M3 Chip for Macs and A17 Chip for iPhone 15 Pro Will Reportedly Use TSMC's Second-Generation 3nm Process

newmacxc said:
The process nodes these days have just become a marketing gimmick which different companies use differently.

comparing intel and TSMC process nodes are like comparing apple to oranges.
Click to expand...
Making comparisons is possible of course , we can extrapolate density , we can estimate the leakage or the active power consumption , we cannot know the yield , but from leaks it looks like TSMC are doing well while intel are struggling.
You cannot compare the marketing names , that's true , but we most definitely can deduce that TSMC has the superior node at the cutting edge at the moment , that one is easy to see.

Just FYI , Intel also agrees , which is why they moved some of their high end products to TSMC fabrication :)
 
munpip214 said:
In case someone though we would be jumping to 2nm anytime soon:
TSMC's N3 family of process technologies will consist of five nodes in total, all of which will support FinFlex. The lineup includes the original N3, set to enter high-volume manufacturing (HVM) later this year, with the first chips set to be delivered in 2023; N3E with performance-per-watt and process window improvements; N3P with additional performance enhancements; N3S with increased transistor density, and N3X with support for increased voltages, enhanced power deliver; and augmented clock rate potential for ultra-high-performance applications.

TSMC Outlines 3nm Roadmap: Five FinFlex Nodes

TSMC's N3 node looks to have a big and bright future with FinFlex.
www.tomshardware.com www.tomshardware.com
Click to expand...

N3E sounds like the one to jump in.
Apple is doing the right thing starting at N3E for M3.
This is gonna be a quantum leap.
 
Last edited:
iPadified said:
An M3 for iPad Pro makes sense - the M2 does not.
Click to expand...

The A14X had one GPU core binned off its entire run. A14Z turned it on to eek out more time.
M2 with the "additional" GPU cores binned off , cap the turbo , and skip the 24GB RAM configuration is going to cause more problems than the M1 did?

It would have a better fab process node N5P ( vs N5 of M1) and LPDDR5 versus LPDDR4 (again if don't crank clocks to excess also a power savings ). Better economies of scale because using more common parts across products ( as M2 MBA and MBP 13" ramp and deploy. And other LPDDR5 usage across their line up. ).

The iPad Air has an M1. There is less of gap for the iPad Pro is squat for 9 months on the M1.
 
Boil said:
M3 will hopefully bring first-gen hardware ray-tracing...?
Click to expand...
EntropyQ3 said:
To the benefit of who, exactly?
Click to expand...

To anyone doing 3D work and needing/wanting hardware ray-tracing capabilities...?

munpip214 said:
...N3X with support for increased voltages, enhanced power deliver; and augmented clock rate potential for ultra-high-performance applications.
Click to expand...

Well, that sounds like the process the Mac Pro needs to be on...?
 
I think the M3 Macs (bases on a possible TSMC N3E process) will be a big leap in performance per watt and that the years after that progress won't be that big. So if you want to buy an ARM Mac I think you should hold off until Apple launches M3 Macs if it's just for performance. Apple will find other ways to sell new Macs in the years after launching M3 Macs.
 
MacRumors said:


Apple's future M3 chip for Macs and A17 chip for iPhone 15 Pro models will be manufactured based on TSMC's enhanced 3nm process known as N3E next year, according to a new report from Nikkei Asia. The devices are expected to launch throughout 2023.

m3-feature-black.jpg

N3E will offer improved performance and power efficiency compared to TSMC's first-generation 3nm process known as N3, according to the report.

In the meantime, the report claims that Apple plans to use TSMC's first-generation 3nm process for some of its upcoming iPad chips. It's unclear which iPad models the report is referring to, as rumors suggest that Apple will update the iPad Pro next month with the M2 chip, which is manufactured based on TSMC's second-generation 5nm process. A new entry-level iPad with an older A14 chip is also expected later this year.

The report claims that 2023 could mark the second year in a row in which only the Pro models of the new iPhone lineup feature Apple's latest chip. Last week, Apple unveiled iPhone 14 Pro models with an A16 chip based on TSMC's 4nm process, while the standard iPhone 14 and iPhone 14 Plus models are equipped with the previous-generation A15 chip.

Article Link: M3 Chip for Macs and A17 Chip for iPhone 15 Pro Will Reportedly Use TSMC's Second-Generation 3nm Process
Click to expand...
May the 2nm be next year as well :). and may the SSD be in the SOC as well, so it beat even the WD Black NVME and will be as fast as next year's RAM technology.

God bless, Rev. 21:4
 
UltimaKilo said:
Ufff, another person who fell for the “nm” gimmick. Pro tip: “nm” does not actually denote size.

Pretty soon some manufacturer will just say “-17000nm” and these people will go ape-sh*t for it! 🤣
Click to expand...

For someone who doesn't quite understand this, does that mean Apples Chips being 5nm aren't actually 5nm they're just some random arbitrary size but decide to call it 5nm?

Meaning that the release of 3nm in X years won't really mean 3nm, and again just a random figure when they decide they've upgraded their chipset?
 
NotTooLate said:
Making comparisons is possible of course , we can extrapolate density , we can estimate the leakage or the active power consumption , we cannot know the yield , but from leaks it looks like TSMC are doing well while intel are struggling.
You cannot compare the marketing names , that's true , but we most definitely can deduce that TSMC has the superior node at the cutting edge at the moment , that one is easy to see.

Just FYI , Intel also agrees , which is why they moved some of their high end products to TSMC fabrication :)
Click to expand...
Agree with you that TSMC has more advanced nodes than intel. What I implied by my post was a reply to one of the comment which said that TSMC 3nm is 1/3rd the size of intel 10 nm.
 
That's exciting, however it leaves the Mac Studio in a weird spot.
If it won't be updated this year with the M2 Max/Ultra I cannot see it getting an M3 Ultra next year.
 
And I’ll be sticking with my 1TB Sierra Blue 13 Pro.

Apple has pushed out a new iPhone every year to which they showcase little improvements. Years ago this was done with Apple adding an S. 4 and 4S and so on.

Now it’s a wasteful gimmick of having a whole number pushed out yearly by Apple which is suggestive of a completely new lineup oppose to the ever so slight improvements that are reflective of the new iPhone. Apple pushes these iPhones out yearly simply to drive their sales. Don’t be fooled, it’s hilarious at the hysteria some exhibit over the release of a yearly iPhone.

Think I’m wrong? Think about this. Apple charges developers 99.00 bucks a year or so to be in the developer beta program. Yes! Apple charges developers to be a Guinea pig all year long!
 
deconstruct60 said:
The A14X had one GPU core binned off its entire run. A14Z turned it on to eek out more time.
M2 with the "additional" GPU cores binned off , cap the turbo , and skip the 24GB RAM configuration is going to cause more problems than the M1 did?

It would have a better fab process node N5P ( vs N5 of M1) and LPDDR5 versus LPDDR4 (again if don't crank clocks to excess also a power savings ). Better economies of scale because using more common parts across products ( as M2 MBA and MBP 13" ramp and deploy. And other LPDDR5 usage across their line up. ).

The iPad Air has an M1. There is less of gap for the iPad Pro is squat for 9 months on the M1.
Click to expand...
Yes I know, but power consumption and thermal headroom is more sensitive in iPad than in many Macs including the MacBook Air. Will an M2 down clocked to M1 performance levels be significantly less power consuming compared to an M1? If not, there is no point. Is 5-10% power saving enough to make a new board to fit the larger M2?
 
Not sure if most users will really care how the chipsa re designed. They will only care about the resulting additional battery life, but is seems that the switch from 5nm to 4nm did not achieve a lot in that regard. Not sure if the main processor is the biggest energy consumer anyway. Most energy might go into the display and there are some physical limitation of how much energy you can save, if you still want more brightness each year.
 
meaning-matters said:
All thanks to ASML. ASML, the Dutch company that makes the most complex machines in the world putting the 3nm structures on the chip. TSMC is using these machines.
Click to expand...
CNBC did a nice overview of ASML a few months back for someone that used to be unaware of the process further back in the manufacturing process. Interesting stuff.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Back
Top