I’m not sure why no one has mentioned this but, maybe these are for the new MacPro that could be released at WWDC 2023!
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
3nm Chips From Apple Supplier TSMC to Enter Mass Production This Week
- Thread starter MacRumors
- Start date
- Sort by reaction score
The Digitimes says M2 Pro and Max will be 3nm in Aug 2022, this can be correct if Apple decides to delay the 14/16 Macbook releasing until Oct 2023. That will be 2 years update since M1 Pro.
Otherwise, it will be 4nm as same as A16. Obviously, we all want 3nm soon.
Meh I think your topic is close enough 🤪
I don’t want to just complain about how bad things are. I understand the technical challenges to some degree. It just seems like I’ve been more conscious lately about how often I’m trying to use my iPhone, iPad or other tech device and I’m waiting on something to start/load/process. If I’m trying to do some really unusual tech feat I completely accept that it may not work as desired. But geez Louise I just want to brush my teeth and the toothbrush app won’t start or connect. I want to look for a new app on my iPhone and the App Store app just gives me a blank screen. When I can run a search it just does nothing until I force quit and try again.
Geez I got kind of carried away. ☺️🙄
That really depends on what the sales breakdown is for the Studio.
Are the majority of Studios sold entry level, low-spec versions? Then, perhaps a M1pro/M2pro mini kills Studio sales…
Otherwise, I know that there’s plenty of Intel Mac owners, waiting for Mac Minis and iMacs with more oomph than the now 18mo old standard M1, who don’t need the Studio, and don’t want to use a MBP as a desktop replacement.
An M1pro or M2pro mini is an immediate buy for me, one for the home and one for the office. Apple just needs to actually want my money?
Last edited:
Why choose?
Apple could use the less power use at same speed for the iPhones, iPads and the Mb Air, and more speed at the same power usage for the MBPs and desktop Macs.
There’s also a 3rd choice of balancing between the 2. So we can get slightly better performance per watt while also getting slightly better efficiency.
That’s really the tip of the iceberg though because there’s also a lot of other improvements that come from not only 3nm but also Apple (and ARM) making general improvements to their chip designs and infrastructure. Not to mention clock speed adjustments and thermal improvements. Even this early on, we can already safely assume that these 3nm chips will have: Better efficiency cores, more powerful main cores, higher number of CPU and GPU cores. Based on the rumor of what the A16 was supposed to be, we can also assume that these 3nm chipsets will have hardware ray tracing cores as well.
Things are looking GOOD for the VR headset, A17 powered iPhones and the M3 Mac devices.
The real question now is if Apple can make the M3 outperform above its weight class like they did with the M1… The competition is catching up so let’s see if Apple can widen that gap once again.
It's well-known, from TSMC's own press releases, that the N4P process (likely used in the A16) is 5 nm, not 4 nm:
"As the third major enhancement of TSMC’s 5nm family, N4P will deliver an 11% performance boost over the original N5 technology and a 6% boost over N4."TSMC Expands Advanced Technology Leadership with N4P Process|Taiwan Semiconductor Manufacturing Company Limited
Hsinchu, Taiwan, R.O.C., Oct. 26, 2021 - TSMC (TWSE: 2330, NYSE: TSM) today introduced its N4P process, a performance-focused enhancement of the 5-nanometer technology platform.pr.tsmc.com
So how much weight should we give to DigiTimes' TSMC reporting when they aren't even aware of something so basic?
This is ridiculous. You're quibbling over whether the sky is orange or green, when you can easily see that it's neither.
In fact, none of these technologies are truly 5nm, much less anything smaller. Various features are 5-10x larger than that. But that's irrelevant, as everyone in the business knows that these sizes have been marketing labels for years.
N4 and N4P each achieve better density than their predecessors. So in some intuitive sense they're "smaller". No, of course they're not 4nm, but since N5 isn't 5nm, who cares?
DigiTimes is full of terrible errors, but complaining about this is just silly.
I know what you meant to say, but...
While there's no doubt we will see some changes in the P and E cores (and I really really hope for 15-20% IPC in the P cores, though that's probably unrealistic), that's not where Apple needs to do the most work. Nor, fortunately, where they're likely to find the low-hanging fruit. They're breaking new ground on massive high-bandwidth interconnects (think the InFO-LI tech used in the M1 Ultra) and they've got huge inefficiencies there which means tons of opportunity to improve. They also have to continue to scale up their NoC and the rest of their uncore, which in addition to a much-needed multicore performance boost, is likely to do more for their GPU performance than any raytracing hardware (though it would be nice to see that too).
BTW, to the poster who talked about the new Snapdragon beating the A16's GPU: Don't be too hasty. Synthetic benchmarks are interesting, but real-world results matter. The jury is out but there are strong indications the Gen2 is not performing as well as claimed, leaving the A16 still the winner.
As for CPU: The first 80% of the performance difference between the X-series and the A16 was always going to be the easiest part to make up. So it's not surprising that ARM's cores are catching up (though it is surprising that it took them so long, if you don't take into account non-engineering-related industry trends). Don't expect that making up that last 20% will be as easy, or that Apple won't be moving ahead in the meanwhile. We're in early innings, yet, with a lot of game still to play out.
I disagree, $699 to $1999 is too big of a gap to not have something in between, plus I firmly believe that Apple would've nixed the intel Mac mini already if they weren't planning on replacing it with a Pro chip model. The Mac Studio will continue to get better with additional features in order to entice people to get it over an M#Pro Mac mini.
Oh Jesus Christ. Do you also want to tell us the breaking news that nothing on N5 is actually sized at 5nm, and nothing on N3 is sized at 3nm?It's well-known, from TSMC's own press releases, that the N4P process (likely used in the A16) is 5 nm, not 4 nm:
"As the third major enhancement of TSMC’s 5nm family, N4P will deliver an 11% performance boost over the original N5 technology and a 6% boost over N4."TSMC Expands Advanced Technology Leadership with N4P Process|Taiwan Semiconductor Manufacturing Company Limited
Hsinchu, Taiwan, R.O.C., Oct. 26, 2021 - TSMC (TWSE: 2330, NYSE: TSM) today introduced its N4P process, a performance-focused enhancement of the 5-nanometer technology platform.
So how much weight should we give to DigiTimes' TSMC reporting when they aren't even aware of something so basic?
And why doesn't MR do some basic fact-checking before posting this stuff? Given that the whole point of this article is to report on a process change, you'd think they want to get the process descriptions right.
This doesn't make sense. First they say that mass production on the first gen 3 nm chip (N3) starts in a few days, but now they're saying that production on 3 nm is unlikely to ramp up until the next gen 3 nm (N3E) gets underway. This is just screaming out for clarification, yet MR gives us none. Come on guys, how putting some thought into these articles, rather than just cutting and pasting a few quotes from another source (which, incidentally is behind a paywall, preventing us from checking it ourselves)?
Look that ship has sailed. Anyone who hasn't been dead for the past ten years is WELL AWARE that N5 means '5nm class" which means fsckall 5nm related except that it's a set of improvements that give you about 1.8x density increase over N7.
Likewise anyone who actually matters and cares is well aware that N4 is an optical shrink of N5. You don't need to act like you're Woodward and Bernstein here revealing some scandal.
Finally the N3/N3E situation is also well understood.
- TSMC began by designing N3. The yield was disappointing and this has not improved as much as was hoped.
- Given this (and somewhat in parallel, because TSMC always has a plan B in hand) they designed a second process N3E that is not quite as dense as N3, but better in other ways (definitely better yield, slightly better power and frequency); overall a better a set of tradeoffs.
- OK, so that at this point, which process gets rolled out? Answer
+ Apparently Apple will use N3 for SOME products (probably low volume, probably those that have been in the pipeline as designed for N3 from the start, so likely the M2 Pro, Max, Ultra? Possibly also the first round of chips for glasses?) Apple will likely do this to meet schedule. But everyone, Apple and TSMC included, want to get off N3 as soon as N3E is ready.
+ OTHER COMPANIES, and Apple's main volume (ie A17, M3), will move to N3E as soon as it is available.
This all makes perfect sense, and fits with everything we know about N3 and N3E.
If you want to follow this stuff, read the people who explain it well , like Dylan Patel:
TSMC’s 3nm Conundrum, Does It Even Make Sense? – N3 & N3E Process Technology & Cost Detailed
Shrinking finally costs more, Moore’s Law is now dead in economic termsA couple of weeks ago, we were able to attend IEDM, where TSMC presented many details about their N3B and N3E, 3nm class…
www.semianalysis.com
................................
Going forward you need to realize that lithography is only one part of chip manufacturing, and the significant next few steps are not captured by obsessing over "5nm" vs "3nm".
What mattered was
- N7 introduced EUV
- N5 used more EUV
- N3 tried to use double-patterned EUV, but that was too ambitious and so
- N3E uses all the other improvements of N3 (eg chemical improvements) but single-pattered EUV
-N2 will use high-NA EUV which allows for smaller lithography while still not yet requiring double patterning.
- N2 will also introduce GAA FETs which are not smaller than FinFETs, but will have lower power and higher frequencies.
- HOWEVER right now the biggest limitation in designs is not transistor size, it is wiring density. Lithography can't really help with that. What's needed is to move as many wires as possible to BELOW the transistors rather than having them all above. This is so-called BSPD (back-side power delivery).
In other words the same moron crowd that have recently become hysterical about lack of SRAM density improvements for N3 need to get ready for another round of hysteria for N2 (even though this was al understood and expected years ago). N2, like N3,
- improves transistor QUALITY (power, frequency)
- allows for smaller transistor which gives improved logic density BUT
- does not allow for much density improvement where wires dominate (ie SRAM)
You can see some of the numbers here:
TSMC: N2 To Start With Just GAAFETs, Add Backside Power Delivery Later
BSPD is currently scheduled for a second phase of N2, an N2E or N2P or whatever it's named. For what I care about, it's a more interesting development than N2 and GAA, and I suspect that TSMC want to introduce it as soon as possible; but it is a very different sort of tech, and requires a whole lot of novel fab tooling.
Intel is in essentially the same place. The names of i4, i3 and i18 just do not matter, they're just labels (cf N4 vs N5). What matters are the concrete changes of
- use of EUV
- how many layers of EUV
- high NA EUV
- GAA
- BSPD
Each of those is a real step forward, not minor tweaks. Minor tweaks have their value (if I can get eg a 6% performance improvement for free, who's complaining!? That's why an N4 or an N3E have real value!) but it's the big steps forward that build the future.
Am I the only one who feels this would have to be for an M3 chip? Shouldn't an M2 Pro/Max chip be based on the exact same architecture as the M2 chip but with more cores etc. I would hope, and expect, Apple to name this chip M3 Pro/Max and just have the MacBook Pros skip the M2 Pro/Max.
Apple is free to name anything they want any way they want. It is all marketing. No one here knows what Apple is going to do.
Angstroms (2nm = 20A)
Intel’s foundry roadmap lays out the post-nanometer “Angstrom” era
Could the “Intel 20A” process node (that’s 20 Angstroms) rival TSMC and Samsung?
arstechnica.com
A lot of us are just waiting for new MacBook Pros at this point no matter what chip they come with.
No.
The A15 is on N5 and the M2 (as far as is known) is also on N5.
The A16 is on N4.
The M2 Pro and Max are wildcards. They COULD be on N5 (and be basically expanded A15/2's)
They could be on N4 (and be basically expanded A16's).
They could even be on N3 (and be basically expanded A17's, where the A17 was what should have shipped this year but was delayed because N3 was delayed, and A15 was slightly improved to take advantage of N4).
We honestly do not know. My GUESS is that Apple took the gamble of waiting for N3 and using the design that was ready for N3 (ie what will become the A17 design), and that this is, in part, why they are not worried about how the upper end of the line "looks stagnant" (so they haven't felt a compulsion to make the lines look "more exciting" by introducing an M2 mac mini or some sort of MacBook Pro minor improvement [better WiFi/BT chip, or whatever], or price reductions); I think they are confident that everything will be forgiven with the N3/A17-based SoC's in April or March.
And I think they don't especially care that the "meaning" of M2 vs M2 Pro/Max/Ultra will be something very different in terms of CPU; it was already the case that M1 vs M1 P/M/U used different LPDDR DRAM and had not just better GPU but also the ProRes blocks.
You're right that N5/N4P are just marketing terms. And you're right that the 5 nm process family isn't really 5 nm, and the 3 nm process family isn't really 3 nm. But you're wrong in concluding that it therefore makes no difference whether you call N4P 5 nm or 4 nm, for a very basic reason:
Different nm designations, e.g., "5 nm" and "3 nm", mean the chips are on different processes—different nodes. If you say a chip is "3 nm", it means it's a different process from chips that are 5 nm; 3 nm is not merely a 5 nm node optimization. Thus problem with calling N4P "4 nm" is that it incorrectly communicates that this chip is on a different process from 5 nm, which it's not. Instead, it's merely a process optimization of 5 nm.
I.e., to say there's no difference between calling N4P "5 nm" or "4 nm" is to say there's no difference between identifying it as an optimization of an existing process (which it is) vs. an entirely new process (which it is not).
TSMC themselves understand this, which is why they are careful to say that N4P is part of the 5 nm family, and does not represent a new node. They can play games with the numbers they use, but calling something an entirely new node technology when it's not is, fortunately, a game they're not (yet) playing.
Hence I conclude it's your post that is ridiculous and silly, not mine.
Last edited:
Damn! I need one of those Wired magazine explainer videos to 5 levels of difficulty. I think you just skipped to the expert level! Nicely done!
I don't think anyone who actively avoids buying Gen 1 anything (cars, computers, phones, TVs) want Apple to skip the M2 Pro/Max for the MBPs (and minis and iMacs?) and wait for the M3 Pro/Max a year from now?
And, I don't think Apple wants to wait another year to take those people's money, either?
Apple isn't going to let it's best selling Macs sit at Gen 1 much longer... The Air has the M2, hell an iPad has the fraking M2... they need the MBPs to get updated sooner than the M3 Pro/Max will be available in quantity ...
Yes and no. To some extent you get the two alternatives (lower energy or higher frequency) by using a different standard cell. So you can't design a SINGLE chip that gives you both the lowest possible energy behavior (at low voltage) and the highest possible frequency behavior (at higher voltages).
Of course Apple could use different designs for the MBA/iPad chips vs the chips to go into iMac and Mac Studio (though higher frequency cells are larger area, so does Apple want to pay the area price...)
The complication is that (at least for now...) Apple mostly wants to use those same chips (Pro and Max) in MacBook Pro's...
So my guess is they *at some point* they will probably tweak the designs to lean to a slightly higher frequency, and slightly larger area, for Max and Pro, and just accept that MacBook Pro's have slightly lower battery life than MBA's.
But that point may not be this year...
This year we may still have the essence of the baseline design (for iPhone, always choosing energy over frequency) just because there are so many other things they have needed to get sorted out (the product line is still a mess, IO is still a mess with best of breed PCIe, USB and TB required, not the mid-level IO that we got with M1, there are many things related to scaling to Ultra and beyond that need to be fixed, and covid of course messed up all the timelines). It's more important to get all that sorted out than to tweak designs so that the Studio's can run at 4.0 rather than 3.8 (or whatever) GHz.
100% agree. I'm totally fine with new mini with standard M2, the sooner the better.
If only the facts supported your argument.
To quote from TSMC's own website: https://www.tsmc.com/english/dedicatedFoundry/technology/logic/l_5nm#l_5nm
"In addition, TSMC plans to launch 4nm (N4) technology, an enhanced version of N5 technology. N4 provides further enhancement in performance, power and density for the next wave of N5 products."
Yes, N4 is widely considered part of the N5 family, including by TSMC. No, you are wrong to say it's "5nm", as that's just a label, and the one chosen by TSMC is "4nm".
In fact I wouldn't be at all surprised to see TSMC refer to it as 5nm in other places on their web site. Either way, my point is proven: To declare it to *not* be 4nm is silly. It's as much 4nm as N5 is 5nm. (Which is to say, not at all, really, but pay no attention to that man behind the curtain. He's only 4.5nm tall, after all.)