Hacker News Reader

Comment by ChuckMcM | Hacker News Reader

I found the news of Intel releasing this chip quite encouraging. If they have enough capacity on their 10nm node to put it into production then they have tamed many of the problems that were holding them back. My hope is that Gelsinger's renewed attention to engineering excellence will allow the folks who know how to iron out a process to work more freely than they did under the previous leadership.

That said, fixing Intel is a three step process right? First they have to get their process issues under control (seems like they are making progress there). Second, they need to figure out the third party use of that process so that they can bank some some of revenue that is out there from the chip shortage. And finally, they need to answer the "jelly bean" market, and by that we know that "jelly bean" type processors have become powerful enough to be the only processor in a system so Intel needs to play there or it will lose that whole segment to Nvidia/ARM.

View on HN

Forget all the technical problems, fixing Intel is mostly about figuring out how they let a bunch of MBAs mismanage the company to the point that they squandered a two-decade long global monopoly on computing, in an era where computing absolutely dominated global economic development.

I mean, when you put it that way, it's genuinely astounding that they managed to find themselves technically so far behind, and so quickly. This was not a minor mismanagement fuck up, this is complete incompetence. Things clearly need to be burnt to the ground, they need a lot of churn.

If they can get back to being engineering-driven rather than financially driven, it will all fall into place.

Once you invite the vampire into the house should you really be surprised that you find it sucking your blood? The answer is that like most large companies they decided technologist leadership no longer mattered. It’s a pattern that repeats again and again and that is because the people who need to learn the lesson most are counter-incentivized to do so.

Several conversations on this topic have brought up that pay rates at Intel are now less than the median of the industry. This would be fallout from actions by the MBAs in charge, but I don't think it's just an MBA issue. Something is clearly rotten in middle management in Intel. Either the people investigating the problems with their process weren't given the freedom to experiment to narrow down the scope of the problem, or they lost all their competent process engineers and are left with insufficient clue to fix the problem. If Intel has to rebuild their process group from scratch, they could be in for additional years of pain as new people come up to speed. Given that the United States is no longer a leader in bleeding edge process technology, this could take far longer than expected. How long will it take Intel to poach enough talent from TSMC to make a difference? And will they be able to integrate into Intel's culture?

The Israel group saved their behind in the days of the Pentium M. It seems like skunk works projects have been crucial to Intel, but the last decade has been a lot of tick tock strategy and not a lot of hard engineering...

I don't have any insider info, but to my eyes a lot of this change would be risk aversion caused by the Itanium disaster.

> figuring out how they let a bunch of MBAs mismanage the company

I think this is often stated as a cause, but more usually a symptom.

> fixing Intel is mostly about figuring out how they let a bunch of MBAs mismanage the company to the point that they squandered a two-decade long global monopoly on computing, in an era where computing absolutely dominated global economic development.

Isn’t the situation you described precisely where we would expect stagnation to be very likely?

As with many algorithms, coin flips might have been better.

> bunch of MBAs mismanage the company to the point that they squandered a two-decade long global monopoly on computing

I'm generally happy to blame MBAs for quite a bit. But isn't this also part of the nature of monopolies? A lack of competition means that in the short term, strategy doesn't matter at all. There was a period where Intel execs could have run the company via astrological chart or coin flip and done just as well for themselves. Managerialism and other MBA ideologies make that worse for sure.

It is strange that the MBA was such a coveted designation for such a long time. Is there literally no value in management?

I'm management, albeit not at Intel, and I don't have/have no interest in an MBA, so read in whatever biases you'd like. Or decide I'm full of shit, that's possible too. There is value in management, but management has to realize that it doesn't exist for its own sake. That's a tough one when "management is the company" is effectively what anyone's been taught for decades now. The problem is, I think, that it is an oversimplification. Nominally management makes the decisions, but we aren't where the product gets built or sold. The company is as successful as its products, and so maybe the product is the company, and we're all doing our part to make it the best thing that people want.

I'd like to think I'm a good leader. Probably not the best one - there are plenty of folks I look up to. Probably not the worst one, but that's not for me to decide. In any case, as middle management, [0] I see three major components to my role:

1. Hire and retain really talented people who are good at what we do;

2. Decide what we'll work on now, what we'll do later, and what we won't do; [1] and

3. Sponsor my team and provide air cover for #2

I generally have a pretty good idea of _what_ to do, but not always exactly. I used to know _how_ to do it, but I'm probably not the best at that today. My goal is to hire or develop a team that can put a fine point on the _what_, or challenge me when I'm wrong, and to own the _how_.

In terms of sponsorship and air cover, it's taking 1 and 2 above, and keeping it on the rails with minimal disruption. I see good middle management as needing to be very team-focused, caring for the health and composition of the team above all. Sometimes that means absorbing a lot coming from above, and sometimes it means being really pushy upwards to get what we need to be successful. But the team should be free to execute if we're doing our jobs well, and their end of the bargain is to deliver what we agreed to do.

This is not the philosophy of all middle managers, and I'm sure there are other models that work. This is the one I can speak to.

[0]: "Middle management" defined as someone who manages other managers but is not C-level.

[1]: This doesn't mean I cook this shit up. I'm listening to the team I manage, as well as the priorities and directives coming from above, and applying my filters. Gotta figure they hired me because I know generally what to do in my space, and the team agreed to work with me because I have the ability to point them toward the critical path and push the debris out of the way before they run into it. Without management, there are either too many ideas to harmonize, or you get lucky and the team converges on the right one. I think good middle management helps us converge on the right ideas faster.

MBAs teach you a load of useful skills, in particular a great breadth of business problems. HR. Accounting. Strategic planning. International expansion. Supply chain considerations. Micro and macroeconomics. Negotiation. The list goes on.

It doesn't make you experienced, or a leader. But it does give you the essential skills to run a business, that are so diverse that they are hard to acquire organically.

A good MBA is perhaps like a good coding Boot Camp. It will basically teach you enough that you can survive 3 months of a junior dev job, and if you're smart, bootstrap yourself up from that. But if you staff a company with no-one but people from such Boot Camps with no other experience, who will just then resonate against each other, that won't go anywhere.

About as much detail as I can learn from 1 week on coursera.

More often than not it teaches them the arrogance to believe they can do everything on their own and that they know the law because they have an MBA. I'm currently closely observing a sequoia invested startup run by a bunch of TOP 3 MBA program invested people.

Questionable legal decision, terrible HR, abusive bullying management. It's all a big circle jerk.

Let's not kid ourselves. The real thing MBAs learn in school is how to network and score their next company they will destroy.

Is there any experience to corroborate it, or is this your impression?

My wife is doing an MBA right now, on a top-100 uni globally; so not Harvard but good. I am genuinely impressed at the breadth and detail they go into.

As I say, these don’t turn you into experts or experienced business people, but from my perspective, very valuable.

It's my personal observation. In my experience there is a strong difference between people that study MBA for the sake of studying MBAs and people that are engineers or have other experiencing, want to create something but feel like they're lacking from a business perspective. Those people seem to do comparatively better, and seem to have better functioning companies if they choose to run them after the MBA.

Yeah, I'm not questioning this. If you want to start a company, you need a good product, and you won't magic one out from an MBA.

Equally, running a large corporation without people who understand so many complex business dynamics (and crucially, you'll need some who appreciate how they interact) seems crazy-suboptimal to me.

It's a mistake to think the only way to understand something complex is with an MBA.

A good example is Toyota versus the US Big 3 automakers. GM basically invented the modern American management the MBA came out of. Toyota was run by engineers. Toyota came out of the post WWII decimation of Japan to become the world's largest automaker, kicking the asses of US automakers in terms of cost and quality.

Toyota was happy to teach how they did it, and even set up a joint venture with GM, the NUMMI plant. The plant worked fine, but GM was unable to adopt the lessons. This American Life tells the tail well: https://www.thisamericanlife.org/561/nummi-2015

I think this is directly related to the MBA ideology, an important part of which is that you can take privileged people, give them a short general education covering a number of disciplines, and then make them part of the managerial caste that controls everybody else. A core notion of the MBA is that management is a universal skill. This is both historically novel and distinct from elements of the Toyota Way, which strongly honor the people doing the actual work and their accumulated knowledge.

MBA is not the only way to understand anything complex.

MBA is a good starting place to understand complex business problems. It won't make you a good engineer. And yet even engineering companies have supply chains to manage, contracts to negotiate, labour force to grow and keep motivated, finance deals to secure, multi-year strategic planning etc. I know for a fact engineers are, by and large, not good at that.

Doing an MBA won't magically make you good at that, but at least gives you a head start thinking about those problems and standard industry solutions (if you take more than arrogance and networking out of it). Similar to how an engineering/science degree doesn't tell you how to build a Tesla, but gives you the basics to bootstrap yourself from that point. Even if many a Stanford engineering grad thinks they'll build the next Decacorn because they are so great.

I really can't see what's controversial about it.

Lots of things give people a head start in thinking about those problems. But an MBA is a ticket to power.

If you really can't see what's controversial about giving power to a bunch of young people with a smattering of training and a very specific ideology, you haven't been looking at the harm caused. Maybe start there.

I think you're spot on with the arrogance bit. I've had friends who have done MBAs and some of them came out of it with an enormous performative confidence despite actually knowing much about whatever topic was under discussion. It was exhausting to be around until they learned to moderate it.

It reminds me of a friend's complaint about having done debate in high school. He says it took him years to unlearn the habits it gave him. Including the notion that whatever take first occurred to him in a discussion is one that must be defended at all costs. Or treating all discussions as gladiatorial to-the-death combat.

Learning that sort of faux certitude is certainly helpful in persuading people. So I get why people invest so much in learning it. But it seems to be to be double edged.

Huh. What's the source for your claims?

They teach you about those topics. But it's not clear how much they teach actual skills beyond finance and presentation, in that there's no actual business to run. One can read about negotiation, for example, but actual negotiation skills come from negotiating deals.

A friend of mine went back for a high-end MBA after a few years working for large companies as a (non-software) engineer. His basic take was that the curriculum only had one thing that he couldn't have figured out just by thinking about the topic for a bit. (It was the principle of comparative advantage, which is a bit of a mind-bender.) He says what he really got out of it was strong relationships with a network of people that were soon highly placed in a bunch of major companies.

As per my other comment, my wife is currently doing an MBA.

Again, skills are not replacement for experience, or even good ideas. They won't make you a pro in those things either.

It's a bit like saying a CS degree is useless, because it's impractical, and with programming, you learn by doing anyway. And you can't build a business out of a CS degree alone. Undoubtedly so, and yet a CS degree shows you how large the computer world is, so to speak.

And again, I don't think the MBA program teaches skills, the main exception being presenting confidently when you don't in fact know much.

And I'm not saying an MBA is useless. As I said above, it's very useful for establishing a network. It's also an entry ticket into a bunch of high-caste jobs and a thorough indoctrination into the ideology of managerialism. The starting salary for MBA is proof of its usefulness to their holders.

I agree there's some analogy with a CS degree. The main difference being that nobody puts a fresh CS grad in charge of anything important, because we understand they don't know much yet. An MBA, on the other hand, puts people in positions of significant influence despite them knowing little or nothing about the actual work. It's as if somebody went through the book-learning part of a medical degree and then was turned loose to practice on real patients.

> And again, I don't think the MBA program teaches skills, the main exception being presenting confidently when you don't in fact know much.

What can I say, this is not consistent with my experience. Perhaps it says more about the motivations and attitude of the people attending.

Perhaps a core confusion is thus: running a technology-based business requires, well, technological and business acumen. If you try to supplement one with the other, things go wrong, regardless of the direction. When I hear Toyota is ran by engineers, I'm sure the engineers don't negotiate finance deals, manage HR, career development, accounting etc. These boring things are important. Equally, when I hear about Boeing or IBM becoming corporate quagmires, whose core product is developed without a proper engineering input, I'm not suggesting this can work. My personal experience of companies ran by technical folk was, as said, awful.

The frustration about MBAs I hear in this thread seems to be when they encroach on technical territories they don't know about, and I agree this cannot work.

I think where MBAs work best is where you have mid-career people, already with experience, using this to boost their skills in diverse areas (alongside all other benefits like networking etc.), and I maintain it is great at that. If you take someone who doesn't know anything about anything, give them an MBA, then a random start-up to run, and it all goes wrong, I wouldn't blame the MBA.

I certainly agree the motivations of the people attending are a problem. But again, it's systemic. Locke and Spender cover it well.

You keep being sure about Toyota without any knowledge. That's very MBA of you, but I don't think it advances the discussion. The executives who turned it into a global power were engineers. And that engineering orientation was vital to the creation of the Toyota Way. Deeply influenced by W Edwards Deming, an engineer and statistician.

Nobody is denying that "boring" things like finance are important. Indeed, I'm saying the opposite. My point is that giving somebody a quick education that covers many "boring" things in a shallow fashion is not a great way to create executives. It is a great way, though, to create a glib executive caste that is good at extracting cash. Which is why executive comp has risen massively over the decades even though few would argue that executives are massively more competent.

You might be right that the MBA is better for people who have already developed actual expertise. But the average age of Harvard MBA applicants is 27, and only 12% are over 30. So if you're correct, then we can at least agree that most MBA grads shouldn't be in charge of anything.

You might enjoy reading "Confronting Managerialism", which looks at the history of the ideology and how it has become basically a caste system. Or read some of the Lean literature on how non-MBA-driven companies approach things. E.g., "Toyota Kata".

As an occasional manager, there's some value to the work that some managers do. But managerialism itself often extractive (that is, part of its function is to justify siphoning a lot of cash into the pockets of managers and execs) and value-destroying (both through optimizing for value-negative things and empowering bad decisions).

A good example is what private equity does to companies: https://www.nytimes.com/2009/10/05/business/economy/05simmon...

Or the way MBA ideology tends to look at everything through a quarter-by-quarter financial lens, leading to underinvestment in R&D, staff, etc.

"MBA" has become a sort of shorthand for everything that is wrong with modern American business culture.

The pattern is typically something like:

1) Management awards themselves a bunch of stock and sets annual cash bonus targets tied to short term financial metrics.

2) Management then focuses entirely on optimizing for short term metric attainment (e.g., quarterly EPS) and current stock price, by minimizing investment in R&D and new products, reducing worker headcount, engaging in financial engineering like stock buybacks, etc.

3) Management extracts as much value for themselves as possible through the above and then leaves before it all falls apart, to repeat the process somewhere else.

MBAs might be over-represented within this group, but this isn't something taught in MBA curricula, not all MBAs are engaged in this abhorrent behavior, and non-MBAs do it as well.

> they squandered a two-decade long global monopoly on computing

This conclusion is extremely premature. Intel has greater market share now than it did in the Athlon 64 days.

Isn’t it obvious they are at the cliff of total collapse? They’re still flying, but all engines are off.

Why, because they are one node behind the competition?

Intel pulled in revenue of $77.9 billion last year. I don't think total collapse is a realistic outcome.

Investors don't think so: stock price is twice what it was 5 years ago.

While I largely agree, it is also true that running a business requires very different skills than building its product.

MBA or not, technologist or not, you can still be good or bad at what you do. Smart or stupid. I think simply looking at a mismanaged tech company and saying “ah too many MBAs in charge” is missing the point.

My guess is, if you look at the well-ran large companies, these will also have an MBA-rich management core.

I worked at a mid-sized company completely driven by technological people, with no one having a solid business background. It wasn’t good.

> I mean, when you put it that way, it's genuinely astounding that they managed to find themselves technically so far behind, and so quickly. This was not a minor mismanagement fuck up, this is complete incompetence.

Is it though?

Or is it just the hardware of equivalent of the risky move of throwing it all away and rewriting from scratch for version 2? Intel have just incrementally improved their design, whereas AMD threw theirs away and almost disappeared in the process. In a parallel dimension, Intel could be exactly where they are now, with no AMD.

But... I agree that there was/is clearly a huge amount of incompetence letting non-engineers run an engineering company.

I wonder.

Large corps do gather intelligence on the competition. Even small ones do.

For example, when a new pizza place opened in town, the one I worked for when in College, immediately ordered one to see what they were like, pricing, taste, etc.

Large corps go way beyond this, and information does leak. Bribes, promised promotions and roles at the new company, theft of code, processes, trade secrets.

You do not have to use trade secrets, IP, or processes in house, and can silo that info from engineers / the main company. Yet having an awareness of these things, gives you an idea of how much to spend on research.

I wonder, did Intel not do any of this? Did they do so, but AMD security/counter espionage worked to provide a fake picture?

I wonder how large corps operate in this realm, these days.

I think they should replace every person in the hierarchy with a randomly-selected subordinate.

>how they let a bunch of MBAs mismanage the company to the point that they squandered a two-decade long global monopoly on computing

A MBA Story.

Once upon a time, a MBA joined a company, worked his way from Finance Director to C grade. He could probably be described as pure breed MBA. He made friends with other rare MBA in the company and improve their career path along the way. Why? Because that is what you do, MBA always help another MBA. Somewhere along the line, the company's first non-technical / engineer MBA CEO was born.

As the company grow there are more MBAs, but not all MBAs are bad, you just need engineers and product people around these MBA. But some of these engineers or product genius [1] are pain in the ass to manage, and since many of them dont have MBA, they dont understand MBA speak and they get driven out of company decision making forum. So what happened was these product engineers got rotted out over the years even when they are C grade and out very publicly.

With leading edge tech company there is a very long lead time in their product R&D roadmap, so they were fine for the first few years, and continue to make record breaking results. Then MBA move his way up to the board and become chairman just in the time before his MBA CEO friend retired. Which means MBA Chairman gets to pick a new CEO again, guess which CEO would a MBA pick? Another MBA! Since MBA always help another MBA, new MBA CEO promoted more MBA to various rolls within the company. And MBA cult now officially dictate the company decision making.

Everything was great for a few more years, the market was growing, they have a monopoly and most important of all, they are doing great because they are MBA. At least that is what those MBA thought. Until one day cracks started appearing, MBA CEO thought it was all fine, but with time external pressure were piling in and the whole industry moved on. The market looked very differently to what company were 10 years ago. And somehow the MBA CEO already had an on going internal investigation which shows he had an inappropriate relationship and was let go. Which is the MBA speak of CEO being fired. And a recently arrived CFO was named as interim CEO while MBA Chairman and Board continue their CEO search. They could hire other MBA from outside the company, but since the leading edge tech company are a gigantic company with so many different specific domain knowledge and culture. The suitable choice outside the company are virtually nil. MBA Chairman looked at CFO, despite not being his first choice and is fairly new to the tech company, but he has a MBA! And the third MBA CEO was born.

The market is moving at a much faster pace than before. To give credit to the new MBA CEO he was trying very hard to steer the ship to a new path and direction. The tech company are already behind. These new direction and requirement are completely new, MBA middle management have no idea what they are doing and cant turn the ship fast enough. And while they are still turning, the industry has moved further along. I guess this is an example of MBA being killed by MBA.

Finally shareholder were so fed up, some of them asked the same question;

>how they mismanage the company to the point that they squandered a two-decade long global monopoly on computing, in an era where computing absolutely dominated global economic development.

There are very little choice of what shareholder could do. They need those product engineers back. And a new CEO. But the perfect candidate had a fallout with MBA Chairman a decade ago, the only option was to rally enough support to get rid of him. In the end, MBA Charimand retired, a new Board Chairman was named, and those product engineers are back to save the company.

As far as the MBA story goes, this is the end. We dont know what happen to the leading edge company, I guess that part is To be continued.

[1] https://www.youtube.com/watch?v=lmFlOd0MGZg

The Amlogic S905X3/4 are "jelly bean" chips under this definition - you can buy full computers with those including a motherboard of sorts, RAM and standard USB/HDMI/Ethernet/eMMC/Wifi for under $30 bulk in China. They're branded as "TV boxes". The S905X3 is also the SoC of the $55 Odroid C4 SBC.

Yet those chips are apparently made on either TSMC's or GloFo's 12nm nodes (which one I can't find). Whilst Intel's 14nm is still superior to either TSMC/GloFo 12nm by all accounts, they're definitely the same ballpark. Competition for Intel even looks rough for cheaper high volume chips. You'd need to consider that Intel still needs some time to adapt to being a third party foundry too.

Good analysis, I didn't say it would be easy :-).

Why do you think Intel will succeed in selling leading edge fab services to fabless companies when GlobalFoundries (the chip making business of AMD) was unable to compete?

To me, it seems like the same story is playing out again.

One could easily flip that question around right? "Why would Intel try to sell fab services like Global Foundry which didn't work out well?"

But more importantly, Intel actually has way more foundry expertise than Global did/does. In the case of AMD selling off Global it was more like cutting loose an under performing part of the company rather than empowering the a significant asset of the company.

Production for a datacenter CPU is not the same as production for datacenter + enthusiast-grade consumer CPUs like Zen 3 currently achieves, unfortunately. Rocket lake being backported to 14nm is still not a good sign for actual production volume, although it probably means next generation will be 10nm all the way.

Datacenter CPUs are much larger than consumer parts and yield goes down with the square of the die area. They start with these because the margins go up faster than square of the die area.

But modern techniques exist to deal with problems in a large die (ie testing and then segmenting off cores with mistakes on them), so the fact they’re starting with large chip die sizes doesn’t really tell you much, no?

The top end most profitable SKUs are fully enabled dies. That they are now able to ship dies this large is a good sign. The 10nm laptop chips they have produced so far were rumored to have atrocious yield.

"The top end most profitable SKUs are fully enabled dies." If you take a part that would be discarded because it didn't make "top end" binning, and sell it as a lesser(core count or clock speed) chip. Isn't that profit from what would otherwise be scrap?

Yes. But that’s a silver lining to a black cloud not the best case scenario.

My impression is the vast, vast majority of chips have at least part of the chip disabled to help solve yield problems. So it seems to be the status quo?

Die harvesting is not perfect. If there is an error in a logic transistor, they can turn off that cpu core and use the rest of the chip. If there is a short between ground and power, the whole chip is useless because it will melt when you plug it in.

Yields on larger chips are still substantially worse than smaller chips, even with harvesting.

Interesting, that’s informative. Well that sounds like a big advantage for the chiplet approach.

And here I always thought it was because, should your large die have an issue, you can likely sell it as a lower-binned SKU with the failed area gated off. This way you end up with more inventory of a chip you intend to sell, or you have a high margin chip - win-win!

I assume for intel, they make more server CPUs per year than the entirety of AMDs output.

There is a slow but seismic shift to AMD within data centers right now.

They still have something like 40% to go to even reach parity with Intel though

It’s harder to go from 0.1% to 10% than it is to go from 10% to 50%.

Yes and no. It’s harder to convince folks to switch going from 0.1% to 10%. But you are also talking about winning the low hanging fruits.

once you get those easy wins the rest take a bit to chip away at.

while initial growth was 100x, that pace is going to slow down as marketshare increases.

Why not ARM? ARM is now penetrating the consumer development devices (M1), why not mirror that in the servers as well?

Because M1s are light years ahead of anything else, and you can't buy M1s outside of a Mac.

Intel has a momentum and some cult following, but they're no match for AMD in certain aspects like PCI lanes, memory channels, and some types of computation which favors AMD's architecture.

Day by day, more data centers get AMD systems by choice or by requirement (Oh, you want 8xA100 nVidia made modules with maximum performance. You need an AMD CPU since it has more PCI lanes, for example).

You don't see much AMD server CPUs around because first generation and most of the second generation has completely bought by FAANG, Dropbox, et al.

As the productions ramps up with newer generation, we can buy the overflowing parts after most of the production is gobbled up by these buyers.

In terms of PCI lanes efficiency, there is no competition between Intel and AMD. Intel is much ahead wrt AMD. Don't not be impressed about number of lanes available on the board.

> Don't not be impressed about number of lanes available on the board.

When you configure the system full-out with GPUs & HBAs, the number of lanes becomes a matter of necessity rather than a spec which you drool over.

A PCIe lane is a PCIe lane. Its capacity, latency and speed is fixed, and you need these with minimum number of PCIe switches to saturate the devices and servers you have, at least in our scenario.

I can only assume you're referring to Intel's Rocket Lake storage demonstration they tweeted out. This was using PCMark 10's Quick Storage Benchmark which is more CPU bound than anything else.

All of the other benchmarks in the PCMark test suite push the bottleneck down to the storage device.

One would think Intel might want to build a storage array that could stress the PCIe lanes but then that might show an entirely different picture than the one Intel is portraying.

What is PCI lane efficiency? I thought they were fixed speed and bandwidth.

Either put up the real reason to be impressed by Intel's much, much delayed PCIe 4, or stop shilling for them.

PCIe efficiency is such buzz-none-word.

> Rocket lake being backported to 14nm is still not a good sign for actual production volume

I'm genuinely having trouble understanding what you mean by this.

Rocket Lake being backported to 14nm means that 10nm can be allocated in greater proportion toward higher-priced chips like Alder Lake and Ice Lake SP. Seems like it would be good for production volume.

I think they mean that the fact that they needed to backport Rocket Lake, versus just having all their production on 10nm, implies a much more limited production capacity than the other situation.

Production volume referring to production volume of the node still being low as a whole not risks to the being able to get volume of these particular SKUs using the node.

I was under the impression the issue with 10nm is frequency, which lead to the rocket lake backport. Unfortunately it seems that the 10nm node's efficiency point is lower on the frequency curve. The reviewed ice lake processor was 300mhz lower then the previous generation (though with much higher core count), despite higher power draw and the process node shrink.

In laptop processors they can easily show efficiency gains from the 10nm process and IPC improvement, It appears most laptop processors end up running a power envelope lower then the ideal performance per watt efficiency. Server Processors with higher core counts means you can run more workloads per server, again providing efficiency gains. However desktop / gaming tends to be smaller core count + higher frequency with little concern of efficiency outside of quality of life factors (i.e. don't make me use a 1KW chiller). Intel has been pushing 5ghz processor frequency for years, and rocket lake continues that push (5.3ghz boost), when they drop frequency to move to 10nm, its hard to see an IPC improvement that is able to paper over that.

However alder lake CPUs will have a thread count advantage, so at least with 24 threads it should be able to show generational improvement over the current 8c/16 rocket lake parts. That will allow them to at least argue their value with select benchmarks and intel only features. Those 8 efficiency cores will likely be a BIG win on laptop, but on desktop I doubt they will compare favourably to the full fat cores on a current Ryzen 5900x (i.e. a currently available 24 core processor).

Intel is going to have at least 1 more BAD mainstream desktop generation before they can truely compete on the mainstream high end, however there is a chance they have something like a HEDT part that would allow them to at least save face. That being said, given a choice, Intel will give up desktop market share for the faster growing laptop and server markets.

> Rocket lake being backported to 14nm is still not a good sign for actual production volume,

I'm not seeing a good reason for thinking this is the case. Server CPUs are harder to fab (much larger die area) and they need to fab more of them (desktop CPUs are relatively niche compared to mobile and server CPUs).

If anything this is a sign that 10nm is fully ready.

What's a "jelly bean" processor? Trying to search for that just gets a bunch of hits about Android 4.1.

https://news.ycombinator.com/item?id=17376874

> [1] Jelly Bean chips are those that are made in batches of 1 - 10 million with a set of functions that are fairly specific to their application.

Is that the chip-on-board packaging like described here: https://electronics.stackexchange.com/questions/9137/what-ki... ?

Sometimes referred to as "applications specific processor" (ASP) or "System on chip" (SoC). These are the bulk of semiconductor sales these days as they have replaced all of the miscellaneous gate logic on devices with a single programmable block that has a bunch of built in peripherals.

Think Atmel AtMega parts, there are trillions of these in various roles. When you think of something like a 555 timer[1] that is now more cost effectively and capably replaced with an 8 pin micro-processor you can get an idea of the shift.

While these are rarely built on the "leading edge" process node, when a process node takes over for high margin chips, the previous node gets used for lower margin chips, which effectively does a shrink on their die increasing their cost (most of these chips seem to keep their performance specs fairly constant, preferring cost reduction over performance improvement.)

Anyway, the zillions of these chips in lots of different "flavors" are colloquially referred to as "jelly bean" chips.

http://sparks.gogo.co.nz/assets/_site_/downloads/smd-discret...

> Jellybean is a common term for components that you keep in your parts inventory for when your project just needs “a transistor” or “a diode” or “a mosfet”

-----------

For many hobbyists, a Raspberry Pi or Arduino is a good example of a Jellybean. You buy 10x Raspberry Pis and stuff your drawer full of them, because they're cheap enough to do most tasks. You don't really know what you're going to use all 10x Rasp. Pi for, but you know you'll find a use of it a few weeks from now.

---------

At least, in my Comp. Engineering brain, I think N2222 transistors or 3904-transistors, or the 741 Op-amp. There are better op-amps and better transistors for any particular job. But I chose these parts because they're familiar, comfortable, cheap and well understood by a wide variety of engineers.

Well, not the 741 OpAmp anymore anyway. 741 was a jellybean back in the 12V days. Today I think 5V compatibility has become the standard voltage (because of USB). So 5V op-amps are a more important "jellybean".

I don't know how old you are, but the 741 was obsolete in the 1980s. It sticks around in EE textbooks because it's such an easy way to demonstrate problems with op-amps... high input current, low gain-bandwidth product, low slew rate, etc.

I think your jellybean op-amps would more likely be TL072, LM358, or NE5532.

Old, beaten up textbooks from the corner of my neighborhood library was talking 741 back in the 2000s, when I was in high school and started dabbling around with electricity more seriously.

Maybe it was fully obsolete by that point, but high school + neighborhood libraries aren't exactly filled with up-to-date textbooks or the latest and greatest.

I remember that Radio Shack was still selling kits with 741 in them, as well as breadboards and common components... 12V wall-warts and the like. Online shopping was beginning to get popular, but I was still a mallrat who picked up components and dug through old Radio Shack manuals into 2005 or 2006.

It was the ability to walk around, and see those component shelves sitting there in Radio Shack that got me curious about the hobby and start researching it. I do wonder how modern children are supposed to get interested into hobbies now that malls are less popular (and electronic shops like Radio Shack are basically disappeared).

------------

I don't remember what we used in college. I knew that I was more selective and understood the kinds of problems various OpAmps had back then. Also you're not really rich enough to invest into a private stockpile of chips, and instead just use whatever the labs are stocked with in college.

LM358 is the jellybean that I keep in my drawer today. If you're curious. Old habits die hard though, I still think 741 as the jellybean even though it really is obsolete today.

I've got a tube each of 358's and 1458's (dual version) in my parts supplies. But my microwave stuff is finding them lacking.

+1 for LM358

That's the one my text book often referred to.

Yeah, I got about a hundred of them for $5 total when my Radioshack went under. Very jellybean.

I'm not sure if this is the same as what you mention with 12V vs 5V, but Intel is pushing a 12V standard PSU design that is also used by the new Nvidia Founders Edition GPUs:

https://www.intel.com/content/dam/www/public/us/en/documents...

My interpretation: Small buck converters are cheap and efficient, and so many parts need different voltages anyway (1.8V, 1.5V, 1.2V, etc.), that what you want to do is distribute the 12V from the main PSU and then locally convert it to the voltage you need.

If you are already putting buck converters everywhere, it makes sense to raise the voltage coming from the PSU, because you can reduce the current supplied by the PSU. Only the power supply will be running at 12V, almost everything else will be running at a lower voltage. (Would you rather supply 120W at 12V/10A or 5V/24A?)

So when you work with analog electronics, you learn that voltage is your friend. Playing with +/- 12V (a 24V range) for "analog logic" is innately cleaner than playing with 0V to 5V (5V range).

Lets say you make a design that has +/- 1V error. On the 5V circuit, that's 20% error. But on the +/-12V circuit, it is only 4.1% error. (Going much better than 5% or 1% error is nonsensical. Most resistors you'll buy are 5% error, and capacitors are maybe +/- 20% error). So you can see, +/-1V error is acceptable on +/-12V circuits, but maybe unacceptable on 5V circuits.

If precision is important, there are 1% or 0.1% resistors available, as well as matched-resistors (which have say 1% error, but all the resistors have the same degree of error, so your "proportions" remain consistent. You can manually-match resistors together with an accurate ohm-meter as well)

As such, +/- 12V is simply easier to make "precise" electronics on. Of course, the downside is that you've got leakage all over the place (more voltage means more power-draw)

---------

But now its 2021. Most parts, even extremely cheap "Jellybean" parts like the LM358, have low errors and low-bias. And USB's popularity as a 5V delivery mechanism has grown, everyone has a USB plug somewhere to use as the basis of electricity experiments.

So while electronic engineers way back played with +/-12V, today the assumption is that you play with 0V to 5V.

What I don't understand is: ASML is building these machines for making ICs. Why can TSMC use them for 7nm but Intel can only use them for 10 right now? Doesn't ASML make the lenses as well so that you're "only" stuck making the etching thingy (forgot what it's called, but the reflective template of a CPU).

It seems like nobody is talking about this, could anyone shine some light?

Consider that the wavelength of red light is 700 nm, and the wavelength of UV-C is 100nm to 280nm.

And immediately, we see the problem about dropping to 10nm: that's literally smaller than the distance that photons vibrate on their way to the final target.

And yeah, 10nm and 7nm is a marketing term, but that doesn't change the fact that these processes are all smaller than the wavelength of light.

-------

So there are two ways to get around this problem.

1. Use smaller light: "Extreme UV" is even smaller than normal UV at 13.5nm. Kind of the obvious solution, but higher energy and changes the chemistry slightly, since the light is a different color. Things are getting mighty close to literal "X-Ray Lasers" as they are, so the power requirements are getting quite substantial.

2. Multipatterning -- Instead of developing the entire thing in one shot, do it in multiple shots, and "carefully line up" the chips between different shots. As difficult as it sounds, its been done before at 40nm and other processes. (https://en.wikipedia.org/wiki/Multiple_patterning#EUV_Multip...)

3. Do both at the same time to reach 5nm, 4nm, or 3nm. Either way, 10nm and 7nm is the point where the various companies had to decide to do #1 first or #2 first. Either way, your company needs to learn to do both in the long term. TSMC and Samsung went with #1 EUV, and I think Intel though that #2 multi-patterning would be easier.

And the rest is history. Seems like EUV was easier after all, and TSMC / Samsung's bets paid off.

Mind you, I barely know any of the stuff I'm talking about. I'm not a physicist or chemist. But the above is my general understanding of the issues. I'm sure Intel had their reasons to believe why multipatterning would be easier. Maybe it was easier, but other company issues drove away engineers and something unrelated caused Intel to fall behind.

Thanks, the multipattering choice was an interesting thing to take home regarding why X is ahead of Y.

Could you not simply take the 2D fourier transform of the desired on-chip pattern and use that as the mask?

It seems like the diffraction pattern of the "long" wave laser would give you exactly what you want on the chip... and if you are putting hundreds of chips on a single wafer, it seems like you might not even need to worry too much about ringing at the edges.

You might be confusing the somewhat similar shapes arising from the Fourier transform of step functions etc. or from diffraction from an individual hole or obstacle. Unlike the Fourier transform, which is a purely mathematical operation, diffraction has no "inverse": if you add more diffraction you just add more waves to the projected pattern.

The only ways to defeat diffraction are reducing it with shorter wavelengths and compensating it with multiple exposures with different patterns in which light and dark fringes compensate each other: exactly the two general approaches (EUV and multipatterning) taken by the semiconductor industry.

I don't think I am, actually.

In the far field with a narrow bandwidth coherent light source (i.e., a laser), the projected image should be the FT of the aperture. That limiting condition is sometimes known as Fraunhofer diffraction, and generalizes to arbitrary apertures (not just a single hole).

Consider a narrow-band laser incident upon a diffraction grating for example. It produces a single point (well, two or three points, mirrored across the grating), not the uniform smudge that you'd expect by naively adding up the diffraction patterns of a bunch of slits. You should actually try this experiment for yourself!

The only trick is that you need a collimated laser and you need it to illuminate the entire inverse-FT-chip-grating aperture at once.

This deck has some nice examples of multi-hole apertures on slides 20 and 22: https://www.brown.edu/research/labs/mittleman/sites/brown.ed...

I'll take a crack at it, though I'm only in undergrad (took a course on VLSI this semester).

Making a device at a specific technology node (e.g. 14nm, 10nm, 7nm) isn't just about the lithography, although litho is crucial too. In effect, lithography is what allows you to "draw" patterns onto a wafer, but then you still need to do various things to that patterned wafer (deposition, etching, polishing, cleaning, etc.). Going from "we have litho machines capable of X nm spacing" to "we can manufacture a CPU on this node at scale with good yield" requires a huge amount of low-level design to figure out transistor sizings, spacings, and then how to actually manufacture the designed transistors and gates using the steps listed above.

Thanks, that clarifies it as far as possible for me who don't really know how the full manufacturing process works.

Could we simplify this roughly into "ASML makes the machines to shine light at the right nm, foundries makes the Silicon and packages it into an useful device, architecture designers give you the layout to etch" if my mom were to ask?

A few possible corrections, but one interesting one: the output of a foundry is generally known-good dice -- sliced up and tested bits of a silicon wafer. Packaging is done by IC packaging companies, and is really quite an interesting field in its own right. (Many of the foundries do some packaging in house, especially for more exotic approaches like Intel's Foveros, but this is the exception.)

I would've guessed that they packaged everything inhouse. I'm fascinated by this industry, it's so incredibly omnipresent products, yet "we" know "nothing" about the process of going from "sand" to gigaflops in your machine.

We talk to the piece of sand with lightning (USB Keyboard 5V interface), and then we hear its response back with lightning (HDMI Port interface). And somewhere in the middle, there's a source of lightning that's making it think on its own (Power supply and VRMs).

EDIT: More seriously though: https://www.youtube.com/watch?v=NGFhc8R_uO4

Video saved to watch later, thanks mate. Though I must say I like your first unedited answer too, so I'll put it back here.

> Its pretty simple. We zap sand with lightning until it starts thinking for us.

Lol, thanks.

I think I kept working the joke in my brain, and it was too brutally simple at first. And then I worked it over, and now it reads too complicated. Ah well.

I think I was overthinking the joke.

TSMCs 7nm is roughly equivalent to intels 10nm, the numbers don't really mean anything (anymore) and are not comparable

Like others have said, the numbers aren't comparable between manufacturers.

Here's a neat video where they use an electron microscope to actually compare the transistor sizes for Intel 14nm and AMD 7nm: https://www.youtube.com/watch?v=1kQUXpZpLXI

Thank you, this is not what I was looking for though. Even though they're equivalent TSMC was there way before Intel.

For one, "7nm" and "10nm" are just marketing names at this point, and don't really correspond to the physical dimensions of anything on the chips produced. Intel 10nm and TSMC 7nm are considered to have very comparable density. TSMC 5nm and is quite ahead though.

As to why it's not just a matter of buying a bunch of ASML lithography machines and plugging them in: In addition to what the other replies have noted, there is so much complexity and precision required in a fab. Consider all of the challenges that would be involved with starting with a bunch of industrial robots, and trying to build a fully automated assembly line that manufactures cars. Then scale precision requirements up by many orders of magnitude.

Intel tried and failed at jelly beans at least once in recent history: https://en.wikipedia.org/wiki/Intel_Quark

I'm not sure that is a worthwhile segment for them to try to compete in though. It's high mix, has a ton of competition, and doesn't necessarily leverage their fab strength.

Forgive me for my ignorance, but what is a "jelly bean" processor? Do you mean the "extremely cheap, almost no power consuming" processors?

If they price it right, it could be amazing. Computing is mostly about economics. The new node sizes greatly increase the production capacity. Half the dimension in x and y gets you 4x the transistors on the same wafer. It is like making 4x the number of fabs.

It also has speed and power advantages.

I think this release is excellent news on many levels.

Intel 10nm is really just a marketing term at this point and has nothing to do with transistor density.

I know there is some truth to what you're saying, but if that really was the full story, why wouldn't it be "Intel 2nm"?

Presumably because people aren’t buying chips based on the node size used and they would get bad press from calling it 2nm. My guess is it simply relates to the node refreshes in the old pattern of old node / ~1.414 = new node size. With different companies picking different arbitrarily differences in nodes aka intel’s 14nm++ was better than it’s initial 10nm process.

That said you can read some interesting details here: https://en.wikichip.org/wiki/7_nm_lithography_process

Reply on news.ycombinator.com