for at least a year, so I guess they would be running some 'still being worked on' version in a simulator?

How do they know they are M5/A19 rather than M4/A18 which would make a lot more sense?

 

[Ajay]

How do they know they are M5/A19 rather than M4/A18 which would make a lot more sense?

Yeah, Apple, the company that has practically perfected secrecy, is allowing this info out this early - not. I agree with you that the speculation here is way off, and if anything, we are looking at M4/A18 - or total BS.

 

[Eug]

How are people getting these? Why would this stuff be exposed to the outside world for chips that won't see the light of day for 2+ years?

Can someone explain or guess at what is going on here? They'd need CPIDs for software support on internal test builds of iOS/macOS I guess, right? The chips won't tape out for at least a year, so I guess they would be running some 'still being worked on' version in a simulator?

How do they know they are M5/A19 rather than M4/A18 which would make a lot more sense?

Interestingly, this is not new. This method of determining chip IDs has been out there for years, even if we hadn't been aware of it until recently. My understanding is if you query the server for a non-existent ID, it errors out. However, if you query the server an existing ID, it shows the ID, even if such a named chip hasn't been released yet. So how are people finding these IDs then? By brute force.

What does that tell us? It tells us nothing except that these IDs exist for unreleased chips. Does it confirm that these chips will actually get released? No it doesn't tell us that either. It just confirms the existence of these chip IDs. However, over the years, these chips have been released.

For example:
0x8110 is A15 - This ID was found in 2021 before A15 was released
0x8120 is A16
0x8130 showed up a long time ago, and is believed to be A17
0x8140 also showed up a while ago, and is believed to be A18
0x8150 is new and is believed to be A19 going by existing naming conventions

A19 and M5 chips confirmed by Apple's servers?

If you follow aaronp613's X/Twitter, as I'm sure MacRumors editors do, you may have seen this post from him that quotes a post from Orangera1n. Said post contains a screenshot of a list of numbers that supposedly confirms Apple is working on A-series chips up to the A19 and M-series chips up to...

forums.macrumors.com

BTW, similar inferences have been reported previously for new Apple devices. You can often find references to new Macs and iDevices on regulatory databases, but those references do not name those specific machines. What they do is just tell us new devices are in the pipeline, and sometimes they don't get released for close to a year.

 

[Doug S]

Well now I'm very curous about what 0x6033 and especially 0x6034 are, that's much more interesting than "Apple is working on A19/M5" which is kinda obvious!

 

[poke01]

Well now I'm very curous about what 0x6033 and especially 0x6034 are, that's much more interesting than "Apple is working on A19/M5" which is kinda obvious!

The 0x6030 is M3 Pro, 0x6031 is M3 Max and 0x6032 is M3 Ultra.

The 0x6033 and 0x6034 could be M3 chips, who knows?

 

[Eug]

The 0x6030 is M3 Pro, 0x6031 is M3 Max and 0x6032 is M3 Ultra.

The 0x6033 and 0x6034 could be M3 chips, who knows?

M3 Extreme!

 

[A///]

M3 Extreme!

It isn't too wild of an idea. there was the airport extreme for example.

 

[Doug S]

M3 Extreme!

My assumption would be that there would ONE number reserved for M3 Extreme. If there are two chips that go beyond Ultra that's very interesting news, and could lead speculation in a lot of directions.

I would give more credence to the idea of Apple introducing an I/O chip for interconnection beyond Ultra if there are two levels above Ultra instead of just one, for example.

 

[dr1337]

It's not about the resolution. It's about PPI.

I am really not sure why you think 2560x1600 has more PPI than 3840x2160.


Just because you want a very specific form factor and aren't bothered by bad specs doesn't suddenly make 8/256 workable for everyone. I already conceded that there isn't a windows laptop that doesn't match your every specific requirement at the same price. Even though you can instead get a laptop with a real GPU, higher refresh rate, the same PPI, and actual 16/512 specs.

Just look up Apple's definition.

You said you had one for a year, why not you speak about it? You should know better than marketing kool aid after all.

An M1 Macbook Air can easily do heavy tasks like software development and video editing.

And the 16gb ram model is actually faster at video editing and even photoshop. Hence why its such a shame to me Apple silicon gets choked with 8/256. That extra performance doesn't actually cost the amount they charge us for it.

 

[soresu]

I am really not sure why you think 2560x1600 has more PPI than 3840x2160.

PPI is about pixels per area, not overall pixels.

1920x1080 in 5mm² would be vastly higher PPI than 3840x2160 in 5cm².

 

[mikegg]

MacOS.

Glad that I could help.

In my profession (software engineering), macOS trounces Windows. I'm a software engineer in Silicon Valley. I'm estimating that 80 - 90% of the tech workers here use macOS. Probably 5% are using Macs but with Linux. And the rest are Windows laptops but installed with Linux. Hardly anyone use Windows here.

Gaming isn't the only thing that matters in this world.