I sincerely hope this doesn't turn out the way the announcements for Intel's 10nm and 7nm process nodes turned out. We were promised a lot of fantastic technology and got a lot of disappointment.

Even if Intel manages to adhere to this aggressive technological roadmap, it's still unclear to me what the market for x86-based machines is going to look like 3-5 years down the road. COVID-19-related PC sales are more of an anomaly than an actual trend; the knock-on effect though of that development promoting an embrace of remote work could possibly turn things around for x86, but I'm still left wondering why most people would want / need x86 when Apple's M1 chips clearly demonstrate that ARM-based solutions are not just sufficient, but potentially superior.

On the one hand, x86 is (eventually) going the way of the dodo for some of the biggest datacenter customers, which is a large part of the reason that Intel has once again opened itself up to the prospect of foundry.

On the other hand, x86 is so ubiquitous for PC that it would take a sea change for Apple Silicon to take over. Users and businesses are entrenched and this isn't the first time that Apple has run a different architecture from the rest of the PC crowd. I, for one, prefer my x86 MBP to the newfangled M1 because of RAM, compatibility, external display support, and number of ports. These limitations are all a function of implementation, not inherent to Apple Silicon, but they influence my preferences nonetheless.

COVID-19 has changed the way that work is balanced against everything else in our lives, and I view its impact on the chip market as demand creation rather than a demand spike. We will never go back 100% to the way that things were prior to the pandemic; big shocks all have that in common. The hybrid model of work from home as a complement to office work means that people/companies will need to buy and maintain more hardware.

I share your hope that this doesn't turn out the way the delayed nodes did. Intel's 10nm and 7nm setbacks started in the Krzanich days during the doldrums of the PC glut. People were mobile, phone innovations were booming, and no substantial PC advancements came to the fore during that time. I think we can give Intel some benefit of the doubt and assume that they aren't dumb enough to repeat the error of announcing a road map for which there is no road map...especially because the tick-tock model is already a thing of the past and exerts no expectations nowadays.

It is true that M1 is limited in terms of some functions such as memory capacity, external display support etc. But, you need to consider two things. 1. M1 is Apple's 1st iteration. M1x or M2 is coming around, and Apple could likely get rid of all of those limitations. 2. Apple's tight control over M1 means they can build all sorts of features into the SOC and fast unlike Intel which has to cater to heterogenous needs from its integrators.

Also, if you use a Mac with M1, you'd quickly realise how much of a work intel needs to catch up. Boy, it's so pleasant to use while being cool and quiet. Intel needs to run for the money and hope that no one figures out how to replicate Apple's success in the non-macos space and start eating intel's gold pot of datacenter CPUs. AWS Graviton?

> Users and businesses are entrenched

What dependency do you think most users and business have on the instruction set architecture, even if you ignore that you can translate?

I think you'd be surprised at how many businesses have proprietary applications that their resource stretched & budget starved IT orgs would not want to update.

Consumers probably less so but businesses will definitely care. Translation might be an option for some. However, Rosetta currently only supports limited translation [0].

[0] - https://developer.apple.com/documentation/apple-silicon/abou...

i bet if there is an arm version of the office suite it will be a no brainer. I've seen tons of stuff soley on ms office and maybe a thick client for an erp, which can be run via citrix or other app-v tech

> I think you'd be surprised at how many businesses have proprietary applications that their resource stretched & budget starved IT orgs would not want to update.

Written in assembly language?

Written in VB5. Or in C, but we're not sure where to find the source. Or expecting to be compiled by a 16-bit compiler and run in a DOS box. Or written in some off-brand interpreted language, and the interpreter only runs on x86. Or...

Finding the source isn't free. Finding the archived compiler isn't free. Finding a machine that still can run the compiler isn't free. Re-writing the app isn't free.

Doesn't dynamic binary translation cover these cases? It can't be that performance sensitive if it's written it VB5 and running on modern hardware.

It might. But if you've got a VB5 program that you're still running, are you the kind of person who knows about binary translation? Or do you just say "I'd better keep buying x86 and running Windows forever"?

At the low level you don't know what instruction sets are. At medium to high level you know that despite the instruction set, old non-intensive programs will work fine. I think the range where you're spooked by the instruction set is pretty narrow.

VB6 is still officially under support in Windows Server 2019.

More like their infrastructure is dependent on old versions of tools that they don't want to upgrade, can't build, and only work on x86. Hugely common.

There's an unbelievable amount of software out there for which porting is either not viable or not economical.

At work we use Delphi to produce Windows applications that rely on Win32 and GDI. While Delphi can be used to make say Android apps, if you're going for Win32 you're limited to the x86(-64) compiler.

A lot of the low-level stuff in standard libraries, third-party libraries and in our own code rely on platform details like calling conventions, which will almost certainly be different for a different architecture. Sure it can probably be rewritten, and a lot of code in the standard library already has generic high-level variants as fallback, but it's not going to be a simple recompile for us.

By extension then, our customers rely on the ISA as long as they rely on our application, and our software is key to their operation.

In many cases the dependency is native (not virtualized) Windows and Windows doesn't run natively on any good ARMs.

> it would take a sea change for Apple Silicon to take over

Not really.

Over my career I used x86, MIPS, PPC, SPARC, x86, and ARM machines. With the exception of x86 Windows boxes, they all looked and felt mostly the same. The Mac in front of me and the Linux laptop to my left incidentally use x86 CPUs but, at the core, they are both Unix-like machines and that's what I use them for. I don't care about binary compatibility and what ISA the machine uses inside it.

> On the one hand, x86 is (eventually) going the way of the dodo for some of the biggest datacenter customers

For the most part, I'm convinced the bulk of this transition has already happened. If you were running highly-parallelized code or needed to switch to specialty hardware, you've probably done so already.

> On the other hand, x86 is so ubiquitous for PC that it would take a sea change for Apple Silicon to take over.

Not just Apple silicon, but for ARM to take over. Even the best compatibility layers for x86/ARM still present a massive performance hit, and there's otherwise no real incentive for the bulk of PC owners to switch over. I say this as the kid who grew up with a first-gen Raspberry Pi on their wishlist: x86 is an immovable force, and it will be decades before it loses support in the mainline Linux kernel. Even if Apple, Microsoft, Amazon and every other Fortune 500 company made a mad-dash transition to ARM (which they won't), you still couldn't dethrone x86 in terms of support and install-base.

Sony and Microsoft consoles are running x86 hardware so demand is certainly not just driven by COVID-19, people would want those machines regardless. There is no current consumer option to pair an ARM processor with a high end graphic card. I don't know what the future holds but it's pretty obvious why people want x86 at the moment.

Not for consumers yet, but a demonstration recently made by Nvidia[1] involving ARM+dGPu(+Linux!) proves that world isn't as far away as you think. Assuming pure cloud gaming experiences such as Xbox 365 or Playstation Infinity don't become the only way to play console games over the next decade and half, An ARM-based Xbox or Playstation could happen by the next generation or the one following.

[1] https://www.gamingonlinux.com/2021/07/nvidia-shows-off-rtx-a...

Sure it "could" happen but they would actually need a reason to switch. They aren't selling Laptops where battery life matters.

These are the 10nm and 7nm processes. They're just retconning the names. 10nm+ is Intel 7, 7nm is now Intel 4, 7nm+ is Intel 3, 5nm is Intel 20A.

It makes sense. Intel marketing must be kicking themselves with all the press about them not being on 10nm while TSMC is on 5 even though 7nm is the best anyone but apple can get from TSMC and Intel's 10nm is (marginally)better than TSMC's 7.

So there is a weird factor here that we only know what intel tells us about 10nm. The actual parts that have shipped on 10nm have not been competitive with AMD parts at 7nm, or Apple at 5nm.

When Apple/AMD were shipping Intel beating parts folks were quick to dismiss it as a byproduct of TSMC's process. Is intel's chip design so much worse that AMD can trounce them on a substantially worse process? Or is Intel 10nm just not very good.

When the chips worked they had better perf/gate, but Intel had a lot of yield issues, making the perf/$ ratio worse.

Not true.

The 10 nm Intel Ice Lake Server has much worse perf/W than the 7 nm Epyc made by TSMC, and this is what matters.

That means that at the same power limit and the same number of active cores the Ice Lake Server clock frequency is much lower, or at the same clock frequency and same number of active cores the power consumption of Ice Lake Server is much higher.

The SuperFin 10 nm process used by Tiger Lake allows much higher maximum clock frequencies, up to 5.0 GHz, than the Ice Lake Server 10 nm process, and it also has a better perf/W, but even the SuperFin 10 nm does not reach the perf/W of the 7 nm TSMC process.

If the previous claims made by Intel are true, the process used for Alder Lake, now renamed "Intel 7" will be the first Intel process that is better in perf/W than the 7 nm TSMC process.

Nevertheless, it will be still worse than the "5 nm" process that has been used by TSMC in high volume products (Apple) for more than a year.

> The actual parts that have shipped on 10nm have not been competitive with AMD parts at 7nm,

Tiger Lake H closed the gap, it seems. The 11980HK is competitive with 5900HX.

Also, the first Alder Lake benchmarks on QS chips suggest it's faster than Ryzen 5950x. I believe the competition will be fierce from now on.

Sounds like the "equivalent speed" advertising war back in the day between Intel and AMD.

If architecture doesn't matter, why on earth would you think x86 is a negative. The most ARM can ever hope for is that people don't care or notice it's ARM.

Energy efficiency is a thing that some people (including phone manufacturers) care about.

As much as ARM is associated with efficiency, the instruction set itself is largely irrelevant to the performance of modern CPUs. x86s CPUs are translating x86 instructions into the CPU's native instruction set, and while that used to impose some non-trivial overhead, it's been trivial for quite some time.

>t's still unclear to me what the market for x86-based machines is going to look like 3-5 years down the road.

~750M of PC are used in Cooperate and Business. And if IBM mainframe has taught us anything business couldn't give a f what ISA they are using. They want 100% ( when possible ) backward compatibility. Considering most Business PC are already using low end x86 CPU, and they have not been CPU bottlenecked for long, I mean most of them are still using PC with HDD for all sort of reasons. So I dont see them switching any time soon.

Excluding that the vast majority of PC are Laptop and Gaming PCs. Gaming wont switch. Laptop is a complicated issue. And I think it sort of depends on Microsoft and how aggressive will Apple price their MacBook Air. I see no reason why they cant have the lowest cost MacBook starting at $799 while maintaining the same margin.

On Server I think x86 is pretty much done, at least in the HyperScaler market. The whole reason why Amazon is an Intel foundry customer is because Amazon are moving everything they could to Graviton. So you either loss Amazon as a customer leaving them to buy some x86 for some smaller part of their AWS, or you try and Fab them a chip they want.

And I believe, for political reason, Intel will gain lots of foundry customers.

So while a potential lost of 30-40% of their DC market in the next 5 years. x86 should still be a healthy state in 5 years time.

You hit the nail.

Potential fab clients look now at Intel, and keep guessing what's been going on with both their 10nm process, and their fab offering for the last 5 years.

Now people keep seeing Intel setting ambitious new targets, and keeping missing them.

Would they bet their tapeout on a fab like this now? No.

The least worst thing Intel can do for itself now is to disclose what in the world was actually happening which led to the current situation.

They long missed the window when their leading edge offer can capture first few clients ready to pay any premium for it.

Retail PC sales are only a small portion of the x86 market. The vast majority of corporate PCs will continue to use x86 in the foreseeable future for compatibility reasons. On the datacenter / cloud side providers are trying to make inroads with ARM, such as Amazon's Graviton, but are meeting similar resistance.

Additionally we have yet to see x86 on the TSMC 5nm process. Making that leap once Apple's reservations expire may remove much of the performance gap.

I think one big reason for this is how disappointing on the shelves ARM chips are compared with Apple's offering, and the lack of a standardized platform like the PC was for x86.

UEFI on AArch64 is a good start, but we really need to reach a point where a single image can be installed on every ARM desktop machine, like it has been for decades for Intel and AMD. Also, the lack of FOSS drivers for lots of boards mean that you risk getting stuck on a specific kernel version forever.l All of this is absolutely unacceptable, as long as there's x86_64 out there that's competitive and doesn't suffer from these downsides.

Absolutely. x86 standardization has made generations of developers forget what it was like coding for a specific machine. ARM is still far behind despite the enthusiasm.

TSMC Announced a 2nm soon after Intel announced it can catch up https://9to5mac.com/2021/07/28/tsmc-2nm-production-set-for-2...

> I'm still left wondering why most people would want / need x86 when Apple's M1 chips clearly demonstrate that ARM-based solutions are not just sufficient, but potentially superior.

Simple, I use Linux and MacOS cannot run the software that I need to use on a daily basis. The most powerful Apple Silicon you can buy right now still can't quite beat my 2014 beater-PC in raw performance, which also doesn't exactly instill confidence in it.

I have an M1 MacBook Air and I love it. But it's no replacement for my Threadripper 1950X / RTX 3080 for work/gaming.

x86 gives me choices. I can run Windows, Linux, etc. With M1...I'm at the mercy of Apple's closed-system. Sure, there are folks working on getting Linux working on the M1, but there's no guarantee that Apple will let you keep running/installing Linux. You're one Apple firmware update away from bricking your device/alternate-OS.

Who says Apple will be the only vendor of successful ARM desktops? It seems like it's only a matter of time before other vendors catch up. The point is simply that Apple demonstrated the viability of ARM in those use cases.

Windows on ARM laptops existed long before the M1 and they didn't catch on. It's Apple that is late to the party.

But like you said yourself, they didn't catch on. Therefore I wouldn't describe those vendors as "successful". Apple was the first vendor to achieve that. It may be that the Qualcomm chips used by the majority of those devices just aren't competitive enough to replace desktop use cases.

Apple Silicon != Windows on ARM.

Apple Silicon is a product line-spanning transition to home-cooked ARM CPUs. Windows on ARM was an experiment designed to run alongside x86, likely intended to fail. As for the "successful" bit, I couldn't frankly care less. You're welcome to measure success by whatever metric you choose, but the switch to ARM has ostracized me and many other developers I know (some of whom are Apple users currently).

In other words, ARM is good at watching Youtube videos at low wattages: who knew? (Answer: Everyone who used ARM for this exact purpose before Apple did)

Lol, way to play the game I guess. It's not Intel 4nm, but 'Intel 4', which was Intel 7nm and Intel 7nm+ is now Intel 3. Retconning Intel 10nm+ as Intel 7.

Makes you wonder if the same thing with the initial 10nm failures where a lot of management bonuses were tied to a specific number in the release is happening here too, or if it's just them trying to recalibrate those marketing literature when the competition is a node worse for the equivalent nm value.

Edit: It literally says this in the article:

2021 H2, Intel 7: Previously known as 10nm Enhanced Super Fin or 10ESF

2022 H2, Intel 4: Previously known as Intel 7nm.

2023 H2, Intel 3: Previously known as Intel 7+

2024, Intel 20A: Previously known as Intel 5nm.

Dishonesty aside, this hits like a ton of bricks: FOUR YEARS to leadership. Even with Moore's law slowing, that is a loooong time. I remember when Intel had seemingly a two year lead in process, around 65nm I think.

And now, project four years to leadership, and assuming the economics and competition won't respond.

If Moore's Law hits a major wall (it has to ... sometime, right?), then that may be the catch-up.

It won't even be leadership. TSMC 3nm is about equivalent to Intel 20A, and is hitting risk production by the end of this year.

It reads to me as sugarcoating the admission that they're continuing to fall behind.

It is not clear at all that these games work. In April SemiAccurate already warned:

Intel should not rename their processes

Analysis: An unwinnable fight should not be picked

https://www.semiaccurate.com/2021/04/07/intel-should-not-ren...

Intel is moving to GAA on 20A. TSMC is moving to GAA on N2. Just to put that into perspective

4 years is an acceptable timeline. We are not taking about software product roadmaps here. If you consider the amount of time and time and effort it took apple, amazon to come out just with chip designs, this is reasonable. While its true that tsmc “3, 4 nm” are ok track for 2022.. its debatable if that naming means what it actually implies. Also tsmc had priority access to euv solutions this time around.. lets wait and see if Intel can deliver before passing judgement outright.

Yeah, it’s dishonest, but the consumers of their processors most often look at benchmarks, not nm width. The only people this might actually fool will be casual stock traders.

Yeah, I don't blame them that much when tech news always had to be "TSMC 5nm.. essentially equivalent to Intel 7nm" when talking nodes. The numbers have been made up for a while (even on the Intel side).

Some of those numbers are very curious though. Intel has historically had an issue with managers gaming bonus structures and some of this reeks of that too. "We're totally shipping Intel 7 this year (because it's what was 10nm+)"

how is it dishonest? the numbers are basically arbitrary anyway (the only metric even close to these figures is fin width, which is not really a metric of primary interest), and they've chosen to realign to numbers that more closely match comparable TSMC nodes and acknowledge that it doesn't really measure anything by dropping the 'nanometer'.

Well, they didn't align with TSMC. They end up jumping ahead of TSMC by a node.

Angstrom's? I think the connection to actual feature sizes is long gone, and not sure intel 3nm is really comparable with TSMC? Anyone have a process comparison?

Also, key issue - ignore the roadmap and just start shipping you 10 and 7nm process stuff.

Seriously, every time I read about a TSMC process it seems to already be in risk production. Then in the next year all of a sudden Apple and friends are releasing nice product. The hamster wheel here with apple funding huge $$ into TSMC must be amazing as well.

With intel we've been hearing about 7nm forever, how the intel chips are going to crush (2x performance etc) AMD chips. But they never ship these things.

Those numbers refer to what a planar transistor would have to look like to achieve the same density. Of course nobody is actually using planar transistors anymore.

Are you sure? If so TSMC and Intel must have discovered different ways of calculating surface area densities…

Intel 3 is what used to be called Intel 7nm+, which is more or less equivalent to newer batches of TSMC N5.

Perfect - thank you!

I mean, this is nothing new. If you want newer processes, AMD does a pretty good job of keeping their Zen architecture on the cutting edge without under-engineering the end product. Intel's bag has always been about selling very specific things to specific audiences: bare-metal virtualization for people on Windows/MacOS (rip), super-wide SIMD busses, special overclockable models, IMEI, the list goes on. Intel, much like Apple, is selling a product that supposedly offers a better "experience" than their competitors, which moves like gangbusters in the corporate sector.

I wish Intel all the success, but we already know they're excellent at writing roadmaps. They just suck at executing them.

Frankly, this has me excited. I can't wait to see what new, exciting tech AMD invents to eat Intel's lunch, again.

So cool they are past nanometre now, Ångström!

x Nanometer or y Ångström isn't based on physical size anymore and is just a marketing term.

From the article:

>A lot of the industry, for whatever reason, hasn’t learned that these numbers aren’t actually a physical measurement. They used to be, but when we moved from 2D planar transistors to 3D FinFET transistors, the numbers became nothing more than a marketing tool.

> Intel 20A: Previously known as Intel 5nm. Moving to double digit naming, with the A standing for Ångström, or 10A is equal to 1nm

Was going to say wtf.. I could be half asleep but the numbers don't add up. 20a == 5nm != 10a == 1nm

> ... is just a marketing term.

Who are they trying to fool?

Terms like nanometer and Angstrom obviously don't mean much to average consumers, so these are not the target.

But engineers rightfully call bullshit on the use of these terms.

So who are they targeting with this?

I am a big Intel fan and floating point and single core performance have kept me buying. However I think that TSMC have the gap on Intel and that this is widely perceived to be true in Asia, which is where the manufacturing happens these days.

TSMC are ahead in part due to the yields they are getting from their EUV machines and the fantastic customers they are getting. Their business model is working out. Meanwhile Intel are shuffling a few top managers around which isn't the same thing as making the investment in people and processes that TSMC are doing.

The metric of single core performance has been the same for years from Intel and 11th Gen has come in for criticism because they didn't move to the next level process. I am rooting for Intel but the momentum is with TSMC.

Qualcomm is the company named in the press release as a partner for Intel 20A

Totally not going to be confusing to buy an "Intel 3" i7 or an "Intel 7" i3. :D. I'm a little sad they went "Intel 4" instead of "Intel 5" though.

The Intel 3 i7 13N65GK is pretty good, but the Intel 3 Xeon Platinum 9451 v6 is better for virtualization and stuff.

Do you mean you would prefer Intel mini Pro (late October 2022)?

20 ampere? That's a lot of current for a microprocessor.

It’s actually quite little. Most modern CPUs run at ~1V, which means 20A ~ 20W CPU. Ultrabooks are able to pull off 45W, so double of that.

You joke, but CPUs can draw hundreds of amps.

1 angstrom = 0.1 nm

10 A = 1 nm

That'd be Å not A

Showing my ignorance...

Is intel really competing with TSMC or with ASML? (ASML make the EUV machines TSMC uses)

Why can't Intel buy ASML machines and do what TSMC does? I understand there is a lot more that goes on around the ASML machines, but Intel should be able to do that bit right?

With TSMC.

One of the announcements here is that Intel will be getting literally the first of the new batch of ASML's machines.

They do buy from them (and invest in them). But as far as understand, ASML is one piece of the puzzle.

There are several other parts of the process that are not touched by ASML

Intel owns just under 5% of ASML. TSMC owns a bit of ASML as well. So does Samsung.

When I saw "20A" and "18A" I thought maybe they'd moved to using silicon atom diameters as a unit of measure, which seemed to make sense. Turns out it means 10^-10m and a silicon atom is actually ~2A diameter.

Ha, I first thought they meant “Ampere” (vaporizing the conductor at those numbers). Using “Å” would make it clear they mean “Ångström”. Someone should tell Intel that Unicode exists.

> Someone should tell Intel that Unicode exists.

How easy/hard is it to type accented characters on Windows?

On Apple systems you simply press-and-hold the letter, and you get pop-up with various options. Holding <shift-a> gives: À, Á, Â, Ä, Ã, Ã, Å, Ā.

They have an app buried somewhere that lets you select and insert Unicode characters from a scrollable table, but you have to find it first. In MS Word you can open that window from a menu command.

On Linux (and Windows too IIRC), you hit AltGr + accent key (number row on my layout) -> ÃǍÂĂÅÀĄÀȦÁÄ.