Apple Silicon: The Passing of Wintel
mondaynote.com
mondaynote.com
The US already lost the DRAM industry, the disk industry, the display panel industry, most of the small component industry, and the consumer electronics industry. In ten years, the US won't be able to make electronics in volume.
[1] https://en.wikipedia.org/wiki/List_of_semiconductor_fabricat...
Or will that happen, just further down the line?
It still remains to be seen how successful they would be, but at least they recognize this situation as a problem in nearest future.
I wish the US had seen this with equal clarity 30 years ago.
Step 2 is to steadily increase prices once they have a manufacturing monopoly, and impose tariffs on imports.
More like it transferred that wealth to the managerial class, while the US worker is flipping burgers for minimum wage now.
We're too busy arguing over irrelevant stuff, unfortunately.
Some people did. Even many politicians did. But they were shouted down as being racist and protectionist, and "globalism" was the future. Anyone who didn't get on board was called old and stupid.
IMO if a foreign adversary spends 20+ years and tens of thousands of dollars (at a minimum) training a highly skilled worker we would be insane not to take that skilled worker for free, if they want to move and work here.
Limiting immigration of highly skilled workers is a massive national security failure. It's a huge mistake. If a Chinese or Russian top tier engineer / AI researcher / wants to immigrate to the US we should roll out the red carpet because they're a strategic asset.
In Canada the only party to consistently oppose NAFTA and the FTA was the NDP, the left wing party.
In my experience, the right was very opposed to NAFTA (while the left was pushing it), in favor of free trade but not at the expense of jobs and manufacturing ability in the USA, and thought all these deals with China were going to enrich a possibly future adversary. By experience, I simply mean social interaction with middle class conservatives/Republicans, not experience in politics or "the party". And, I didn't really start paying attention to what was going on until the early 90s. (edited for a paragraph break)
: liberal is not considered "Left" outside of the United States
I see this repeated as a talking point by multiple individuals. I don’t really get what the point of the comment is or what I’m supposed to learn from it or do about it. It feels pretty close to a Scotsman but to what end?
And because Americans are being lied to by their own politicans, and part of that lie is that they have a left wing opposition. They don't, really, apart from the Sanders wing of the Democrats. That Fox News or whatever in the US can with a straight face call Biden "left wing" isn't just a political vocabulary corruption, or a relative shift of frames of reference -- it is an ideological lie, and the purpose of it is to advance an agenda.
And because it would help if Americans looked outside of their own borders at the rest of the world, and gained frame of reference from that.
I'll just quibble with this and say that if we're talking about US politics then we are talking about the same thing when we speak of the "left" in the US, because that's what it is regardless of the state of the rest of the world.
I also am not sure there is much value in arguing about how much more "left" some countries are. The fundamentals aren't really different until you talk about communists (and if that's what you're talking about, I abhor communism and don't want that in my country). If you're talking about universal healthcare for example, you're not really talking about something radically different between "more left countries" like the colloquial Sweden, it's just a matter of degree of difference.
The reason this matters is because as long as Americans accept this false polarity they are missing choice. Not just because they're a two party system, but because their two parties are defined along an ideological axis that misses entire policy choices. And Americans have been mistakenly educated to believe that "socialism" = "government involvement", and that it is an extreme 'left' pole of the political spectrum, and as such they are increasingly ending up with dysfunctional state systems that do not serve the populace.
That and being able to speak to the rest of the world about politics is why clarity of terms here is, I think, important
I think universal healthcare is a suitable minimum definition of left wing despite that, since a minimum definition is “without this element, a thing cannot be left wing” not “with this element, a thing must be left wing”.
> Universal healthcare is off the table by both parties in the coming federal election.
No, it's not, whether you are referring to the Presidential or Congressional elections, both of which are federal elections in and, unlike in most parliamentary systems, are separate-though-concurrent elections where the candidate platforms even from the same party have no necessary alignment.
> Biden is against it,
No, he's not.
> so is he left wing?
No, he's a center-right neoliberal, just like the rest of the dominant faction of the Democratic Party, and up through at least the early 90s the dominant faction of the Republican Party, too.
A sizable minority of the Democratic Party, though, is (mostly center-)left, though, and they are a not-insignificant factor in Democratic policy stances currently (US major political parties are effectively multifaction coalitions, and just like any coalition the policy preferences of the dominant faction aren't always those of the coalition.)
> (No, sorry, the "public option" tack onto Obamacare doesn't count)
While I don't particularly like it, being part of the more-left-than-Biden part of the Democratic Party, the public option proposed by Biden as recently fleshed out by the Biden-Sanders joint policy group is, in fact, a proposal for universal coverage, not to dissimilar from some other OECD countries, all of which have universal healthcare but far from all of which have public single-payer like Canada's Medicare or public provision like the UK NHS.
> The impetus for a North American free trade zone began with U.S. president Ronald Reagan, who made the idea part of his 1980 presidential campaign. After the signing of the Canada–United States Free Trade Agreement in 1988, the administrations of U.S. president George H. W. Bush, Mexican president Carlos Salinas de Gortari, and Canadian prime minister Brian Mulroney agreed to negotiate what became NAFTA. Each submitted the agreement for ratification in their respective capitals in December 1992, but NAFTA faced significant opposition in both the United States and Canada. All three countries ratified NAFTA in 1993 after the addition of two side agreements, the North American Agreement on Labor Cooperation (NAALC) and the North American Agreement on Environmental Cooperation (NAAEC).
Politically speaking, NAFTA passed in the Clinton administration because he was able to get enough Democrats on board. To this day, free trade doesn't seem to be strongly supported on the left, in part because it's not strongly supported by unions; it's mostly strongly supported by economists, think tanks, and publications that might be characterized more as centrist (the Council on Foreign Relations), center-right (The Economist) or libertarian. I'm not entirely sure why American conservatives in particular moved so strongly against it in the last two decades.
There are a couple of factions that makes up modern American conservationism. I think it was mainly big business/ideological libertarians that were strongly for free trade, but I think they were far from the majority, numbers-wise
The 80s and 90s were the era of the FTA and NAFTA, and the rise of the fortunes of the IMF and the WTO. And apart from Clinton in the late 90s this was an era of (for the time) pretty strongly right wing American politics.
I remember it the same way, and I was a journalist at the time and wrote many stories about GOP politicians afraid sounding warning bells about American jobs going to Mexico. Which they did. And then they went from Mexico to China.
I also remember that conservative companies like Walmart were so against NAFTA that they actively promoted Made In America goods in store and in advertising. My how times have changed.
iPhones are twice as expensive in China and they always have been.
Tim Cook:
> There's a confusion about China. The popular conception is that companies come to China because of low labor cost. I'm not sure what part of China they go to, but the truth is China stopped being the low-labor-cost country many years ago. And that is not the reason to come to China from a supply point of view. The reason is because of the skill, and the quantity of skill in one location and the type of skill it is.
> The products we do require really advanced tooling, and the precision that you have to have, the tooling and working with the materials that we do are state of the art. And the tooling skill is very deep here. In the U.S., you could have a meeting of tooling engineers and I'm not sure we could fill the room. In China, you could fill multiple football fields.
> The vocational expertise is very very deep here, and I give the education system a lot of credit for continuing to push on that even when others were de-emphasizing vocational. Now I think many countries in the world have woke up and said this is a key thing and we've got to correct that. China called that right from the beginning.
Source: https://www.inc.com/glenn-leibowitz/apple-ceo-tim-cook-this-...
First - cheap labour is still a huge part of the value. Point blank. I quote from China and you can see literally the 'labour' input on the spreadsheet, and on that basis alone, it's still very competitive vis-a-vis the West.
Second - the reason 'all those skills' exist there, and not here, is because of the wages and conditions. They do all those things at far less than 1/2 price than they do here. That's a big deal. And for the most part, they are effectively commodities.
If Apple et. al. wanted to pay 'American wages' for most of that activity, then major clusters of those skills would develop here.
Very importantly, there is the 'trade asymmetry' in that China has trade barriers, the US largely does not.
It would be 'good' if both economies kind of opened up, but it's like a 'prisoner's dilemma strategy game' in that, it's even better if you get your opponent to 'open up' while you 'remain closed'.
The US has lost dramatically by allowing China on the same WTO terms, and all the other shenanigans. The US should really simply implement 'tit for tat' trade policies such as requiring Chinese companies give up their IP in the US while the US hands to a strategic competitor, not allow them to have >50% ownership of anything, put up indirect barriers, the US Treasure could direct major US banks to subsidise Cisco's customers for export, yada, yada. It seems dramatic, but that's what China is doing.
Finally, though we think of a lot of this as 'high tech' ... a lot of it is not. We don't worry too much about China making tons of 'plastic stuff' because we believe it's 'low value' - but as certain kinds of screens, disk drives etc. become truly commoditised, well, in some ways, it's a lot like assembling plastic toys, and the surpluses go to the buyers, not the makers.
Edit: I'm working on a project where our factory price in quantity is $60 here, will be in $20 in China and overall, it all boils down to labour input costs and overhead through the value chain (i.e. N. American manufacturer pays more for Internet, largely due to labour, those costs get passed on). In the end, it's not the price of the materials, it's the layers of services by humans on top.
They would, but you're still talking about a 10-20 year gap. First for masses of people to get the adequate education, apprenticeships, get good at their jobs, and then for these people to mentor the second cohort 8-12 years later to get the kind of volume China and Germany have.
Germany, Japan, Taiwan, and South Korea are all developed countries that don’t have “cheap labor” anymore. They found and are holding their niches. Even in terms of “cheap labor”, China’s losing out to places like Bangladesh. A lot of these excuses for why America can’t compete in manufacturing sound like sour grapes to me.
Vocational and apprenticeships are still a thing there.
Exactly. But it’s the “American dream” to send their kids to college, even if that’s the completely inappropriate thing to do for that child’s future prospects.
Sure. That’s ok. It’ll take time to see a return on investment. I don’t see this “but X” as a counterpoint or really an issue at all.
Further, you could make investments now that exceed current capacities and capabilities today with an eye to the future.
Actually the US does a lot of good manufacturing in some fields, like arms. But in other fields, well-intentioned protectionist policies like have backfired. For instance, rather than subsidizing US shipyards, the Jones Act protects them from foreign competition which has led them to significant decline over the past century.
US policy to rebuild or improve US industry has a very mixed history in terms of results. It probably doesn’t help that a lot of these efforts are focused on bailing out the same old manufacturers over and over again to save jobs in battleground states.
And then you have the 'raw labour' - you still need factories with workers making very little, treated like crap.
The bottom 10% of Americans will not do that. For what it's worth.
Now - there is the possibility for 'full automation'. Sony Playstation in made in Japan by this amazingly automated assembly line.
America could feasibly do this if they wanted for some things, but it will take innovation.
The revolution will be: a 'pair of hands'.
When robotics and AI becomes sophisticated enough that you can buy a 'pair of robot hands' for $100K and very easily train them to do 'anything' ... then this will enable automation in rich countries and it could really damage the entire developing world economies.
> The bottom 10% of Americans will not do that. For what it's worth.
OR you need to find other ways to bolster margins without skimping on labor costs at the bottom end. Intelligent cuts to C-Suite pay and golden parachutes could easy pay for thousands of workers at a living wage at the bottom of any company.
The top 10% of Americans will not do that. For what it's worth.
Nothing will make up for the low wages in China.
America could push up minimum wage some, but it won't change the dynamics of export/import.
> Nothing will make up for the low wages in China.
China's wages aren't even that low anymore. Maybe you are thinking of Bangladesh or somewhere that China is now themselves outsourcing "cheap labor" to?
> America could push up minimum wage some, but it won't change the dynamics of export/import.
Correct, it's not a labor cost issue at the end of the day. As I said, companies are more than capable of finding margins that work and still pay labor what it owes them. There's too many externalities companies are willing to, and legally allowed to, ignore that keep them from feeling any pressure or need to actually do so. The point I tried to make is that is not "low wages exist in China" keeping jobs outside of the US. That's a solvable problem with honest dealing with the labor market. Companies just have no interest to solve it.
(The labor market isn't a fair market, for what that is worth, and has almost no liquidity. Corporate structures impact the labor market a lot more than the labor market affects corporate structures, whether or not you agree that the modern American "gentry" are playing too selfish a game with the American labor market and getting away with it.)
No, this is an ideological statement, not a factual statement.
The 'earnings' (not wealth) of Billionaires would not move the needle that much.
But even the wealth: if Jeff Bezos gave $100B up his wealth, it would be a one-time check to every American for $300. That's it. And just once.
Take up all the billionaires and it wouldn't amount to much for Americans, and it would be shortlived.
"China's wages aren't even that low anymore."
Yes, they absolutely are. To the point wherein there is absolutely no way for America to compete with them on a wage-basis.
"Correct, it's not a labor cost issue at the end of the day."
Yes, it is a labour cost, you're misinterpreting my point. Nothing America can do will make the US competitive with China for things involving labour.
" As I said, companies are more than capable of finding margins that work and still pay labor what it owes them."
This is just fantasy business math. There is no 'finding margins'. There is 'increasing prices' or 'lowering costs'. So either they pay suppliers less, or increase prices in order to pay greater salaries.
And 'owes them' - is again, just made up. How much does Starbucks' 'owe' it's Baristas? What you say they owe? What the market will bear? Minimum wage?
"The labor market isn't a fair market, for what that is worth, and has almost no liquidity"
The term 'liquid' is generally not used for labour markets, you might mean to say 'fungibility'?
The US could afford to pay better minimum wages and probably better for the lower 1/3 - but again, it won't change what's happening vis-a-vis China.
Yet we aren't talking about "every American", we are talking about increasing the labor force, so your math isn't that relevant. Instead, let's say that Amazon required Bezos to put that $100B back into wages for low level employees for ten years: $10B/year works out to as many as 320,512 people employed at a $15/hour full time job minimum wage for those 10 years.
You are correct that I suggested wealth rather than earnings because of course most of the wealthy play shell games with their incomes that launder as much as possible of their income into other things that aren't "earnings" on paper (because one is taxed much higher than most of the "capital gains" taxed sorts of passive incomes). I mentioned golden parachutes, I was not trying to hide that wealth is a bigger component of the debate than just "earnings".
That said, even Bezos' over-the-table non-stock "earnings" are still nearly 3x a $15 minimum wage earner.
All of the above is mathematically true, so easier to consider "facts" just as much as yours. No ideological statements there. I could bring out some actual ideological statements if you wish what those really look like.
> This is just fantasy business math. There is no 'finding margins'. There is 'increasing prices' or 'lowering costs'. So either they pay suppliers less, or increase prices in order to pay greater salaries.
I agree, it is fantasy business math, but it is fantasy business math that is the current language of the MBA and Wall Street folks managing this mess. I did not invent it, and I would use more direct language if it were weren't readily apparent this is the preferred language of those currently propping up the established system.
> And 'owes them' - is again, just made up.
By that logic, all of capitalism is made up. All of human history is made up.
Even setting aside the deeper ethical questions about what an honest society owes to its labor class, I did explicitly define it in that context as capable of meeting the requirements to move those jobs back to the US, which in this case would specifically mean "US minimum wage" as the law of the land people keep claiming makes US noncompetitive with China, and you are just needlessly splitting hairs here.
(To be fair, I'm happy to continue splitting hairs, as I did use "owe" to evoke both the indebtedness meaning as well as the philosophical one as well. But it was well defined in context.)
The issue with China now is 25% Trump Tax, which is huge. For less sophisticated goods, Malaysia, Vietnam, etc. are becoming options.
I wouldn't be surprised if Apple being a major contributor in propping up China for so long before taking action (like their recent diversifications to India and Brazil for manufacturing) is going to be a stain on his legacy long term.
What does that mean? Are you talking about opinions he has expressed, or are you just saying he's responsible for the opinions of other people you associate him with?
I watched the same thing happen to Japan way back.
This is absolutely the case, and you don't even necessarily need to look to "other [geographic] areas" to find them.
But one important factor why China is ruling this industry right now is not only because the cheap labor, but also the massive supply chain. If you're prototyping some electronic hardware, you're in Huaqiangbei, Shenzhen, you can probably go there and find all the parts you want in one hour.
It's extremely slow and expensive for other places to do the same in the world compared to there. And it still holds true for other parts in China (where many of them with even same quality of labor but cheaper).
https://www.reuters.com/article/us-foxconn-india-apple-exclu...
Protective trade polices have a costs, but free trade polices have costs too--they're just different ones.
Add enough local industry to that and there are unspoken pockets of vacuum and two generations of children in poverty and unemployment instead.
Very short sighted.
Apple’s iPhone business collects the majority of worldwide smartphone profits. It’s iPad business collects the majority of tablet profits. It’s Macintosh business collects the majority of pc manufacturing profits.
It does this by focusing on design, development, and usability, primarily through software. It keeps costs down by outsourcing assembly to Asia. If it brought that low value work back, it would make it shareholders and America poorer. It would sell fewer devices at higher prices, losing high paying developer jobs and all for a few more minimum wage level jobs.
If US chip makers have competitive advantages they don’t need our help. If they don’t, helping them is diverting talented resources from better industries hat would have better increased our living standards.
Apple sure makes its shareholders wealthier. I guess you can say that about its 90,000 USA employees, and that's about it- there's pretty much no supplier network to talk about.
Apple pays app developers over $30B a year. It’s created hundreds of thousands of high paying jobs at third party developers.
Disks afaik are still made in the US, although not only in the US.
They've managed to change direction in the past. They were all-in on Itanium and NetBurst until the Opteron showed up. They then kicked both to the curb and went all-in on the Pentium-M/Core microarchitecture.
Intel would be foolish to ignore the ARM market if for no other reason than the server market has started to change their focus on ops/watt rather than just ops/dollar. A 64 core ARM uses less power than a 64 thread Xeon with comparable performance.
That means the lifetime cost of that ARM chip will be much lower than the Xeon since it'll cost less to power and cool it (and the datacenter). Lower power requirements can also change the calculus of datacenter siting.
I don't see Intel getting to ARM ops/watt levels with their x86 designs. The Lakefield designs are promising but the "big.LITTLE" core configuration is more useful in mobile applications than servers.
Intel losing out to ARM designs is really Intel's choice at this point.
These days, I think you’re right with the possible exception of density: if you have comparable resources, vector extensions, etc. an ISA which uses memory more efficiently is going to get a little more out of its cache and memory bandwidth, but that’s probably not going to be huge.
Both Intel and Microsoft made some kind of categorical error in 1995-2005 in assuming that their dominance was tied to a particular programming interface. Intel clung to the x86 instruction set while Microsoft operated on a “Windows everywhere” mindset, first pushing Win32 and then .Net.
Wintel seems finally dead, and good riddance.
Given the resulting impact (Intel being far behind in the smartphone market), it might qualify as the costliest Excel error of all time.
If Intel managed to convince Apple to go x86, and they had low power chips competitive with ARM at the time, then that would be a different story.
You could even argue that it was a happy accident; Intel providing a high performance ARM core may have forced mobile competitors to react, producing high performance ARM chips of their own, and sped up the erosion of the x86 monopoly we're finally witnessing today by several years.
I doubt TSMC would have been able to afford to catch up to Intel without their current high volume business.
Apple pursued custom silicon very aggressively from the start, with the acquisition of PA Semi less than a year after the launch, and the first custom chip coming in the iPad in 2010.
Although with an Intel chip they wouldn't have needed to do this for performance reasons, Intel's fat gross margins would have been equally compelling.
If Intel was willing to design the chips and take TSMC sized margins on them, I guess they could have kept the iPhone business for years, but that isn't in their DNA.
I can't see Intel ARM iPhone chips as anything other than a shortly lived oddity.
By sending all that business to TSMC and Samsung, they were financing their rival's R&D.
If they were willing to do that, I totally buy in to your scenario. It would have been the right play with the benefit of hindsight.
But it would have represented the biggest shift in their business model since they dropped DRAM. Just agreeing to fab the iPhone 1 CPU wouldn't have been nearly enough.
>It was the only moment I heard regret slip into Otellini's voice during the several hours of conversations I had with him. "The lesson I took away from that was, while we like to speak with data around here, so many times in my career I've ended up making decisions with my gut, and I should have followed my gut," he said. "My gut told me to say yes."
https://www.theatlantic.com/technology/archive/2013/05/paul-...
Keep in mind that they didn't quit their Xeon Phi strategy until 2017, four years after that interview. The parallels are strong.
It's an equally valuable market to mobile, but they weren't willing to get their hands dirty, build a commodity GPU, and either a CUDA bridge ala ROCm or concerted effort on bringing OpenCL up to speed.
Instead, they tried for a decade to bring x86 to GPGPU, not for any technical advantage, but because if they can make it stick they can be an (at worst) duopolistic vendor.
They started with in order cores little better than a Pentium with massive vector units, and progressively improved them to Atom cores once it became clear how much work you'd have to do to port legacy x86 code on to them, at which point why not just port to CUDA?
Now the Xeon Phi torch will be passed to... absolutely standard Xeons with HBM. And they're building that general purpose GPGPU, but it's going up against a version of CUDA that's had a decade longer to mature and a nVidia with 102% of their market cap, instead of 9.9%.
The reason for this insanity is Intel realized just how profitable monopolies are, and they're willing to make risky (almost stupid) bets in the hope that they'll yield their next one. They're searching for the home runs with little enthusiasm or staying power for anything else.
This is the Intel I know, and it's not the kind of Intel that's willing to take crappy gross margins to fab commodity ARM chips. It's the kind of Intel that's going to try to do everything in its power to cram so very out of place Atom processors in to mobiles in the vein hope they'll catch on through a combination of incentives and temporary consumer insanity.
Which is exactly what they did try instead of the ARM thing, from 2008 to 2016. RIP.
If Intel did make ARM CPU for the original iPhone and other early mobiles, it would have followed the same trajectory as Intel's original SSDs. They'd be years ahead of their competition at launch, but left to stagnate as the rest of the market catches up, once Intel realizes they're actually going to have to work for their money with no sustainable advantage.
- He's probably flat-out lying, poopooing the data driven stuff to come off more human, humble, and sympathetic to the interviewer.
- If the numbers really did say they should stick to their monopoly, huge extra revenue be damned, then this is some model the FTC and "...but consumer prices!" crowd needs to be forced to grapple with.
>At the end of the day, there was a chip that they were interested in that they wanted to pay a certain price for and not a nickel more and that price was below our forecasted cost. I couldn't see it. It wasn't one of these things you can make up on volume. And in hindsight, the forecasted cost was wrong and the volume was 100x what anyone thought."
https://www.theatlantic.com/technology/archive/2013/05/paul-...
Itanium launched in 2001 (per Wikipedia); if anything, I'd say they overestimated how readily they could ditch x86. Of course, it might be that replacing x86 was a good idea and it was just that Itanium sucked (I'd accept that possibility, given the whole "needs a smarter compiler than exists" thing). But I do agree with the general feeling that Intel and MS both overestimated how solid/entrenched they were.
HP saw the disaster coming and designed their own McKinley Itanium, but a bit later, and HP obviously didn't want Dell to have it.
The compiler support was not good outside HP-UX, if I remember correctly.
And then AMD64 happened, and Microsoft told Intel that there would be no further 64-bit PC instruction sets, and the market would choose between AMD and Itanium.
Cue the guillotine.
it sounds interesting
Non trivial amount of know-how and - if rumours to be belied[1] - direct design then went from Alpha team at Compaq to AMD, though Intel retained some of it (AMD already had been doing cooperation with Alpha, leading to the funky case of "jig it enough and an Athlon-MP motherboard might just run an Alpha and Athlon at the same time").
[1] - Once heard from semi-reputable source that an early K8 changelog entry contained "dropping support for VAX floating point formats". Unverifiable unless someone somehow publishes those docs, but it did fit with "That opteron looks suspiciously similar to EV7"
As I understand it, these were influenced by SuperH, a 32-bit RISC design that uses 16-bit instructions, so two are always executing at once.
The busybox-armv7m Thumb2 binary is the smallest of all the compiled architectures.
Thumb2 appears to have been much more practical.
However, when AMD released AMD64 it killed this strategy. You could now buy a chip that natively ran x86 32 bit at full speed, plus be able to run code that could address more than 4 GB of RAM.
Once AMD64 came out, Itanium was a dead architecture.
[0] https://devblogs.microsoft.com/oldnewthing/20150727-00/?p=90...
This means that there is some hardware that checks condition that will not happen during normal course of events - compiler won't generate such code and write to register that reads zero is harmless.
That hardware stays there, occupies space (does not matter how little) and drains energy (no matter how little).
This part of hardware is utterly meaningless and shows how priorities were skewed during design phase. It also shows why Itanium was such a bad chip - features were piled up instead of being cut down.
RISC-V did cut down on features first (compare division overflow in RISC-V and OpenRISC) and has a process to add them. This is why RISC-V is easy to implement even in OoO variant and why it is being picked up by everyone and their dog.
You mean 64 GB limitations? 32-bit x86 could handle that much, and I'm sure it would have without AMD64.
All in all, not ideal, but I'm sure we would have managed with it a bit longer.
[0]: https://en.wikipedia.org/wiki/Address_Windowing_Extensions
The downfall of the Itanium had several reasons. They went beyond RISC with EPIC, a technology which had great theoretical promises, but had not been shown to yield those results in practise. Also, the project suffered from bad management and delays. Finally AMD pushed x86 just far enough forward to be a viable alternative.
In the beginning, the Itanium was an expensive, power-hungry monster, but it did perform quite well for heavy computations. It could have had a much better future, if Intel had not abandoned it - the software situation could have been much better, if for example Linux boards had been made available to all interested. The latest version only made it to a 32nm process, the transistor count being smaller than in a modern iPhone. So just redoing the same design in a cutting-edge process would make for a very interesting processor.
Intel misjudged both sides, for phones x86 compatibility was never necessary, for business machines, it absolutely is (for the largest number of customers, there have been exceptions).
Intel should have known this, but you buy processors because they are faster or cheaper. These days the third rail is energy but it’s basically the same, if you pay more you want a lot more and “Different” costs. Itaniums where a whisker faster than their ia32 offerings at the time, they cost a lot more though, and they were different. Beyond that, it was like a geek designed the specs with out a market story, risc chips were coveted by x86 dorks because they had 32 registers to the x86’s 5 or 6, well ia64 had 128... If you’re just doing simple stack frame compiler work without full program optimization, how many functions need 32 registers? Let alone 128? everyone could see that you wanted to simplify things in order to put silicon where it mattered for longer term performance, mid to late 90s reordering conditionals for branch prediction was sort of hot in compiler dork and low level circles, well Merced fixed that by adding a lot of complex logic to take both branches and retire the wrong one. IBM, HP and Sun needed different reliability and serviceability features on their high end systems too, from the big enterprise metal perspective they were toy parts for PCs, but they cost more, didn’t run your existing software and weren’t 2x faster or anything like that. They didn’t take a holistic market view as they made it. Rather than work with ibm, sun and hp on those high end issues, they just went back to the lab and told them when the first parts would be out.
Not really sure what goes on inside intel, they have a lot of brilliant people. I suspect that they cannibalize themselves because nothing will ever be as successful as the x86 was when they owned every aspect of the part and package and things are decided to be a failure before they ship. With ia64 nobody asked why. Xeon phi, they gpu and machine learning stuff reflects that too.
Actually their judgement there was correct. There dominance absolutely was tied to monopoly power, via "particular interfaces". Working on alternatives would have undermined the monopoly.
They error was thinking the monopoly wouldn't have eventually been undermined anyways. They upped the risk for more reward, and got bit.
This incidentally is another reason why monopolization is bad: even when it doesn't last indefinitely, it makes for shitty companies that refuse or cannot adapt, and will be stubbornly stick to their ways until more sudden failure. Stagnation and then a jolt is less growth, and irrespective of growth bad for society. No government should want this.
> In fact, Intel pointed out to me in the wake of the article’s publication that it possesses an ARM architectural license, the highest level of relationship with the British intellectual property (IP) licensing firm, which Intel retained when it sold its communications product line to Marvell in 2006.
Even Microsoft proved that they could build a featurific program without .Net, in VSCode. It’s an electron and JavaScript application. And, they were able to make it run everywhere.
Worth pointing out the two directional changes That Intel made, from DRAM to Microprocessor and IA64, NetBust High Clockspeed to Pentium M IPC. The first one was Andrew Grove, the 2nd one was from his disciple Patrick Gelsinger.
Those people are gone. Retired or pushed out by Intel's politics. Their legacy and work continues for about 5 years, as that is the lead time of their roadmap. And we saw what happen to Intel since 2016.
Except that everybody seems to forget that Apple's ARM chip isn't commodity.
Apple is going to tune their ARM specifically to run portable workloads. For starters, they're going to have humongous caches that make the chips expensive.
And they're going to bury the expensiveness of the chip in the enormous margins of iDevice ecosystem.
Apple isn’t going to release chips that benchmark slower than the Intel equivalents. To sell the transition they’ll need to say something simple like “2X faster” on the big screen. So I expect to see 8+ core chips as standard even if they don’t need that many to deliver a good product.
The question is whether such a chip would be cheaper for Apple.
The interesting question is going to be how Apple's GPU stacks up to the competition. We know from the A13X that it is already quite good and I fully expect it to exceed Intel embedded graphics. But what about higher end machines? How long till Apple GPUs can take the fight to discrete graphics?
I already play Fortnite, Call of Duty, PUBG, etc at 60fps on my iPhone.
https://developer.apple.com/videos/play/wwdc2020/10686/
I'd be surprised to see discrete graphics in anything below an ARM iMac/iMac Pro.
My bet is that for the iMac & MacPro, Apple will use custom AMD on-die GPUs (like the now discontinued Kaby-Lake G). They just haven't announced it yet and AMD seems to be silent about it. I also bet the demo they showcased of Rise of the Tombraider was running on such an unannounced custom chip.
I you think about it, it doesn't make much sense for Apple to enter the highend segment for GPUs and sink a significant amount of research cost into it. What they need is power-efficient GPUs for their mobile line (iPhone, iPad, Macbook Air & low end Macbook Pro) so that is where an all-Apple GPU design will be used. But for professional high end use, they will not be in the position to ship a GPU (no matter if on-die or discrete) that could compete with AMD or Nvidias performance hogs out of nowhere.
Market share for high-end Macbooks + iMac + MacPro is just too little given their overall product portfolio (including mobile devices) and the research effort too high. Also, AMD would probably happily design them a custom GPU chip for on-die usage, just as they do for consoles and did for the Kaby Lake G. It wouldn't hurt AMDs business, even if Apple plans to replace AMsD with custom all-apple based GPUs in the long run (i doubt it), as it is very unlikely Apple will ever license it to other CPU vendors or even enter the market for discrete GPUs for Wintel machines.
Arguably the r&d costs would still exist for just their iPhone line. And I’m sure the cost to add a Mac line isn’t as much as the intel premium
For them to not do this would be strange.
Literally everyone has humongous caches. In fact Intel cache architectures are far larger and more elaborate than anything on an Apple chip.
The problem is that Intel no longer has the leading state of the art fabs. They used to be two generations ahead of everyone else, and now they're two behind. The inertia of x86 has kept them afloat, but the longer Intel fabs lag the more ground ARM has been able to make up.
Intel fabbing ARM parts isn't going to do them much good unless they also manage to get their process shrink delays worked out and behind them.
There's been some suggestions to replace node size claims with better density metrics https://ieeexplore.ieee.org/document/9063714. It's not like TSMC's fabs are some order of magnitude more advanced than Intel's fabs. If Intel had an ARM chip to produce today they could use their considerable fab volume to produce them profitably. Even at large process sizes they could still produce competitive ARM chips. It's not like their current x86 chips at 14nm are slouches.
Did they have such competition in the past, plus their main market (desktop) becoming irrelevant and mobile (and maybe server) going to another CPU?
So Intel's got the income to make big shifts. There's also enough momentum of continuing x86 sales that their revenue from them won't dry up immediately if they did start building ARM chips.
I don't think absolute numbers are what matters. For a company in the stock market, the fact that mobile has grown 10x the size of desktop sales, where the desktop has become stagnant means they lost big on a market...
While they could be a player in the smartphone CPU market it's not like they had it tied up and lost it. Intel's real challenge now is recognizing that the desktop/laptop and server markets are no longer guaranteed sales for them.
ex-ARM employee here.
Also Intel clearly has some sort of license that covers modern cores as well, given it manufactures FPGA's with various newer ARM cores based on its purchase of Altera.
Historically this seems to be the case most of the time. Market leaders miss the next iteration of their industry. The "Innovators Dilemma" is ever present.
"This 'telephone' has too many shortcomings to be seriously considered as a means of communication. Western Union internal memo, 1876.
Is a developer in this context a third party who goes out and licenses the different IP Cores for a customer and then designs a chip using them for that customer? If so I'm curious why would someone use a third party developer over ARM itself? Who are some of the bigger ARM developers today?
> In ten years, the US won't be able to make electronics in volume.
Quite possibly, but if this happens, it'll be Intel's fault. Expecting their 500% margin on x86 to last was complete lunacy. They've had 40 years. Anything they manage to invent on the architecture side will not maintain their margin in a world where ARM is good enough. That margin is never coming back.
They need to get their act together, swallow their pride, and learn how to become a foundry like every other state-of-the-art fab on the planet.
IMHO the problem is in the C suites, period. The executives have not provided enough leadership, nor have they recruited others to do so if they themselves are not able.
We still have a ton of great semiconductor design. AMD, NVidia, and Apple's semiconductor division are here. They just don't fab here.
What we have seen with China is that if manufacturing leaves eventually R&D and design will start to leave too. The manufacturers realize they could own the whole shebang if they stand up their own design, product, and marketing, and the US has no talent monopoly.
I am waiting for a Chinese laptop that is built to Apple quality or better, has a fast ARM CPU, and can run either Linux or Windows on ARM. That could be disastrous for the domestic PC industry or what remains of it, especially if Apple flubs the ARM switch.
Edit: Samsung could do it too. They have a sleeper line of ARM laptops, but they are underpowered. Put out one of those with 16-32 cores, 16-64GiB of RAM, 1TB+ SSD, a (maybe optional) high end GPU, and that can run ARM Linux or Windows and Samsung could have a customer. If they invested in helping Linux desktop work very well on the machine they could take a lot of developers from Apple... or if MS removes their head from their posterior and stops stuffing Windows with shit. MS could offer a reasonably priced "no BS edition" without foistware or needless telemetry and I would consider it.
Yep, Apple et al. are reaping what they sow here. All their employees go find suppliers in Asia for their hardware. The travel schedules are so grueling that you basically have to be from the country/province where the supplier is located for it to be a tolerable job (free trips home!). So basically, we have Asian engineers seeking out Asian suppliers on behalf of an American company where only the top few levels are American. The chance those Asian engineers are going to cut out the American company in 10 years is high.
By pushing the PC, the nucleus of all development, to ARM they are helping push the whole software industry onto a heavily globalized platform with many vendor choices and generally available IP.
To me this is Apple trying to stay ahead of ARM Chromebooks and the desktop-ification of Android/Chrome. If left alone another half decade, it’s possible that Chromium via either Google or Microsoft could have made more of a dent in the wider market than just education uses. Adobe is already facilitating creatives to cheaper Windows workstations.
Frankly, Apple investing in unique silicon for their software simply means it will be more unique, just as the iPad has defended its place through an unbeatable software ecosystem and “cheap enough” hardware. If the Mac can go “cheap enough” to iPad pricing, this is Apple carving out its niche and market for the decade to come...
Right. ChromeOS is getting interesting. At first glance it seems to be even more limited and appliance-like than iOS, but now you can run both Android and Linux software in containers, which opens up many more potential uses.
The second justification though is that the CPU architecture is less significant on the system scale. The SOC really is a ”system” (we used to predict this in the 1990s but couldn’t get there then) with a lot of other some what specialized functional units designed to work together.
In that the A series are more reminiscent of the mainframes of old, where the CPU really was merely the central processing unit and a tremendous amount of processing went on in other processing units.
But I was lazy in conflating the two cases so thanks for pointing that out.
Lee Hutchinson of Ars Technica once wrote that a junior engineer now at Boeing has no viable path for promotion towards the C-suite. The middle rank engineering teams have all been divested and the only way to rise up the ranks is to actually leave Boeing, work at a different company, rise up the ranks there and come back a decade later.
The only difference is that you're less likely to be sent to Gulag, or worse.
AMD and TSMC both have always had technical leaders. Intel had a series of legendary technical CEOs during their most formidable years.
The first non-technical-PhD CEO they had was Otellini, but he had been there for 30 years before becoming CEO.
I advise you to see how huge are the overseas captive outsourcing centres of all of said companies are.
Think of it like a career change at age 42.
Imagine being married with a couple of children and at 42 you tell your wife that you are going to spend all of your free time, and a lot of your money working on becoming a lawyer. You want to do this because you think your current career is topped out and you want to be prepared for the future. In 6-8 years you'll be ready for the career change, but you'll need to start at the bottom of the laywer pool and work your way back up, your expenses won't change, nor will your obligations, just your income, and it will go down substantially.
I worked in a startup and we had a product that cost $1k per seat and it was full featured and slick as hell, but we started getting competition from a company that was selling a bare bones product for $99 a seat. Our customers really liked their price, but they wanted our features, in the end price won and we started losing business to the competitor.
Because of my company we actually legitimatized a market for an certain type of after market computer cards, we spent a ton of time and money helping to build the surrounding ecosystem so that our product could have a market. Once we invested all the money, the competitor came along and could ride the ecosystem we built, but we still needed to cover all of our costs, hence our high price.
I spent a ton of time arguing with the BOD about pricing and volumes, but they had financial models they had to meet and an $89 product had no place, luckily for us we got acquired because of our expertise and the work that we did to create the market. If we hadn't been acquired we would have been overrun by our competitor, and it had noting to do with lazy.
As Buffett says, "price is what you pay, value is what you get." Your ex-co failed to provide reasons why customers should pay for the value it offers.
Apple excels at this, despite naysayers saying that a premium consumer electronics is a "bad business".
They want to stay the Ferrari, Bang & Olufsen of mobile phones and laptops.
Intel does both designing and fabbing in the US.
And I think the situation is similar for other US-based social media sites such as Twitter.
But, this list could very easily grow. There are almost 200 countries in the world, and about one-third of those are authoritarian regimes, and at least another third could be classified as semi-authoritarian. That's a lot of countries which might be inclined to ban US social media and could do so in the future.
I think China could export its own social media services to authoritarian and semi-authoritarian governments in other parts of the world – a lot of those governments would be very interested in social media offerings with built-in tools to enable censorship and information manipulation.
ha-ha-ha-ha-ha, Zuckerberg likes to have a word with you
How does one have a social media industry without computers?
do you really think social media(s) appear(s) only with computers?
TikTok is the social media darling not long ago. Now it's in hot water because of privacy issues.
Chrome is the only major browser left that doesn't block 3rd party cookies, tracking and fingerprinting. We’re probably at peak Facebook right now.
Meanwhile, virtually every device needs semiconductors.
It's not glamorous, but Texas Instruments still operates a pile of fabs in the US, and is well positioned for the anticipated IoT boom.
If it's IOT, yes definitely. Some things you don't want to plug in every day--they have to run 10yrs on battery. TI is pretty good in that space.
That's how old empires fall and get replaced by new - and this while everybody in the old remains complacent...
Most of this is out of greed - the corporations got bigger margins if they outsourced production. And the public could short term buy more for less - to the detriment of the local working class, and ultimately, the middle class. And domestic stuff like rent, health, education, etc, would remain expensive.
First you lose the factory jobs, then the factories, then the readily available infrastructure to build one soon as you need it, then the investments to new technology, and ultimately, even the know-how (and meanwhile, lots of patents also go elsewhere, to your subcontractors).
"Corporations" are profit-seeking for the entire lifetime; there isn't a point where the making-money comes to the fore and ruins everything.
Automation explains almost all loss of manufacturing jobs. America today manufactures more than it ever has, with very few jobs needed. A car factory today requires a few hundred people, rather than many thousands.
As for why temporary monopoly companies become less innovative, the answer there is somewhat obvious: because innovation isn't the best profit-making strategy. But that's fine: hence the breakdown of temporary monopolies via competition.
Then there's the question of "why china" is where genuine competition comes from, and that's largely due to economies of scale and large pools of talent. The labour costs are secondary for the capacity to automate and integrate supply lines.
America is a massive country with low density, poor infrastructure and limited talent base for relevant skills. The latter is a choice of its citizens to educate themselves in a particular way. The former is the choice of its citizens to vote for particular government programmes: low-density sprawl, and tax-cuts.
If you want to "bring manufacturing back" (whatever that means) you need significant tax raises to build and subsidise significant amounts of infrastructure and subsidise education and training in the relevant skills.
Of course, that isn't as satisfying for a voting for a rable-rouser.
IMHO large investments in infrastructure and education are the playbook China has followed for the last 40 years to become the world's biggest factory.
In China's case, they didn't fund it by raising taxes on an already-wealthy populace, but rather by going "lol, workers' rights?" and forcing an entire generation (or two) work 12 hours a day for 6 days a week (996) making pennies per hour. [1]
The PPP of Chinese citizens has gone from ~$1000/year in 1990 to ~$17,000/year today. [2] That's a 17 fold increase, compared to just ~2.7 for the US over the same period. [3] I highly doubt any employee in a country like the US, UK, or Germany is going to accept a 996-style work schedule (in fact it would be illegal in many, if not all, of these jurisdictions) and especially not for the current PPP of a Chinese citizen.
> Automation explains almost all loss of manufacturing jobs.
This is also applying to China now. You can't have labour costs increase 17x over 30 years and not consider automation to reduce your costs. However, since China has historically been the destination of outsourced manufacturing, they have built up supply chains that the US used to have back when the US did "manufacturing."
Why would a profit-driven corporation spend money to bring the supply chain back to the US when there is no economic reason to do so? To me, this is a very large issue, separate from "infrastructure" that the US government has yet to seriously discuss. Raising taxes is a politically toxic move in many political systems, and it's incredibly unlikely a corporate entity would voluntarily choose to forego profit to help an OECD country re-develop its lost supply chains.
[1] https://en.wikipedia.org/wiki/996_working_hour_system
[2] https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?locat...
[3] https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?locat...
Err... no. 996 is a recent culture of work in Chinese IT.
The reason subsitance-living agricultural workers moved to cities and factories is the same reason they have always done so in history: farming is dangerous, risky toil.
A wage is highly stabilising, and highly attractive. You do not need to "force" people from farms into factories.
> Why would a profit-driven corporation spend money to bring the supply chain back to the US when there is no economic reason to do so?
The US can engage in the same state-funded programme of infrastructure investment, skills subsidy, and massive sector-targeted tax cuts. That it chooses not to is the explanation of why the US isn't competing successfully. NOT "996".
Its just scape-goating to blame "the despicable Chinese!" when at the same time you're doing nothing to compete with them.
The left? That's just reality, not a left/right thing. I think it's now common sense in both the left and right. Heck, Trump (which one would call right) admitted more or less the same (that globalization lost America jobs and sending things to China was a bad idea).
>"Corporations" are profit-seeking for the entire lifetime; there isn't a point where the making-money comes to the fore and ruins everything.
That corporations are profit-seeking for the entire lifetime is neither here, nor there. You can be profit seeking and selling good products and fostering a healthy community, or you can be profit-seeking and fucking customers and polluting and so on.
So, there might not be a "a point where the making-money comes to the fore and ruins everything" but there's absolutely a case that how a corporation goes about making-money can ruin everything.
Which also means there a point in time, where corporations were let (culturally, legally, etc) to run wild with money-making to the detriment of other things (society, externalities, etc).
>America is a massive country with low density, poor infrastructure and limited talent base for relevant skills.
China is an even more massive country and didn't have an educated workforce (tons of farmers etc) before it began investing on building one. As for the population density, that's a red herring - China got workers brought over from the countryside to work in big cities, and America can do the same.
> before it began investing on building one.
That is my point. Americans chose not to do that. A loss of manufacturing didnt "happen to america" like some crime. American society chose to be educated differently, chose to spend and raise taxes differently, and presently, continues to choose to do so.
The "blame china and outsourcing" narrative is just misdirected scape-goating. Trump's repetition of it is a pretty strong signal of its mumbo-jumbo narcissim.
China has invested in its infrastructure to an incomparable degree. The US has not, does not, and will not. It is your fault.
>state of the art
You must work at intel
Welcome to Europe, I guess.
What happens to this national infrastructure in a time of war? If one country places it’s infrastructure inside another’s then in a state of war the Emirate would certainly lose access to their chip fab. They’d want to see no one else can use it either.
Would something like the Global Foundries site somehow rigged to be permanently disabled, maybe even remotely, so as to not let it fall into the host state’s hands?
And that would not fix their process either...
Plus to be competitive on perfs Apple is using some tricks (larger page size => huge impact on L1) that are maybe difficult to apply to Windows, given the Windows ARM ecosystem already exists (even if no very striving for now).
Also MS is attempting a somehow closed ARM hw ecosystem, so that's not a full replacement for x86.
To finish what to compare to is currently likely to be AMD, not Intel.
So I think x86 is here to stay, or at least that's not the Apple switch that is going to initiate a revolution.
Isn’t Google also designing custom ARM chips for machine learning and their data centers?
Clearly, there can be shared elements, but, if you want something optimized, not everything.
And ARMs fixed length instructions allow apple to just duplication their instruction decoder logic until they are decoding (currently) 7 instructions per cycle.
With x86, you have to work out how long each instruction is to know where the next instruction starts. Intel and AMD actually try to decode an instruction starting at every single byte (over 16 bytes) letting the earlier instructions cancel out any incorrect instructions.
With this scheme, Intel can only decode 5 instructions per cycle, while AMD are currently doing 4. And they both require certain complex instructions to always go in the first decoder.
To overcome this bottleneck, Intel and AMD both cache decoded uops in an L0 cache to speed up tight loops. Intel can do 6 uops per cycle, AMD can max out at 8 uops.
Apple can skip all this complexity. Their decoders operate at full speed and there is no need to cache uops.
Beyond these two areas, the rest of the CPU core is reasonably ISA agnostic.
The scheduling and execution units can stay the same, but the optimal balance of execution units might change based on the ISA.
That balance also might change based on the type of code the CPU is expected to execute, and the compiler.
Could you elaborate why?
On modern Superscalar designs, this requires the instruction cache to actively watch for any writes to cachelines it holds and invalidate them. If that cacheline is currently being executed, the core needs to go further.
The scheduler needs to decide if the instruction modifing the cacheline is before or after that position in the instruction stream. If it's after, the whole pipeline needs to be flushed and the instruction fetcher reset to the modified instruction and execution restarted.
Not a fast operation, but it needs to be supported due to the spec (and legacy code)
With ARM, the spec says the CPU is allowed to do whatever it feels like in this sitatuion. The cheapest option is to just ignore the write and execute whatever old code is currently in the instruction cache.
But to allow self modify code (and loading of libraries) ARM provides a special "invalidate the instruction cache" instruction which explictly tells the CPU something funky has happened and it need to flush data out of the instruction cache before continuing.
Note that ARM still generally supports ldar to match Intel’s semantics.
Also, even with those instructions there are still things ARM can't currently do. X86 can and will do misaligned atomic updates (as long as it doesn't cross a cacheline). ARM can't.
Only the A32 and A64 instruction sets of ARM have fixed 32-bit-length instructions. T32 instructions are either 16 or 32 bits long.
With T32 you just look at bits 15-13. If they are 111, it's a 32bit instruction. If they are anything else, it's a 16 bit instruction.
x86 is somewhat insane. Instructions can be any length from 1 to 15 bytes, and thanks to prefix bytes, the bits that specify the length can be found at any byte offset and might be spread over multiple bytes and interact with each other.
Even for 15 byte instructions, the 15th byte can tell the difference between "this is a valid 15 byte instruction" or "this instruction is actually 16 bytes long and therefore invalid"
Why it has an impact on L1? Because you don't want to perform address translation before beginning your L1 lookup (else the perfs would be shit). So you index your L1 cache by the LSb, which are the same in virtual and physical. So a bigger page size means more bits to index, and a bigger L1 without having to give it more ways. And if you can avoid more ways, its better because you have to compare the tags in all the ways, then mux the data out of the good one, so having tons of ways draw some power. That's why L1D is typically 32k on computers with 4kB pages: that gives 8 ways, which is OK. Intel went with 48k and 12 ways on Sunny Cove, but that's starting to be high. A12 is 128k L1D, with 16kB pages, if I'm not mistaken that's conveniently 8 ways.
It's worth pointing out that L1 is universally split into L1I and L1D of usually the same size, so for x86 mainstream L1 is specified as 64 KB, meaning 32K+32K for instructions and data, respectively.
Of course just about every JIT runtime does this.
It will, indirectly. When MacBooks offer 1.5x the performance at 2x the battery life, people using Windows or Linux will start to demand similar devices from Dell, Lenovo, HP, Acer, etc;
It makes me the saddest for AMD. They’ve just clawed back a few victories from a decade of loss, and starting within 5 years that will hardly matter.
As for the customer: this will be an amazing time, especially if we will finally get some standardization around ARM SoCs. Apparently every single one of them (at least phones) does their own little boot quirks, instead of their being a unified BIOS or EFI or whatever to guide it all.
Well, people can demand that, but if only Apple knows how to design the chips, and given they will not sell them to competitors...
And yeah it'll exist but if they decide not to make the OS replaceable then there's going to be a severe dampening effect on people switching to Apple hardware. So we'll see on that front.
It's like saying BMW or Mercedes is a better car than a Kia. I'm sure it is but neither my sister nor my 2 adult nephews can afford the Mercedes and neither can they afford the MBA.
Just like Kia will still get them to work and to the grocery store the cheaper laptops will still let them facebook/twitter/ms-office/slack/google/maps just fine.
Going on Amazon an typing "laptop" 6 of the first 7 hits are under $500 so less than 1/2 the cost of the cheapest Mac. The 1 that is over is $579 so nearly 1/2
I'm typing this on a $3600 MBP and I have a $2700 MBA as well so it's not like I'm not an Apple fan but I see what my family buys they'll opt for the perfectly adequate $550 computer over the probably superior $999 one.
In my mind the small wildcard is going to be how effectively Apple can scale up CPU performance as their thermal budgets are increased. But the big wildcard is going to be how their in-house GPU tech scales up. It's going to be strange times ahead if Apple can bring the fight to AMD and Nvidia's mid-tier discrete products.
While Apple doesn’t need to be chasing tier 1 AAA game performance, they do need graphics horsepower to handle two or three 6K displays. They also need to be a compelling platform for occasional gamers. And they need performance to power their push into AR.
> It makes me the saddest for AMD. They’ve just clawed back a few victories from a decade of loss, and starting within 5 years that will hardly matter.
I wouldn't be sad for AMD at all. This transition presents them with an amazing opportunity to take more of the client market.
The performance deficit you cite is more of an Intel performance deficit, as current AMD processors also substantially outperform Intel's offerings. I would expect AMD Zen 2/3 and Apple Silicon to perform similarly in terms of performance/watt, as they're both modern processor designs manufactured at TSMC.
Unless Intel can quickly resolve the manufacturing woes that keep them from moving to a competitive semiconductor process, I expect the PC OEMs threatened by Apple to turn more often to AMD to stay competitive.
It's early, but these benchmarks from the DTK looks really promising: https://gizmodo.com/a-wild-apple-arm-benchmark-appears-18442...
The collapse of so-called "Wintel" will not be directly from Apple's move to switch to their own silicon, but as a second order effect on the other OEMs such as HP, Dell, Lenovo. Just as they're reaching build parity (e.g. XPS 13, X1 Carbon) with Apple's Macbook offerings, Apple is potentially about to pivot and take a lead in where they can't follow - in CPU performance. Make no mistake that ARM-based Macbooks will easily surpass the performance on the existing Intel-based offerings. What will Dell, HP, and Lenovo do to be competitive?
Microsoft has arguably seen this risk and trying to mitigate with shifting Windows to ARM. But there's a big problem - Qualcomm is not Intel and their chips as-is (e.g. SD845) have both compatibility and performance issues. Are there any alternatives? AMD could bring back their shelved "Seattle" chips, or maybe Intel's Lakefield platform will be fruitful, or even Nvidia will join the fray with a worthy Orin-successor?
Very interesting times ahead indeed...
1. When Intel released HT as a way to recover 15% idleness in CPUs, the Linux scheduler took a while before it realized it should schedule work on all full cores before assigning something to a hyperthread.
2. Intel has been adding specialized SIMD vector operations, and wide vector operations, yet popular compilers like GCC were really slow to take advantage of them. One or two blog posts a month will hit the HN front page about inlining assembly or using specialized math APIs like the Intel MKL to extract performance from the cores you're running on.
Apple can add neat blocks like audio, low power video playback, always-on processor, and neural engine because they also release the software APIs for it. Maybe that's Intel's sin. They never realized how hands-off the white box industry needed to be to stay alive and relied too much on OSS developers to fuel general-purpose Linux performance outside of large HPC computers. Intel made add-on things, people didn't use them, and now Intel is perceived as the company that can't innovate.
On the bigger point, it ends up hurting them because companies often use gcc to benchmark across systems, so the Intel performance comes out lower than it might be if Intel really invested heavily in gcc rather than icc.
So now Intel chips are limiting battery life because every application is executing unoptimized garbage that runs longer and therefore uses more energy. Meanwhile, Apple Xcode will produce optimized binaries for Apple Silicon and the press will go wild.
According to statistics maintained by LWN [1], Intel was the number #1 employer of Linux kernel developers in the latest kernel, and has been for a while. If you go back into previous years, Intel was not always as high up, but even back in 2007 (when LWN started tracking this) they were in the top 5-10 (depending on how you count it.)
Intel released HT way back in 2002. I don't know how active Intel was in Linux kernel development back then. But what Intel was or wasn't doing 18 years ago probably isn't very relevant to the question of Intel's current performance.
Keep in mind that Intel's ability to properly light up new silicon features on your PC often relies upon your willingness to update your software. For instance, even if the SMT scheduler fix has been written and upstreamed, the next Ubuntu LTS to integrate the change may be 18 months away. The end user will believe that Intel is taking a long time to develop the fix.
My naive POV - It would seem worthwhile for Intel to participate in GCC/Linux to accelerate adoption of these innovations... PRs welcome, etc. (though I realize it’s often more complicated than that)
Standardise the features you add, work with AMD and I’m sure you’ll get more support.
- nVidia GPU with CUDA
- CPU with SSEv2 (or whatever the second version of vector was called) optimizations
- CPU with SSE optimizations
- CPU without vector optimizations
In the block diagram of the SoC for the new Macs, Apple very clearly signaled what they are doing. They are working towards Steve Jobs's long-term vision.
Onboard "advanced audio processing", "enhanced camera", "neural engine", "AI processing": this is about making Macs that support development of the next generation of software for iPhones.
Mobile phones right now are stuck. Apple wants to make them more useful to their users, and the way to to that is to make them context aware, continuously listening and occasionally looking around. so the phone knows what's going on.
Context-aware phones is what this is about. "Siri, write minutes of the meeting we just had, and email them to everyone."
Users will never accept a phone that's continuously uploading everything in their lives to Apple's servers (or anyone else's), so all of this capability has to be on board the device.
Intel is incapable of making a SoC that will support this, and no-one else has the capacity, so Apple has to do it themselves. (Also, Apple wants to keep the IP to themselves.)
My guess is that they want the development platform to be a beefed up version of the target platform.
Intel throws a lot of money at software, much of it open source, and this has played out fairly well for them. AMD is far, far behind in this regard.
I fully expect a cutting edge implementation of macOS on ARM, and I fully expect it to languish post release.
I think a fair criticism would be the noise created by developers about the quantity and average age of bugs in their APIs.
In 2015, 10% of Intel’s engineers were software. I don’t know how that’s changed in recent years.
https://medium.learningbyshipping.com/apples-relentless-stra...
This doesn't sound right to me. You seem to be looking for ulterior motives where none need exist. Isn't it a sufficient explanation that Apple believes they can build a superior product at a lower cost? Why does it need to be some conspiracy about re-framing the Mac as an iOS devkit?
For this conspiracy to hold, you'd first have to believe that Apple builds Macs with developers in mind. Nobody with a passing familiarity with recent Mac history could believe that. It's been a long time since Apple has focused any Mac product on the needs of developers. Why would they start now?
Besides, so what if future Macs are more like the iPhone? While it might be of occasional benefit to iOS developers by increasing feature parity of their iPhone Simulator, that's an extraordinarily marginal reason to overhaul a critical nexus of Apple's product stack.
Oh man, I genuinely want you to be right in that prediction. But I wouldn't bet on it.
IMO we already see many (most?) users and even tech industry pros (who should "know better"?) showing comfort in uploading/streaming/routing large swaths of their lives to many companies. Companies arguably less "good" than Apple. (And many of these are Apple users who also have no problem hosting their lives on Google and Zoom). Because convenience, shiny things, and lifestyle marketing. And "free".
I would love a future where all personal data and heavy computing was done on local hardware but I bet most people, given the choice, would choose a cheaper option that puts streams and storage and computing in someone else's hands. Because "free" and/or "cheap". i.e. the product is you
In 2016 ARM sold to SoftBank for $31B.
In 2016 Intel had a market cap of ~$170B and about $17B cash on hand. I'm no M&A analyst but surely given their fundamentals at the time Intel could have bought ARM without financial difficulty, no?
If even Facebook can look two steps ahead and pay an eye watering $19B for WhatsApp couldn't someone in the C-level at Intel could have seen that owning ARM would have been a jolly good strategic purchase - or am I missing something?
1/ Intel could be considered a monopoly
2/ Intel think they can beat ARM in the long term
This is according to the internal rumor mill at Intel Labs circa 2010, so take it with a grain of salt.
At no point does Intel selling ARM make financial sense until Intel is on the verge of death.
https://www.anandtech.com/show/15603/the-samsung-galaxy-s20-...
The tablet situation doesn't look much different.
Who's going to design equally capable SoCs for Windows portables? Doesn't look like it's going to be Qualcom or Samsung.
They've got Tegra and experience targeting 5nm, or at least, they appear to be growing it. It could be reasonable to see nVidia willing to jump to 5nm for Orin or whatever is after it.
Yeah I don't know how realistic that is, but, I struggle to guess at any other company that could stand up and go toe-to-toe with Apple at least in some target markets on ARM... if they wanted.
The real question is mobile GPUs, and whether NVidia or AMD want to step into that market. On the desktop, Ampere and RDNA2 absolutely crush Apple's offerings, Apple is multiple generations behind. However this is at the expense of not really carrying too much about TDP. The real question is can NV or AMD squeeze a cut down version of their architectures into a SoC thats power competitive.
That is, I might be willing to buy an AMD "Fusion" that has a Ryzen and RDNA2 architecture in it, even if it is a severely cut down version, over a MacBook if the battery life is reasonable.
https://news.ycombinator.com/item?id=5587283 ("The rise and fall of AMD...")
With comments like these:
https://news.ycombinator.com/item?id=5588069 ("I could go on and on about all the reasons I believe AMD went down the tubes... but I think a lot of it reduces to the disfunctional corporate culture alluded to in this piece.")
Guess what happened with AMD? So I suggest waiting with the obituaries, at least until Apple actually has a working system a normal person can buy, and until Intel/AMD shows us their response.
The entire idea behind the obituaries is that Apple's whatever will be so performant that ordinary PCs will switch on mass. This misses a few things:
A) Where will the PCs get their ARM chips from? Not Apple, and non-Apple ARM is still obviously inferior to x86. Intel/AMD will rather keep their x86 monopoly, so it's in their interests to keep producing x86 as long as it's competitive even just against non-Apple ARMs.
B) The typical Desktop/Laptop user doesn't care that much about performance anymore - not as much as they care about running their favourite apps. x86 has a huge lead here.
All it takes is for Intel/AMD to stay/become slightly competitive with whatever Apple put out, and Apple will stay in their silo - which actually wouldn't bother Apple Inc. at all; They always ran their own ecosystem.
C) Really, CISC/RISC doesn't matter that much. Apple's advantage (if there's one) would be a result of other architectural decisions, we should ask ourselves if there's anything preventing Intel/AMD from copying those decisions.
It doesn't need to and is unlikely to happen en masse. Apple creaming off the most valuable consumers is a problem for everyone else. You can see that in the smartphone market where Apple dominates the profits.
Regarding your thesis:
'Creaming off most valuable consumers' somehow didn't stop Huawei, Samsung, etc. from being huge profitable enterprises. Partly because it's not quite true (Apple has a position in some countries, in others it might as well not exist), partly because there's more to things than a benchmark comparison (Apple survived for a long time when Mac processors were inferior), and partly because we'll see a response from Intel and AMD. It's going to be an interesting few years in this sector.
https://gizmodo.com/so-just-how-powerful-are-apples-new-lapt...
The ARM TDK is basically an iPad Pro running macOS, without a screen and with extra ram. A12Z themselves are just re-binned A12X (with just one extra GPU core activated), processors that are nearly two years old.
No way Apple is shipping that in their first MacBook ARM...
No doubt. However, I don't think the article is so much "getting it very wrong" as it is working with what they've got. We've seen what a 10-18 watt soc from apple can do in the iPad. And it's pretty cool. But the point is it's not truth defining jaw dropping revolutionary. We can look at what AMD is selling right now. It's also pretty cool and starting to turn some eyeballs. It's very unlikely Apple is going to release something that outright spanks the status quo. So my expectations are that whatever Apple releases will be competitive with what we currently see from AMD and, maybe, with what we'll see later this year. That's all. I'm skeptical that the shift to ARM is the crux of the matter, is all.
So I know why I can't get a good mini-ITX board, but I still really want one :-)
As is, my best option for the living room NAS/HTPC turned out to be a J4105 board from ASRock. 10w TDP and apparently it can do 4 simultaneous GPU-accelerated 4K transcodes, which is pretty neat. Perhaps I'll upgrade to an AMD board at some point.
I realize the situation changes every time a new CPU comes out, but I have never personally seen a real workload where ARM won on energy efficiency and had reasonable performance. Tests like [1] and [2] showing x86 having an orders-of-magnitude lead on database performance vs. AWS Gravitron2 should give you serious pause.
1: https://openbenchmarking.org/embed.php?i=2005220-NI-GRAVITON...
2: https://openbenchmarking.org/embed.php?i=2005220-NI-GRAVITON...
If you're wondering why ARM needs to have both competitive performance and energy efficiency, see Urs Holzle's comments on wimpy vs. brawny[3].
3: https://static.googleusercontent.com/media/research.google.c...
Google is laggard when it comes to server tech. Amazon and others have leapfrogged them.
The poster above says that Amazons "leapfrogs." They question is "leapfrogs where?" The fact that ARM cores cost 100+ times less than Intel, and are n times more power efficient was well known for the whole eternity.
What people don't get is that you get the whole platform on x86, and ARM is a very, very DIY thing, even if you are a multinational with lots of money on RnD.
They've already brandished their ability to port their whole product from x86 to POWER[1], and deploy POWER at scale if they need to[2]. My personal interpretation of these announcements is they are made with the purpose of keeping their Intel sales representatives in order, but the fact that you don't also see them or anyone else brandishing their AArch64 port should tell you something.
1: https://www.cnet.com/news/google-acquires-a-taste-for-ibms-p...
2: https://www.forbes.com/sites/patrickmoorhead/2018/03/19/head...
"Alphabet operating expenses for the twelve months ending March 31, 2020 were $131.077B, a 13.47% increase year-over-year."
Operating expenses is everything, including salaries, etc. Google isn't dealing with $2B/wk in power/server purchases, though they're still huge, no doubt.
If asked whether google will succeed at something new (say, Fuschia), given those priors, my response will be: "no. it would be a surprising first in many many years. the company is on decline"
What we're missing is the connection between the services of the large companies: Google, Amazon, Microsoft all have an offering made of devices (hardware), websites (software) and cloud services. There seems to be a synergy, where you benefit from doing all 3 things in-house to reduce costs on your core product or to capture consumer minds. Microsoft is getting back in phones, with an Android offering. Amazon is not giving up on Kindle.
Notice how Apple is missing on the cloud services part here. They have some internally (for Siri) but they do not sell them.
Even if they don't start a cloud offering, they may sell their CPUs to others who will, before eventually rolling their own hardware.
This will give time to people who adapt existing server software to work better on Apple ARM CPUs (recompiling is the tip of the iceberg, thing about the differing architecture, what can be accelerated etc.)
We are seeing SIMD/AVX optimization for database like computation just now. It may take a while.
Same with Hangouts and video calling.
Tiktok and zoom are both flash in the pan fads.
Youtube requires a lot of server space compared to tiktok (<10 min means no ad money, so people make videos at least 10 min long!), and Zoom requires almost no space, while it can sell corporate subscriptions.
The only reason youtube still enjoys some success now is because it wasn't made in house, and the acquisition wasn't too badly managed. Grandcentral (parts of which still live as google voice) was a different story.
But it only shows how the last success google made "in-house" was a long, long time ago. The alphabet rebranding changed nothing. Since youtube, Google has turned into another yahoo for startup: a place they go to shrivel and die.
Hello,
There definitely is that power use advantage.
Those tests are pathological cases because of bugs, and aren't representative of the performance of the hardware. (for that one I suspect that ARMv8.1-A atomics weren't used in compiler options...)
> I'm also not sure why atomics would be important in a single-client mysql benchmark
They are because exclusive load & store operations are quite expensive on Arm, even if you actually just run an ST workload.
PostgreSQL appears to have had ARM support for sometime. Mysql only added it with 8.0. As far as the other database options are concerned, ARM isn't even supported yet.
I don't think they would do that.
It is not wise to venture into microelectronics business, let alone CPUs.
https://youtu.be/Nb2tebYAaOA (Starting from ~6:30)
Right click -> "Copy video URL at current time"
Energy Efficiency Wins on Server Workload is a very recent thing. To the point it is new as it basically start with N1 / AWS Graviton2. And even that is not ALL workload.
Not all software are optimised on ARM, compared to decades of optimisation on x86. Not to mention Compiler options and EPYC was running on bare Metal with NVMe SSD compared to Graviton 2 running on Amazon EBS.
And most importantly, you would be paying for 64 Thread on Amazon for the price of 64 Core Graviton 2. i.e It should be tested against a 32 Core EPYC 2 with SMT.
I still doubt G2 would win in the fair test, but it would be close, and it would be cheaper. And that is the point.
EPYC is an innovative product and comparing it with Graviton clearly shows that AMD has done a great job and that Graviton is not quite fully competitive yet (but not to the extent that those benchmarks seem to indicate, as others have commented).
I think its possible to overstate the energy efficiency gains from using ARM but all the indications are that ARM cores with fully competitive performance will emerge and that they will have some efficiency advantage - after all why would AWS be investing in ARM if not?
We already know what this is like, and the challenges are usually in the other direction (write/test on x86, attempt to deploy on ARM).
This is due to things like word alignment and memory ordering requirements. For the most part, the "extra complexity" in x86 allows you to ignore a lot of the stuff you need to pay more attention to on ARM.
As an iOS developer, you don't need to think about memory barriers if you don't want to, but Apple's simulator and frameworks absolutely do. That's the kind of work that awaits systems programmers who want to port high performance x86 server software to ARM.
x86 hardware is physically more permissive than ARM. Accesses that must be aligned on ARM can be left unaligned on x86, and they'll still work correctly. But x86 isn't going to explode either if you do happen align them. ARM is weakly ordered, x86 is strongly ordered, so you generally need more fences on ARM than you do on x86, but leaving these fences in place doesn't break things on x86.
My point was that code that works correctly on ARM usually works correctly without modification on x86. The opposite is generally not the case, even if the iOS simulator hides this for you.
The real problem is memory concurrency model. You will hit subtle concurrency bugs with ARM which were not manifested on x86. And those bugs will be present in any language with threads and shared data. They are hard to debug and frustrating, because they'll lead to rare deadlocks, and resource leaks.
While the big cloud providers are certainly able to internalize design of their hardware, you still need to ship millions of servers to smaller players.
While Apple is not traditionally considered a HyperScaler ( Amazon, Microsoft, Google, Facebook, Alibaba etc ), their own Datacenter is still pretty big. You would be using it when you are using Siri and iCloud. And they use Xeon.
>Margins will inevitably suffer as the ARM-based SoC field is filled with sharp competitors such as Qualcomm and Nvidia, sure to be joined by arch-enemy AMD and others, all ushering in a new era of PCs.
That is the problem. It isn't ARM vs x86. Nor Intel's Fab vs TSMC ( Although they are part of the equation ). It is their Margin.
Intel is not willing to risk their current healthy margin. And there is nothing inherently cheaper with ARM design. Intel could have lower their x86 Chip with less Margin to compete. In that case the compatibility of x86 with slightly higher premium will be able to fence off any ARM attack into PC market.
In the short term everything will still be fine for Intel. The market moves very slowly unless you are Apple which control everything in their stack. But in the long term the unit economics of x86 chip with current Intel's margin stand zero chance to compete. Unless Intel's Fab does some miracle and suddenly leap ahead of TSMC by a generation. Which as far as I can see there is zero chance happening in the next 5 years. Further out we will have to wait and see.
That is of course assuming the cost of running those Fab and their Tech being equal. Which we know that isn't the case.
And I haven't even mention AWS Graviton 2. I would not be surprised if Graviton 3 is already sampling.
Even worse for Intel, most computers over $1000 are Macs leaving Intel with the low end.
Modern phones have 2+ GB RAM and an SoC that could easily be used in a tablet or laptop (think low end Chromebook). The PinePhone literally runs mainline Linux. Mobile is PC, it's only the physical form factor (and associated power usage constraints) that differ.
I am not sure what you are trying to suggest.
You either go in to the bigger market and complete with lower margin. Or you stay within the market with margin and milk it for as long as you could.
https://www.amazon.com/Innovators-Dilemma-Technologies-Cause...
The solution?
https://www.amazon.com/Innovators-Solution-Creating-Sustaini...
But will their margin of superiority be large enough to cause the seismic shifts in Windows/x86 land that this article suggests?
I'm not sure.
If Apple's laptops are 25% or perhaps 50% better, I'm really not sure that's enough to cause a major upheaval.
I think we will wind up in a middle ground different from the upheaval JLG suggests.
We are already at a place in which your average laptop is "good enough for most people"; being 25%-50% better than "good enough" is a nice competitive advantage but that's not going to necessarily turn the world upside down.
- "Power users" will appreciate the CPU improvement. At this point, that essentially means video editors. It would mean gamers, but Macs still won't be the first choice for gaming.
- "Extremely mobile users" will appreciate battery improvements. However, I suspect most laptop users use them chained to a desk most of the time anyway. I would love 50% more battery life, but it's not really actually that much of a factor for me personally.
- "Cheap laptop buyers" won't be buying Macs anyway as Apple (wisely) doesn't seem like it wants to play in the sub-$1000 space. Though... maybe it will now.
If you ask me, Apple is going to make a ~$700 passively cooled (sealed like an iPad) MacBook Air with similar/better performance to the current model.
Most importantly, it will have the same build quality and “expensive” feeling.
Sure, there are plenty of laptops now with similar build quality to Macs, but none of them are actually cheaper than a Mac. Load up a ThinkPad X1 Carbon or XPS 13 with a high DPI display and comparable specs and you’ll see what I mean.
But now you’re going to have Apple saving $200 compared to every other computer by not giving Intel a dime. The A13 chip is a $10 line item in the component cost of an iPhone.
This barely even costs Apple R&D money since it was already a part of the much larger iPhone and iPad business.
Literally nobody will be able to touch Apple on the value end of the spectrum unless they go into bargain basement territory. Are customers going to buy a $600 plastic 1080p laptop with big fans, keyboard deck flex, and 5 hour battery life or will they buy a $700, thin, fanless, retina aluminum Mac with the same performance, 10 hour battery, and a much better fit and finish?
If this all sounds a little unrealistic, I’ve got an iPad Air to sell you!
There is zero question in my mind that Apple can do what you say, for the reasons you say.
The only potential obstacle is institutional will. You would now have a larger overlap between low-end Macs and the higher-end iPads. In my opinion, this is not a problem because I don't view those product lines as conflicting. But does Apple?
I find this path unlikely and it depends on Intel's foundries not falling further behind TSMC. The most likely path is to compete on price-per-peformance (i.e. slash prices and internal costs) rather than performance-per-watt.
It is pretty much the biggest semiconductor company out there, focusing solely on 3-4 main product lines.
The most surprising is them under-performing so much while them having resources comparable to the next few competitors combined.
It is not only their fabs that got an arrow to the knee with 10nm.
Their newest microarchitectures deliver tiny IPC gains in comparison to AMDs products even on comparable nodes, while having significantly more transistors per core.
Tiger Lake throws 40-45% more transistors, and only gets 10% IPC gain at max, and the rest of performance gains come from better thermals.
A few years down the line, Intel will look back and thank AMD for their decision to spin off GF and go with TSMC with Zen 2.
EYPC and Ryzen are keeping the x86 arch relevant until Intel is able to catch up on their next-gen process nodes.
AMD may be eating Intel's market share for lunch in a few markets, but in the long term keep ARM far enough away for it to be a niche player, both in the server space and the laptop space.
https://www.gartner.com/en/newsroom/press-releases/2020-01-1...
Apple alone sold more ARM based iPhones and iPads than that last year at higher profit margins. Not to mention all of the ARM based chips used for the Watch, Airpods, Apple TV’s, and HomePods.
What this misses, gravely I think, is that from the armchair CEO angle, I would want to go after _Qualcomm_
They’re the biggest vendor in this space for mobile as far as I can tell. All the major Android vendors use them, including in high end devices, and they’re apparently incredibly hard to work with from what I understand. This would be the way in for them to muscle into the ARM market but supplanting Qualcomms dominance.
That would make more sense to me if I was running Intel. Then again, they ditched their cellular unit so it seems this might be questionable all together
Exactly my point, in so far as what I said: Apple’s never going to be Intels customer for CPUs once they the last Mac requiring Intel CPUs is no longer supported. That’s a dead market.
Qualcomm on the other hand, is ripe for some real competition. If Intel wants to swing big into the ARM CPU game, they should look to dethrone Qualcomm. Apple is pretty much irrelevant here. That’s been clear since they brought all their CPU designs in house
https://browser.geekbench.com/macs/macbook-air-early-2020-in...
Is significantly slower than a $400 iPhone SE in single core, and only slightly faster in Multi-core.
https://browser.geekbench.com/ios_devices/iphone-se-2nd-gene...
The entry level $900 MacBook Air is 30% slower in single and multi-core than an SE:
https://browser.geekbench.com/macs/macbook-air-early-2020-in...
Now imagine next years MacBook Airs running the next generation A14 processors on a 5 nm process instead of 7 nm, with much more thermal head room than a phone, with far faster integrated GPUs than Intels, and with functionality like T2 integrated in the SOC. And using processors that are at least $100 cheaper.
It’s mind boggling to think of how much faster they will be, how much longer their battery life will be, and that they likely will be even cheaper.
1) Either they are going to sell those chips on the wider market, which probably won't ever happen, or
2) Eat the costs and make the machine unprofitable to retain a market, which isn't too likely either, or
3) Make the machine less expensive and trying to do a high-volume play out of it, or
4) Massively raise the costs.
There's no way out apart from those, which raises questions on what they'll do.
Instead of having such a foolish dream it makes more sense to just find an existing ARM server vendor.
There are many innovative and competitive hosting companies in Europe that take advantage of the cheap bandwidth. They won't sneeze at cheaper hardware.
6) Maybe Apple will never build "Mac Pro" class CPU. Instead the Mac Pro ekes out an existence with a higher thermal budget and more GPU/FPGA accelerator cards. (Imagine if Apple released many more FPGA programs for their Afterburner and then let you add two or even four into a Mac?)
7) Use their most powerful chips to build a games console and crush out Nintendo like Sony did to Sega twenty years ago.
And additional R&D costs are a drop in the bucket. They already have a big team to design processors, next year will be their 14th version. Every year they have a large team rev MacOS.
If they spent $1B a year on R&D for the Apple Silicon transition, they will get double that back the first year on lower CPU costs on a $25B sales rate. Except their sakes are going to increase substantially. And they won’t even spend half a billion total on R&D for the entire two year transition, Rosetta already existed, the A14 was already being designed, etc, etc.
So they kinda already have the volume to cover R&D costs thanks to how well their phones and tablets have sold.
Also dare say we may see a revival of the hybrid chip design become more fashionable, with virtualization matured, x86 and arm on the same box may become a thing in some area's. Actually a few companies have tried that in the past and one that jumped out (was looking for Nvidia who was at one stage working on such a chip) https://www.extremetech.com/computing/185888-vias-new-isaiah...
>It’s mind boggling to think of how much faster they will be, how much longer their battery life will be, and that they likely will be even cheaper.
I hope so as there has been for a while (seems to of hit the apex of change) been a case of lower power chips meant they could use a smaller battery and make it lighter and smaller. That with advances in screen resolutions and panels soaking up much of that gain in CPU power savings for laptops and mobiles, we have or seem to of hit that point were gains in power are now being handed back to the user and the drive/marking of last a day or two seems to be more weighted over - hey - this is as light as a feather drive.
I'm particularly excited by the idea of Apple adding iPhone type security to their laptops.
I get all the counterarguments about owning your own device and freedom to tinker.
But I, personally, would be willing to make that sacrifice. Particularly since it's pretty easy to get totally open devices, whereas getting a secure laptop is not so easy.
Geekbench is super misleading because they do things like include benchmarks which are hardware-accelerated on the CPU in the iPhone but not the Intel ones. The weak little low power CPU in the Air is also hardly the fastest single thread Intel processor, much less multi-core, so I don't know what they're planning to do on the desktop.
> Now imagine next years MacBook Airs running the next generation A14 processors on a 5 nm process instead of 7 nm
But then it's competing with not only whatever Intel manages to come up with in the meantime but also AMD processors on the same 5nm process.
> with far faster integrated GPUs than Intels
But, again, faster than AMD's? Or what Intel will have at the time, given that they're now developing discrete GPUs and will have that in house to improve their iGPUs?
> And using processors that are at least $100 cheaper.
Which Apple has no incentive to give to you instead of putting in their own pocket, especially if what they offer is actually better.
> how much longer their battery life will be
"Much more thermal head room than a phone" and "much longer battery life" don't go together, they're a trade off against one another.
The amount of hype around this is getting to a level that people are going to be very disappointed if it doesn't live up to it.
I'm waiting for some independent third party benchmarks before getting so excited. And not the first ones which were given early access in exchange for favorable skewing.
Intel is years behind on process and GPUs. Apple is not just getting faster CPU and GPU performance in lower power envelopes, it’s getting those processors at less than half the price. Apple could have switched to AMD to get better faster processors cheaper, but nowhere near half as cheap, and nowhere near as power efficient.
And Apple has no reason to use this to increase margins, when they can use it to dramatically increase volume and market penetration, which will increase margins on its own. Fir example, the iPhone SE is faster than any Android phone ever made, yet Apple still priced it at $399. They wouldn’t risk cannabalizing their higher end phones unless SE margins were comparable to higher end phones. Which again shows you how cheap their Ax CPUs are.
And iPhones and iPads have all day batteries a fraction of the size of laptop batteries, and without the space, mass, or fans of laptops. The A14 will be able to run in MacBooks at a higher power level while still remaining cool, providing for more cores and faster processor speeds.
It will still be using significantly less power than the latest generation Intel laptop CPUs. Also because the SOC will integrate a performance GPU, T2 functionality, Neural Engine, etc all in one package.
If Geekbench wasn't so misleading, why do they do worse on other benchmarks?
https://benchmarks.ul.com/compare/best-smartphones
https://www.gsmarena.com/benchmark-test.php3
> And “Hardware acceleration” isn’t misleading if it’s accelerating commonly done tasks.
Hardware acceleration is misleading because it can give you huge speedups in a single measurement which significantly affect the overall score when it's averaged in, and because it's easy for the competitor to hardware accelerate the same thing if it really matters without losing any of their advantage anywhere else.
> Apple could have switched to AMD to get better faster processors cheaper, but nowhere near half as cheap, and nowhere near as power efficient.
With the amount they have to spend on R&D in order to do this, it's hard to argue "cheap" and the AMD processors have TDPs down to 10W, so it's also hard to argue efficiency.
> And Apple has no reason to use this to increase margins, when they can use it to dramatically increase volume and market penetration, which will increase margins on its own.
Lowering prices never increases margins. It may increase volumes, which reduces amortized costs, but there is a grisly trade off there. If you're selling a phone for $400 which has a $200 unit cost, lowering the price to $300 is probably not going to double your sales but it is going to half your margins.
> Fir example, the iPhone SE is faster than any Android phone ever made, yet Apple still priced it at $399.
Meanwhile this is the cheapest iPhone and it still costs significantly more than the average Android phone, which is around $250.
> And iPhones and iPads have all day batteries a fraction of the size of laptop batteries, and without the space, mass, or fans of laptops.
Predominantly because they have smaller screens, and screens consume more battery than a mostly-idle CPU.
> The A14 will be able to run in MacBooks at a higher power level while still remaining cool, providing for more cores and faster processor speeds.
You still seem to be confused by this. "A higher power level" and "remaining cool" are the opposite of one another. If they use the power budget to have as many cores as e.g. a Ryzen 7 4700U then it will probably use about as much power on the same process node. They're not magic, they're just making different design trade offs for phones than AMD and Intel make for laptops.
Also, you can install Android on x64 tablets with basically equivalent single-thread performance and significantly better multi-thread performance. Hardly anybody does this because hardly anybody cares that much about performance on that class of device as long as it meets a particular threshold that any non-garbage modern CPU (including many Qualcomm chips) already does.
But like they did with the SE, they will lower their parts costs to chase volume. Like they did here, by keeping the same iPhone production lines that built the iPhone 8 going (I'd presume it would be the same production lines that built their phones all the way back to the iPhone 6).
> But like they did with the SE, they will lower their parts costs to chase volume.
This has very little to do with their parts cost. Their problem is that smartphones are increasingly becoming a low margin commodity because the existing hardware is "good enough" and many people would rather save hundreds of dollars than have a phone which is marginally faster in a way they can barely notice. So the price of Android phones goes down over time and on some level they have to compete with that.
But that's true independent of their costs. If Android phone prices get lower and Apple's production costs stay the same, they have to lower prices at the expense of margins or they'll start losing volume to Android. If Android phone prices stay the same and Apple's costs go down, nothing is making them sell for a lower price, so they just get to enjoy higher margins.
In theory there is an optimal sale price which maximizes profits and changing unit costs can affect what it is, but it isn't even guaranteed to be lower when margins are higher. Apple's brand is as a luxury product. If they lowered prices too much at the low end for any reason, that impairs exclusivity and their ability to extract very high margins at the high end.
SPEC is the industry standard for comparing performance across platforms.
>Last year I’ve noted that the A12 was margins off the best desktop CPU cores. This year, the A13 has essentially matched best that AMD and Intel have to offer – in SPECint2006 at least. In SPECfp2006 the A13 is still roughly 15% behind.
https://www.anandtech.com/show/14892/the-apple-iphone-11-pro...
Apple's laptop part will be the first time they have a level playing field on both.
The Apple laptop part is also going to have a process node advantage. It will be built on TSMC's 5nm node while AMD is sticking with an enhanced version of TSMC's 7nm node and Intel will still be on 10nm.
The word was that Apple's entry level laptop part will have 8 next generation Big cores and 4 next generation little cores, which is quite aggressive for an entry level laptop part. The A13, in comparison, only has two big cores.
We'll see how Apple's laptop chip fares when it ships.
This is much less of an issue for single thread performance because a single thread isn't going to exhaust an 8-core CPU's full power budget anyway, or if it does it's by boosting to very high (i.e. less power efficient) clocks which only give a few extra percent more performance. The main thing the extra power budget gives you these days is the ability to have more cores, which x64 processors already have.
> It will be built on TSMC's 5nm node while AMD is sticking with an enhanced version of TSMC's 7nm node and Intel will still be on 10nm.
AMD already has 5nm processors on their roadmap. Apple might release theirs first, but having a modest advantage for a few months until your competitors release their next design isn't especially groundbreaking. It's basically the expected state of affairs when companies compete with each other and neither has a clear advantage -- the one released most recently is a bit faster and then soon the competitor has a release and it goes back the other way.
In the next generation, Apple is going to gain a process node advantage, and level the playing field for the first time in respect to power and thermal envelope.
Given that they were already in the same ballpark as Intel and AMD without any of those advantages, we're in for a sea change.
The point is that "current generation" doesn't work like that. If AMD releases Zen3 on 7nm and Zen4 on 5nm and the ARM Macs arrive half way between one release date and the other, which one do you want to compare them to? Neither would be strictly "fair" because one or the other would have an advantage of several months of advancement one way or the other. But doing all of this just because they couldn't wait a few more months for Zen4 doesn't make much sense.
We're not going to really know until ordinary people can get it in their hands and benchmark it.
The whole thing strikes me as Apple seeing Intel stagnating and starting a long-term process in motion several years ago which was too far along to stop by the time AMD brought competition back to the market again.
Zen 3 and Apple's laptop chips are both due in the fourth quarter of this year.
Apple is going to have a process node advantage at least until Zen 4 ships, which definitely will not happen only a couple of months after Zen 3 ships.
However, despite Apple having a better hand to play than AMD or Intel for the first time, I agree that the proof is in the pudding.
This was a problem in Geekbench 4 (especially so on Android) and was resolved in Geekbench 5.
You can believe the numbers to be roughly accurate within the sampling variation.
Every time I see someone write about the benefits of "Apple Silicon" they always mix this up. What they do is say that the ARM chip is much more power efficient and therefore it will have a longer battery life and can be cooled passively. In the next sentence they say you can sacrifice battery life for higher performance and add more powerful cooling. Yet for some reason people seem to think that you can have both at once without any of the downsides. ARM based SoCs are usually power efficient because of asymmetric cores and low frequencies.
AMD created a symmetric 8 core chip running at 10W simply by cranking the frequency down to 1.8GHz. If they add asymmetric cores they could lower power consumption even further.
Equally the Linux desktop offering have advanced and by that, more palatable for the average user for a desktop.
Certainly when it comes to most desktops in business, something that can open word and email and odd spreadsheet and powerpoint, that's the bulk of usage right there. Small lower power usage alternative like Rpi4 would do those jobs fine as for most users. So Apple may actually shift that cultural hurdle that prevails in business, maybe.
> think faster, svelter laptops actually lasting 10 hours on a battery charge.
...and I instantly think of HP's Envy x360 w/Ryzen 4500U pulling almost 14 hours of battery life in productivity tests.
I Apple wants to stand out, it will have to push closer to 20hs on a charge. Not saying they can't, saying it'll be a stretch. And great if they pull it.
[0] I'm considering Zen 2 "the best performing x86" (with a small grain of salt). If Intel wants to take that crown with Tiger Lake I'm all for it.
like what are the odds of qualcomm today matching the A12 in 5 years even?
I still worry about us entering in to a monoculture.
Mediatek, Samsung, HiSilicon (Huawei) are the high-end competition that produce competitive smartphone SoCs with the latest ARM cores.
Rockchip and Allwinner are the low-mid-end competition that produce a huge volume of low cost smartphone SoCs.
For smartphones, since anyone can license an A76 (or better) core from ARM themselves, Qualcomm's key moat is really in their modem IP - where the competition is Huawei, Unisoc, and whoever ends up purchasing Intel's out-to-pasture 5G division.
Amlogic, Broadcom (Raspberry Pi), Cavium (Servers + NAS appliances), Renesas (Automotive), NVIDIA, Ampere (Servers), and now Amazon (Graviton2) are all major ARM SoC manufacturers somewhat outside the direct Android smartphone space.
You can see how little it threw at creating competitive processors for wearables because the market just wasn’t there compared to Apple.
Look at https://www.google.com/amp/s/www.androidcentral.com/samsung-...
https://www.gizchina.com/2020/06/14/the-average-selling-pric...
The article is bragging that the ASP of all Samsung phones is $292. That’s $100 lower than the lowest cost iPhone.
Apple also captures 66% Of mobile profits.
https://www.counterpointresearch.com/apple-continues-lead-gl...
For the workstation they don't have anything much above a RPi4 but they have an interesting MiniITX form factor for a 7-way cluster and are even working on an IP camera. Really cool set of devices.
Multicore benchmarks say it is around the performance of 3 year old desktop chips. Not bad but the single core performance is probably awful.
https://www.phoronix.com/scan.php?page=news_item&px=HoneyCom...
Intel will be fine for at least the next decade. Businesses will still run on Intel, as will cloud. They'll figure out their process trouble and retake the performance crown again from AMD, although by a narrower margin. All of this has happened in the past. This is due to one simple truth: businesses couldn't give less of a shit what hardware their software runs on, and moving everything to ARM is so expensive, it's not going to happen for existing apps.
Most of us here would have a bout of OCD knowing we're not using the latest and greatest thing, but businesses don't have such a problem.
Apple has never seemed to really be into the concept of backward compatibility; but historically the rest of the industry has been.
Whatever happens, I'll bet it won't be boring.
Hold your horses. No one even knows how it will perform, how Rosetta 2 will perform and if Intel is not going to hit back as it did with Qualcomm and Microsoft.
Seeing is believing, after all.
And Rosetta 2 has already been benchmarked, and the results were amazing even on a two year old iPad processor.
Intel has nothing to hit back with.
Rosetta: usually recompiles once ahead of time, only supports 64-bit
Microsoft: seems to recompile on the fly, only supports 32-bit
Also notable is the fact that the leaked benchmarks show the A12Z is generally faster when emulating an Intel CPU with Rosetta 2, than the Surface Pro X is when running natively, and the A12Z isn’t even Apple’s real desktop cpu, it’s just what they had on hand from the iPad (!)
Is this marketing move?
I think this ignores the historical reasons for x86 success. Intel made their bones by crushing Unix boxes and mainframes with high volume x86 parts. PC CPUs were the highest volume microprocessor for decades, and they leveraged that into making high performance, yet low cost server and workstations parts to buff up their margins.
Well, the tables have turned - ARM is far, far higher volume than x86 due to its ubiquitous use in smart phones. And ARM is going to do to Intel what Intel did to IBM, Sun, SGI, HP, DEC, etc. The problem is the same as it always has been: NRE (fixed costs) for new chips climbs ever higher and needs to be offset by even more volume to be acceptable to consumers.
ARM won that game in 2007. It's just taken this long before market forces and product maturity have forced Intel's hand. To be fair, Intel may have staved off the inevitable for a few years without their well publicized 10nm production stall.
While this is certainly true for bleeding edge nodes, hasn't the NRE become significantly lower for mature process chips? I would guess that EDA tools and a greater availability of IP blocks would drive down the real dollar cost of designing your own ASIC on a node 1-2 generations back.
Honestly, I'm not familiar enough to have a nuanced opinion, but from what I understand, that's probably true.
However, in the battle for cloud VPS and PC CPU sockets, using a 1-2 generation old process doesn't appear to be a winning strategy.
Smart phones, tablets, TVs, projectors, routers and also refrigerators, 3D printers and ovens.
Their volume is 10x to 20x times bigger than PC and laptops.
If this is the case then the problem for Microsoft is losing Windows market share to macOS.
The $329 10.2 inch iPad is a great buy. The $499 inch iPad Air is basically a rebranded first generation iPad Pro with a better processor.
Apple bought a lot of the features from the iPad Pro to the Air and low end iPad including pencil support and the keyboard attachment. They also finally left the 9.7 inch form factor to change the size to 10.2.
The iPad is plenty powerful enough for most users.
The MacBook Air has been $999 forever outside of a brief time when it was more expensive and a usable Mac Mini has usually cost $799. Not $699.
The iMac have also maintained their prices.
>Kuo predicts that the releases will "follow iPhone SE's product strategy." Specifically, he says that the models will be affordable, and adopt Apple's "fast chips" similar to how the iPhone SE used the chip found in the flagship iPhone 11 and iPhone 11 Pro models from the fall of 2019.
https://appleinsider.com/articles/20/05/14/108-inch-ipad-exp...
It would certainly be interesting to see Apple apply this same strategy to an entry level laptop.
See "History of the 8086" from here: http://www.righto.com/2020/06/a-look-at-die-of-8086-processo...
See "History of the ARM chip" from here: http://www.righto.com/2015/12/reverse-engineering-arm1-ances...
It's depressing to think about how Intel was putting so much energy into a chip that never succeeded and succeeded with a chip that they never intended to keep around.
But…does that matter all that much to Apple? They seemingly have made a recommitment as of late to the pro and prosumer market (see: Mac Pro), but for a long time there, they sold laptop Macs, and begrudgingly sold desktop Macs because, I don’t know, education sales. If they nail MacBook Air-class power envelopes, which I have absolutely zero doubt they can, and they don’t knock a new iMac Pro out of the park…is that really going to hit their bottom line all that much?
(Yes, in some way, there’s a halo effect, and developers won’t be happy, nor will creative pros, but if they get great MacBook Airs with higher profit margins, then they probably consider that a win.)
For example, it was practically impossible to get Zen 1 CPUs over 4 GHz on air, not because they got too hot, but because they'd just crash almost immediately, regardless of voltage. Similiarly you can't push the current-est Intel generation beyond ~5.3 GHz, regardless of voltage. Current Zen 2 CPUs are highly binned, some bins cannot go beyond ~4.2 GHz, regardless of voltage and cooling, some can do 4.7 GHz.
I can't wait to see what Johnny Srouji's team can do when they're not targeting a mobile device.
Apple alone sells more ARM based devices than the entire personal computer market and probably the server market. Apple also has more money to throw at it. Intel couldn’t make 4G/5G work and sold that division to Apple.
On the server side, AWS is designing its own chips (disclaimer I work for AWS but nowhere near that division).
Even today, AMD is very competitive against Intel and I think JLG's point was that Apple's move to ARM will boost the entire ARM ecosystem significantly.
Qualcomm and Nvidia won't miss a chance to try to dethrone Intel. AMD, too.
Intel could be in danger.
https://news.ycombinator.com/item?id=5589410
Kinda prophetic right?
Clayton Christensen‘S Innovators Dilemma at work.
It’s amazing how Apple with just 7% of market share can/could change the entire industry. They did the same with Flash(for the better) and this time with the PC. Interesting times ahead
(2) Apple's announcement had no discernible effect on Intel's market cap, which, even after adjusting for inflation, is higher today than it has ever been except for an interval of about 2 or 3 years at the end of the dot-com boom.
(I used a CPI calculator, a market-cap chart and one of my eyeballs to estimate the duration of this interval.)
But Intel already had been beaten by AMD, and for the foreseeable future as well. Even on the mobile side, AMD has more performances with less power.
How will Apple stack up with AMD is the real question. This move is great for their low powered devices. But when it comes to The Pro devices, I'm sure they would have more performance for the next half decade if they just moved to AMD.
I know this meme gets brought up a lot, but has this claim been backed up by benchmarks that's not geekbench?
At IDA Pro, the DTK is actually faster than a 16 inch MBP in practice... (and that's with an x86_64 copy of IDA... as an arm64 one doesn't exist yet, workload being tested was loading an iOS kernelcache)
1. The A12Z in my iPad has 2 more vortex cores than the base A12 and 4 more GPU cores.
2. I'm comparing to my laptop, a T470 with i7-7600U in it.
3. My (I did explicitly say that) laptop's modern equivalent, an i7-8565u, is dog shit compared to the iPad still and costs 100% more.
Some benchmarks if we need them:
https://browser.geekbench.com/processors/intel-core-i7-8565u
https://browser.geekbench.com/v5/cpu/2898525
Now if you do any video processing on the ipad, it'll chunk through 4k like my Ryzen 3700X does if it's HEVC and barely even get warm, but that's the T2 ASIC in it.
The point being, if you compare a modernish "professional laptop" with an i7 in it, to an iPad, the iPad will rip it a new orifice.
ARM + ASICs for other functions (ML/security/codec) run rings around any general x86 CPU on performance per watt for sure.
Comparing as you did to a Xeon with 165W TDP is comedy but illustrative yes.
>https://browser.geekbench.com/processors/intel-core-i7-8565u
>https://browser.geekbench.com/v5/cpu/2898525
My original question: "but has this claim been backed up by benchmarks that's not geekbench?"
>Now if you do any video processing on the ipad, it'll chunk through 4k like my Ryzen 3700X does if it's HEVC and barely even get warm, but that's the T2 ASIC in it.
>ARM + ASICs for other functions (ML/security/codec) run rings around any general x86 CPU on performance per watt for sure.
Comparing hardware accelerated HEVC performance to software cpu performance isn't fair and is shifting the goalposts. If you're going by that logic you can also say that a snapdragon 865 (hardware rendering) has "considerably more grunt" a 64 core threadripper (software rendering). While we're at it, we can also compare A12 to intel/amd in other aspects, such as memory support, floating point performance, and sustained performance, all of which I'm confident A12 wouldn't do well in.
>Comparing as you did to a Xeon with 165W TDP is comedy but illustrative yes.
So? The xeon cpu also has 28 cores and 56 threads, whereas the A12 only has 2 "big" cores (the ones being benchmarked). Dividing 165 by 28 gets you to a more reasonable TDP of 5.89W per core.
This review[1] compares the SPEC2006 score of the A13 to high end modern desktop chips. The A13 is very close despite its power consumption being much lower.
[0]: https://www.anandtech.com/show/13661/the-2018-apple-ipad-pro...
[1]: https://www.anandtech.com/show/14892/the-apple-iphone-11-pro...
The 3700X contains its own encoding blocks which would make a better comparison.
I'm sure the A14 or whatever comes along, with higher core count and better thermals for desktop use, will do the same for desktop CPUs.
An i7-10510U (which is the closest thing I could find, an upper mid range ultrabook part) is significantly faster than an A12Z.
I noticed because I wanted to come back to read more comments and had a hard time finding it.
Apple will be using TSMC's new 5nm process for the first version of it's laptop class chips, so they will have a process node advantage in the upcoming generation.
>Last year I’ve noted that the A12 was margins off the best desktop CPU cores. This year, the A13 has essentially matched best that AMD and Intel have to offer – in SPECint2006 at least. In SPECfp2006 the A13 is still roughly 15% behind.
https://www.anandtech.com/show/14892/the-apple-iphone-11-pro...
Just read the article you linked.
>Apple’s marketing materials describe the A13 as being 20% faster along with also stating that it uses 30% less power than the A12, which unfortunately is phrased in a deceiving (or at least unclear) manner. While we suspect that a lot of people will interpret it to mean that A13 is 20% faster while simultaneously using 30% less power, it’s actually either one or the other.
>In the face-off against a Cortex-A55 implementation such as on the Snapdragon 855, the new Thunder cores represent a 2.5-3x performance lead while at the same time using less than half the energy.
https://www.anandtech.com/show/14892/the-apple-iphone-11-pro...
When compared to other designs that operate under the same cooling and power constraints, there is absolutely no competition.
No surprises if Windows eventually moves to ARM. (And if servers do too.)
I disagree that Apple is only better at marketing things like this, they're much also better at doing it. And their processors are presumably going to be ahead of anyone else's ARM based computers, just like they've been with phones.
https://docs.microsoft.com/en-us/surface/surface-pro-arm-app...
"Some of your software will work on this" is a throwback to Windows XP 64-bit edition, people don't want to deal with that.
The limited emulation support means it's more capable than Windows RT's "native ARM software only" was, sure. But it's even less straightforward to explain to a potential user whether or not they could use it.
This was confirmed coming soon, and will almost certainly be available by the time the new Apple Silicon devices are released.
> Nor will games using OpenGL versions above 1.1, among other things.
OpenGL is just a userspace thing that gets loaded into your address space, and the specific installed dll (ICD) will convert it into hardware ioctls. Microsoft only ever implemented GL 1.1 (as a DX wrapper) and newer versions were always provided by AMD/NVIDIA/Mesa as an ICD.
If the vendor drivers didn't implement anything newer than 1.1 you can still do it yourself. What accelerated interface is available? ANGLE will give you 3.0/3.1 over DX9/DX11 (ships with Chrome and some Windows features); Zink will give you OpenGL 3.0 over Vulkan; the old QindieGL will give you GL 1.4 over DX9. And of course Mesa llvmpipe will give you 4.2 in software.
If Apple’s latest ARM CPU in 2020 was the A7 instead of A13, ARM chips in MacBooks would’ve been equally uninteresting.
It's more complicated than that: It's all about understanding disruptive technology. Arm is disrupting x86 / x64.
The thing with disruptive technology is that being ahead of the curve is just as bad as being behind the curve. Microsoft has no benefit from "pushing" Arm, all they need to do is make sure that Windows is ready to run on Arm when someone's ready to sell an Arm desktop computer.
But, Apple, as a hardware company, needs to be much more proactive about where the market is going and actively switch its hardware. If this means that Microsoft's first major Arm market for Windows are VMs on Apple, so be it. The rest of the PC makers will switch or be disrupted.
Readers of this site will of course understand exactly what is being discussed, but I'm concerned that the overuse of this marketing term confuses things for non-technical readers. There are already more common, and more accurate terms that can be used instead.
GenuineIntel
AuthenticAMD
AppleSiliconMy point was that we don't refer to AMD CPU's as "Authentic AMD" in general speech. However, in Apple's case they have succeeded in promoting their own narrative.
I understand that this happens, but I was just intrigued by the fact that they managed to do it. The difference in terms does have a subtle effect in conversation however.
But it’s disingenuous to compare benchmarks for single threaded performance between the ipad and macbook.
I can’t do any ‘very’ serious productivity work on the ipad that requires multiple apps, period. (try moving a photo between multiple photo-editing apps, ha.)
I hope that is just a ‘feature’ of iOS, or else macOS will suffer greatly as people leave it for more productivity based OS’s.
[0]: https://www.anandtech.com/show/14892/the-apple-iphone-11-pro...
Steve Jobs reveal transition to Intel: https://www.youtube.com/watch?v=ghdTqnYnFyg
i don't know if the history will repeat itself but it will be interesting to watch.
If the last few years are to be a guide, it will crash and burn. Apple hasn't been developer and power user friendly for some time now. Focusing on thinner laptops while sacrificing the keyboard and the touchbar to name two. I've been hearing from developers being skeptic who are planning to buy before they switch to Arm. Which is a different vibe when they switched to intel.
It could be a massive marketing stunt from MS when they announce a more like MacOS version of Windows which actually runs on Intel.
macOS is the last major OS getting ARM support. iOS and Android have been built with ARM support from the very beginning. The first commits working on ARM support for Linux have been made more than a decade ago. The first publicly released Windows on ARM hardware came out in 2012, and MS has undoubtedly have been working on it for years before the commercial release.
If anything would trigger an ARM transition on a massive scale, it would likely be the work invested in the high performance ARM hardware by the companies such as Amazon and Fujitsu.
edit, source: https://github.com/mpe/linux-fullhistory/commit/3c2a0eb7ba89... ("initial ARM support", Linux 2.1.80, 20.1.1998)
Debian has had an ARM port since 2000.
People have a choice with the first two, not the third. This is why Apple going forward with this is a big deal.
It's highly likely that Apple has maintained internal macOS ARM builds for every successive release since, if not since before the Intel transition. They had similarly maintained x86 builds of macOS from the time they began its development after the NeXT acquisition.