• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion Apple Silicon SoC thread

Page 484 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Eug

Lifer
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:

Screen-Shot-2021-10-18-at-1.20.47-PM.jpg

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:


M1 Ultra discussion here:


M2 discussion here:


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:


M4 Family discussion here:


M5 Family discussion here:

 
Last edited:
also what’s the point of buying Intel Macs for bootcamp these days unless it’s a 2019 Mac Pro? Wouldn’t a windows PC be better and faster?
Crossover will be even faster than a windows PC in many cases, but in my experience outside of games, there's basically no software benefit to Windows. Every Windows user I've worked with considering a Mac had specific Windows apps they were accustomed to that didn't have Mac versions, but had Mac alternatives that they liked as much. And to be honest, like half of them got stuck on 'are there web browsers on the Mac' - that's sort of the level of unfamiliarity most consumers have with the software on their computer, and the level of need they have. Mail is a website, office apps are a website, tax prep is a website. Maybe they need Acrobat.
 
I’ll just say my biggest gripes are, non-native ability to close out windows in Mission Control ala Win + Tab, the Command + Tab isn’t as fluid as Alt + Tab, non-universal command + scroll wheel for zooming in and out.

Everything else though I’m cool with.

For the record I use both.
And to counter that, having a proper unix is critical along with all of the tools that typically come with it. Plus running python on Windows is a nightmare. An OS that can't reliably run python isn't functional in my eyes.
 
Automatic sync with your iPhone, same credentials, messages/SMS, etc. To the extent that people spend more personal time on their phones than their computers, that's a pretty compelling benefit for the ⅔ of Americans that are iPhone users, the 25% of global users. Go watch some of the Windows users try the Neo videos and they all note how amazed they are that they can open Photos and every photo from their phone is there, without doing anything, how they can receive and reply to SMSs, etc. things they cannot do on Windows.
you can do those on a windows pc if you have an android.

The one thing I found Windows does worse over macos is working sleep states, maybe cause MS is poor at kernel level optimisation these days
 
Last edited:
The switch to A19P will mean 12 GB instead of 8 GB which solves one of the big things people have complained about. That extra 4 GB raises costs, so I suppose they could bump the Neo's price to $649/$749 and it wouldn't really hurt sales, but I have to think they took that into account when they decided on the $599 price. The only reason I could see why they would have planned to almost immediately turn around and raise the Neo's price would be to reduce the butthurt among people who bought the A18P version with less RAM.
Agreed, I don’t buy that article either. I would also not be surprised to see a price increase with the 12 GB 19 Pro Neo. They have done that several times with the iPhone and Macs already when they increased RAM or storage. However, a $50 price increase would be worth it. Even $100 would be tolerable, although not preferred.

I also hope it gets better cooling than just graphite paper. 🤪
 
Agreed, I don’t buy that article either. I would also not be surprised to see a price increase with the 12 GB 19 Pro Neo. They have done that several times with the iPhone and Macs already when they increased RAM or storage. However, a $50 price increase would be worth it. Even $100 would be tolerable, although not preferred.

I also hope it gets better cooling than just graphite paper. 🤪
When in doubt apply a Thermal Grizzly cooling pad 🙂
 
Aside from the fact that running Python is a nightmare (or rather managing its dependencies), why running it on Windows is a bigger nightmare that running it elsewhere? Asking out of curiosity.
No POSIX support. No fork(). Python generally runs identically on linux/Mac due to having the same system libraries and requires platform conditionals for Windows.
 
It's such an abomination that they had to develop ways to manage dependencies. And not only one, multiple ones: venv, uv, poetry, etc. This language is a disgrace and a pain.
That's not quite fair. Much of those changes were to support native virtualization so you could easily deploy different versions of python without having to tip up docker schemes on some users computer. Python drives a HUGE amount of small team development/data science. Groups that rely on a single developer, or even a self-trained one, no devops, little or no IT support even. Having a platform that an epidemiologist can tip up a workable environment, build models with OS integration, and ship them off to a colleague across the country who also has no IT support and have it just run is something you can't do in any other environment. And those tools are widely shared and pretty critical to the scientific community. I'm not an epidemiologist but when I was tasked with our university covid response, those community tools were critical and I could build up a model using them in a few hours, push out results and iterate.

Career developers hate the language, but if you're on a clock and need to crank out fairly complex data projects very quickly, it's unmatched. It is the tool for domain experts doing data work.
 
It's such an abomination that they had to develop ways to manage dependencies. And not only one, multiple ones: venv, uv, poetry, etc. This language is a disgrace and a pain.

Every time I have to use Python for much of anything, I realize how bad users of non-Tcl (and non-Perl, tbh, but large Perl projects get ugly) scripting languages have it, and it makes me shed tears for the world.
 
No POSIX support. No fork(). Python generally runs identically on linux/Mac due to having the same system libraries and requires platform conditionals for Windows.
Well, since I work 99% time in Windows, might be I don't know what I am missing but it sounds it's mostly about efficiency rather than lack of functionality?

but if you're on a clock and need to crank out fairly complex data projects very quickly
Yes, it's useful for that. Personally I write most of my little scripts in Python, still does not make me love the language 😉

Though to be fair, I am afraid people will find this discussion to be off-topic...😉
 
Recently Microsoft executed a 180 degree spin and promised consumers everything they ever wanted: improved performance and reliability, less memory usage, control over updates, less push for AI, no more mandatory MS account etc. Their plan is to roll out these changes in the next few months, up to a year from now. They're desperate, mostly because the PC segment will take a nosedive in sales this year, but I bet they're also very nervous about Apple.

They can promise users control over updates, less push for AI and no more mandatory MS account - but those are simple switches they could deliver as part of the next Patch Tuesday if they were serious about it.

The other stuff you listed like improved performance and reliability and less memory usage are issues that Windows has needed improvement on for ages. Just because they give lip service to it doesn't mean they are going to actually deliver.

Maybe AI helps them find memory leaks but since a lot of people suspect the reason why Windows 11 updates have been so problematic for the past year is Microsoft increasingly using AI for coding that may be a double edged sword.
 
Well, since I work 99% time in Windows, might be I don't know what I am missing but it sounds it's mostly about efficiency rather than lack of functionality?
You're generally working around an open source toolchain. The Mac has all of the expected GNU tools, etc. In 15 years of developing python data tools I never once had to modify a tool that I developed on the Mac to run on a linux server. For instance there's no equivalent to fork() in windows, and even WSL didn't support it, though maybe they've addressed that by now. So it's entirely about lack of functionality.

I'm not trying to defend the design philosophy of the language, merely noting that it is effectively computing's Swiss Army knife - for good or bad - and in decades of working with colleagues Mac and linux users worked so freely that we often didn't know what platform each other was using, and every effort to collaborate with Windows users resulted eventually in them installing linux or buying a Mac. I'm sure python when isolated to Windows is just fine, but a LOT of community packages just don't run on Windows because of these things.
 
The other stuff you listed like improved performance and reliability and less memory usage are issues that Windows has needed improvement on for ages. Just because they give lip service to it doesn't mean they are going to actually deliver.
I mentioned all this to underline just how nervous they got recently, my expectations for MS actually delivering are slim: their management usually pivots before teams get a chance to make meaningful change.
 
The problem is all of the W10 hardware out there that can't upgrade.

How is inability to upgrade from W10 a problem?

I ditched my work PC for a Mac when the ultimatum came down to upgrade because they didn't want to pay for extended W10 support.

Since this is an Apple thread and not one for complaining about Windows, Python, or Python on Windows, does anyone else think that Apple will release a Neo Pro in order to tack on an extra $100 or $200 and cash in on the demand? It seems like a very Apple thing to do even if it's an it silly.
 
I mentioned all this to underline just how nervous they got recently, my expectations for MS actually delivering are slim: their management usually pivots before teams get a chance to make meaningful change.

What was the timing of this? Did it happen after the "Neo is $599" announcement or before? Because if it was before I can't see it how they are linked.

I could see Microsoft worrying about all the negative publicity around every biannual Windows 11 release having major issues, along with many Patch Tuesdays - or just that maybe it is personally affecting enough people that word of mouth is becoming a bigger issue than media reports.

I mean yes I think the Neo will take some PC market share, but consumer/small business Windows licenses are such a tiny tiny part of Microsoft's business these days compared to their server and cloud stuff, along with their AI ambitions, that I don't find it credible Microsoft has any real concern about the Neo.

That goes both ways. Apple seeing wild success with the Neo and doubling its PC market share would be a big deal for Mac developers, but hardly move the needle for Apple's overall business since the Mac is such a small segment of their overall revenue/profit.
 
How is inability to upgrade from W10 a problem?
Because Microsoft dead-ended their own hardware. So consumers then face a choice to move forward - a new PC that Microsoft may dead-end again, or a Mac. Enter one at $599.
I ditched my work PC for a Mac when the ultimatum came down to upgrade
How on earth do you not understand the problem for Microsoft when you made that choice?
 
What does Apple use for AI compute needs? just Cloud credits from Amazon/Alphabet? They are not creating their own models and so its all inference at the moment.
 
Back
Top