Slashdot Mirror
Intel Is in an Increasingly Bad Position in Part Because It Has Been Captive To Its Integrated Model (stratechery.com)
Once one of the Valley's most important companies, Intel is increasingly finding itself in a bad position, in part because of its major bet on integration model. Ben Thompson, writing for Stratechery: When Krzanich was appointed CEO in 2013 it was already clear that arguably the most important company in Silicon Valley's history was in trouble: PCs, long Intel's chief money-maker, were in decline, leaving the company ever more reliant on the sale of high-end chips to data centers; Intel had effectively zero presence in mobile, the industry's other major growth area. [...] [Analyst] Ben Bajarin wrote last week in Intel's Moment of Truth. As Bajarin notes, 7nm for TSMC (or Samsung or Global Foundries) isn't necessarily better than Intel's 10nm; chip-labeling isn't what it used to be. The problem is that Intel's 10nm process isn't close to shipping at volume, and the competition's 7nm processes are. Intel is behind, and its insistence on integration bears a large part of the blame.
The first major miss [for Intel] was mobile: instead of simply manufacturing ARM chips for the iPhone the company presumed it could win by leveraging its manufacturing to create a more-efficient x86 chip; it was a decision that evinced too much knowledge of Intel's margins and not nearly enough reflection on the importance of the integration between DOS/Windows and x86. Intel took the same mistaken approach to non general-purpose processors, particularly graphics: the company's Larrabee architecture was a graphics chip based on -- you guessed it -- x86; it was predicated on leveraging Intel's integration, instead of actually meeting a market need. Once the project predictably failed Intel limped along with graphics that were barely passable for general purpose displays, and worthless for all of the new use cases that were emerging. The latest crisis, though, is in design: AMD is genuinely innovating with its Ryzen processors (manufactured by both GlobalFoundries and TSMC), while Intel is still selling varations on Skylake, a three year-old design.
The first major miss [for Intel] was mobile: instead of simply manufacturing ARM chips for the iPhone the company presumed it could win by leveraging its manufacturing to create a more-efficient x86 chip; it was a decision that evinced too much knowledge of Intel's margins and not nearly enough reflection on the importance of the integration between DOS/Windows and x86. Intel took the same mistaken approach to non general-purpose processors, particularly graphics: the company's Larrabee architecture was a graphics chip based on -- you guessed it -- x86; it was predicated on leveraging Intel's integration, instead of actually meeting a market need. Once the project predictably failed Intel limped along with graphics that were barely passable for general purpose displays, and worthless for all of the new use cases that were emerging. The latest crisis, though, is in design: AMD is genuinely innovating with its Ryzen processors (manufactured by both GlobalFoundries and TSMC), while Intel is still selling varations on Skylake, a three year-old design.
"instead of simply manufacturing ARM chips for the iPhone"
What's simple about it? Intel's ARM was Xscale, which was based directly on DEC's StrongARM (which they purchased.) It was the fastest ARM core at the time, but while it [x]scaled up, it didn't [x]scale down. It had the highest power consumption at low clock rates of all the ARM cores.
Intel did not have an ARM-based product which would have been a viable core for the iPhone.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
I think this kind of analysis is quite premature. Presently, there is no mobile-worthy x86 option -- for lots of reasons. Until there is, I don't think you can judge Intel for their direction.
Presume, for a moment, that in a few years, Intel successfully produces an x86 proc for mobile specifications. It's distinctly possible, indeed even probable, that ARM becomes useless, and the entire mobile market moves to x86. What a boon for Intel to have not wasted time and effort during these middle-ground years.
We've lived through this before. I refer you to WAP. How many web developers spent how many hours fumbling through WAP-limited options, before the entire mobile market moved to full web technologies? What a wasted investment for any small company. And what a horrible experience in was for consumers.
We'll wait and see.
And that power-per-watt disadvantage vs ARM predates Intel's integration strategy and also their current process-size disadvantage. I don't see any evidence to the contrary in the linked story.
This just marks the end of an era. Moores Law is dead (and has been dead for quite some time). Intel will need some other way to innovate. All they have been doing is adding cores and trying to push up clock speeds.
Moore's law is about transistor counts. Adding cores adds transistors.
Even this is running into a dead end: because of physics.
They can still add cores for some time, if they can improve yields. First there is a process shrink and cores shrink, then the process is improved and cores grow again. Then we get a new process...
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
I thought Intel was in a bad position because it decided to dump $300m at diversity initiatives and fire a bunch of engineers instead of investing that money in R&D like AMD has.
Om, nomnomnom...
I still say it's Compaq, not MS, that deserves the credit here. Compaq created the PC compatible. Microsoft did have the foresight to make sure its software worked too, but at the time the money was in hardware and you more or less expected incompatibilities between generations of the machine. The idea of cross-vendor compatibility in the micro market was Compaq's, for sure.
Wrong. Moore's law is about transistors *per unit area*.
In other words, increasing transistor *density*.
That hit a wall some years back which is why they starting adding multiple cores per CPU.
My point still applies: Intel is an important company in Silicon Valley, but not the "most important" (because their success depended on another company; the location of that other company is irrelevant).
Please let me know if I can help you with anything else.
It won't die, but it will need to find new markets in order to continue growing like investors demand.
This is why Microsoft is doing it. The realized they are not beholden to Intel. They made Windows RT (port of Win32 to ARM) so if the Intel x86-64 ship ever sank, it wouldn't take Windows down with it. They don't need it to sell like hotcakes; heck they don't need it to sell at all. They just need to to be there and ready if ARM overtakes Intel. It's insurance - a hedge against Intel imploding. If that should happen, they'll just transition to Windows for ARM, and all the software companies making Windows apps will (more or less) simply recompile their programs for ARM64, and Windows will carry on as if Intel never existed.
Intel is a company that has a lot of good things floating around, choked by a management that can only think one way. x86
>> shitload of resources and inertia
Agree they have these, but both of these are working against Intel right now: the inertia is what drove them into the ditch while more nimble chipmakers were passing them by, and their ample resources are blinding them from the danger because they assume they can always write checks to get back on the right track if they ever figure it out.
See "Sears"...
This was clear a long time ago. Intel was making X86 mobile chips for Intel to gain market share. Not because the phone makers wanted x86 chips. It was Intel-focused, not customer focused. Microsoft did similar things with Windows 8 and that metro junk.
Recently Intel has branched out into lots of other growth businesses though, buying Movidius, Altera, and MobileEye. They're making silicon photonics chips for optical networks, DOCSIS chips for cable modems and 3D Xpoint RAM to bridge the gap between DRAM and NAND. They integrated an AMD GPU and they are building a new GPU of their own.
It’s ironic that articles like this gain traction after Intel has already turned around and started to gain traction.
i've pointed this out here on slashdot a number of times, dating back at least... six years possibly more. the first really clear signs were when ARM came out with the first dual-core ARM Cortex A9 side-by-side demonstration of running a web browser (linux desktop OS) side-by-side with a 1.6ghz intel Atom. it kept up and in some cases loaded pages before the intel processor. at the end of the demo they showed the clock rate of the ARM chip: only 600mhz.
intel was a memory company. they're proud of their heritage. they designed the world's most efficient and compact memory-efficient instruction set because memory was damn expensive. if you got more instructions into memory, you ran faster, you needed smaller caches, and your product was cheaper. ... except... decoding those instructions takes time. you now have to run the clock at twice the speed of a RISC core in order to decode those "compact" instructions into the same equivalent RISC ones. and that's where things go wrong for intel, because power consumption is a SQUARE law. if the clock rate has to be double, the power consumption is FOUR times greater.
i've made comments regularly about this: it's only because intel was putting vast sums of money into foundries, staying at least one geometry ahead (28nm when everyone else was using 40nm), that nobody really noticed or complained too much, because by being one geometry ahead you reduce power by a factor of 2. ... but they're no longer ahead, now, are they?
now that the power advantages of geometries are beginning to run out (as well as the cost being higher and the yields lower), intel's *really* in trouble, and it all boils down to the design of the instruction set.
they have one hope left: abandon x86 and start making non-x86 instruction set designs. it'll be a really *really* tough sell, but if they can do that they have a chance.
Since open source applications can simply be recompiled to any processor archtitecture.
That's a neat idea and you might be right.
If it weren't for Compaq, then what we all think of as generic commodity PC would still just be thought of as "IBM's particular computer, which happens to be quite a bit behind most everything else." There were a fuckton of competing computers in the 1980s, most of which were quite a bit more advanced than the x86 stuff. But everything was proprietary, so even if you were in love with something, in the end you couldn't trust it to stick around (e.g. Amiga).
Who knows how things might have gone if Compaq hadn't nailed down generic commodity hardware as being what it turned out to be. Maybe we simply wouldn't have generic commodity hardware (but there might be some generic software layer or something like that), or maybe it would have ended up something completely different (e.g. 68k based?).
Well 'integration' isn't the word I'd pick, they have some lockin if they make a market x86-dependent, and so that was their goal. The assumption would be that if the mobile market became mostly x86, then sure Android would have ARM compatibility, but x86 would be optimal and no one would tolerate the crappy non-native experience. Of course, the glaring flaw is that Intel would have to *live* in that unacceptable non-native experience to begin with, and Intel was right about one thing, no one would put up with such a crappy experience.
Larabee was hubris that a lot of sort-of x86 cores would mitigate the GPU accelerated demand, because even if you couldn't be quite as quick as nvidia, you could use a familiar programming model. Problem was that Phi *also* required developers to be more careful and picky, so it wasn't like programming in x86. By the time Intel could have possibly made it easier, the world was just so used to CUDA that the market was slim. They may have better luck with AVX512 in Xeon Skylake, but who knows. It was always doomed as a GPU because they have no competency.
Another problem is being in denial, taking a long time from changing gears from 'no competitor' to 'oh, AMD is competitive again'. In the datacenter, Intel had crazy high core counts. In the desktop? quad-core, because no competitive pressure. When zen was rumored, Intel was skeptical, and when Ryzen came out they were slow to change. Compared to desktop offerings, AMD was so much better. On the server side, things are a bit more mixed (where Intel actually *has* continued to invest in meaningful advances). For example, desktop core counts have been stagnant, as were clock speeds, and no AVX512, meanwhile server chips moved on and had all those improving. AMD still has more PCIe lanes and memory channels, but it's a caveat, more like 4 processors with 2 memory channels and 16 pcie lanes each rather than 1 processor with 48 pcie lanes. This is a distinction that doesn't matter for many workloads, but for a few, it matters (the memory performance of a single threaded application is much better on intel server than AMD server).
XML is like violence. If it doesn't solve the problem, use more.
Just curious: which one of these are you talking about? Most of them don't mention what you're talking about in the product descriptions or even user flames/reviews.
My personal experience is that even mini-ITX cases usually have slots on the back and room for 2.5" drives, and you just get more and more slots and actually-usable-3.5"-drivebays as you go up to mini towers and mid towers and full towers. You've obviously ended up with some different brand, and I'm just trying to figure out what that is.
As of 11:37 EST, Intel's stock price is $50.16, and AMD's stock price is $14.61.
You are welcome on my lawn.
So you are claiming that Moores Law is alive, even though Intel itself said it is dead? Guys like you are delusional and think that CPUs are continually going to get faster and faster. But yeah, I am an idiot even though you can't even see what is happening.
And the hostility towards diversity here on Slashdot is misguided and idiotic. When I see white boys from upper middle class families complain about being oppressed and how it's a meritocracy in technology, I find it hysterical that they can't see outside their little bubble and realize that they had all of their opportunities handed to them.
Yes, us half-asians(or asians), who are penalized in US(don't live in the US anyway) university admissions akin to whites and are born to poor working class families, where name brand kraft dinner was a luxury sure are 'white boys from upper middle class families.' Nothing like finding out that a university specifically penalizes you because of your race, instead of making the selection based on best candidates? Yeah, those of us who climbed up from the bottom really do like meritocracy, because we know that people got there on skill, ability, and competence. Not because they had the current trendy gender selection/sexual preference, had the right colour of skin, or some other *insert non-selective trait* that they were born with/without.
because you were smart enough to pick the right parents and genes so that you have natural aptitude in a lucrative field.
I'm not sure how you got so far in life being so stupid as to believe that a person picks their parents. Or believing that if a person truly wants to, they'll rise through ability and skill instead of moping around going "woe is me." Ever wonder why those of us who were dirt poor are the most angry at 'diversity' bullshit? Nobody has a problem if programs are open to everyone. A lot of people who worked hard for their job and skillset however are pretty pissed off when someone else coasts along because it looks trendy as fuck to the company they're working for though.
And god forbid if a company doesn't for the bullshit that girls are not as good at math and science as boys.
Uh-huh. So that's why in so many places, they also lower the aptitude and physical requirements for policing, fire fighting, military enlistment and make everyone either pick up the slack for them. Or actually endanger the lives of everyone else because they're unable to deal with the demands of the job. Funny enough the most outspoken people against this, are the ones that passed the actual requirements before all that "diversity" bullshit was being pushed. Why? Because people believe they didn't get there on ability, skill, prowess, but were handed the job because it looked good. I mean, what's it going to take? Another dead fighter pilot that would have been drummed out if they were male. Or another ship nearly sheered in half because two women had a snit, and refused to talk to each other? Or an entire fire dept., refusing to work with someone because they couldn't even carry a hose and put the people they were supposed to rescue in danger.
Om, nomnomnom...
Every single tech company in Silicon Valley depends on the success of other tech companies, in or out of the Valley.
Moore's law is about transistors per unit area. Adding cores increases both. Only new manufacturing techniques to cram in more transistors will let the trend continue, and they are indeed pushing the limits of what's physically possible.
At 7nm we're talking features that are only about three dozen atoms wide. The current roadmap has 5nm production in a few years. This kind of thing is well outside my knowledge but I'm pretty confident you can't make devices smaller than a single atom, so they are rapidly approaching a wall one way or another!
=Smidge=
"Moore's law is the observation that the number of transistors in a dense integrated circuit doubles about every two years"
That was almost coherent.
Guys like you are delusional and think that CPUs are continually going to get faster and faster.
Moore's Law does not say CPUs will get "faster", only that transistor density will increase.
There was a really big turning point with MCA vs PCI as well. That's when it became plain that it was no longer "the IBM compatible".
When IBM couldn't force MCA as the standard it became plain that it had lost control over the direction of the design, and that we were now in a commodity hardware multivendor world.
Nope. Moore's Law is about transistor density per unit area. Practically, that translates into more powerful processors. Moores Law is dead, and with it goes the promise of ever increasing processor power.
It won't die, but it will need to find new markets in order to continue growing like investors demand.
Intel has a PE below 15. The S&P average is over 25. So investors aren't expecting much growth.
I never said Moore's Law was about speed (it is about density). But practically, it means that processors will get more powerful. Guys like the poster think that is going to continue forever. But it won't (and hasn't).
I don't think there is anything they can do. Eventually, they will share the x86 market with AMD about even and maybe even some additional manufacturers. And they will come under increasing pressure from ARM, or at least Intel will. AMD does make ARM chips, even if not at volume at the moment.
It is the typical way giants fail: By sleeping, sleeping, sleeping, until they find that they are not very relevant anymore.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
Moore's law is about transistors per unit area. Adding cores increases both. Only new manufacturing techniques to cram in more transistors will let the trend continue, and they are indeed pushing the limits of what's physically possible.
At 7nm we're talking features that are only about three dozen atoms wide. The current roadmap has 5nm production in a few years. This kind of thing is well outside my knowledge but I'm pretty confident you can't make devices smaller than a single atom, so they are rapidly approaching a wall one way or another! =Smidge=
Let's break this down and make it simple. Moore's law is the observation that the number of transistors in a dense integrated circuit doubles about every two years. Simply put this is the founded definition. Also, Moore's prediction proved accurate for several decades, and has been used in the semiconductor industry to guide long-term planning and to set targets for research and development. Advancements in digital electronics are strongly linked to Moore's law: quality-adjusted microprocessor prices, memory capacity, sensors and even the number and size of pixels in digital cameras. Digital electronics has contributed to world economic growth in the late twentieth and early twenty-first centuries. Moore's law describes a driving force of technological and social change, productivity, and economic growth. For those that need a more in-depth understanding of Moore's Law go here: https://www.britannica.com/tec... Why? That we are all on the same page about the true definition of Moore's Law.
They can add cores, but it will not matter much. Most workloads are not core-limoted these days.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
They won't die. No.
The problem is that their market is shrinking:
there are plenty of people who still need or want high performance processors.
Yup, earning millions on specific contracts to build giant HPC center every few months seems lucrative.
(What will eventually become of Intel according to this trend).
Until you realize that there are billions of people on this planet, thus billions of pocket to fill with a smartphone.
("Pocket computer" metaphor in full force. In some region (older, second hand) smartphones are the only computers that people will ever come into contact with).
Intel didn't manage to become "the x86 of smartphone/tablet/other pocket devices", despite an initial strong (pun intended) start back in the PDA era.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
>> Every single tech company in Silicon Valley depends on the success of other tech companies
Not the ones that tap into the consumers themselves. (Google. Apple. Facebook. Uber. Etc.) Those companies have such market presence that they can line up quality, even name vendors (e.g., Foxconn, Intel) around them, and replace them like interchangable commodities as needed.
I have often seen Moore's Law formulated as transistors per unit cost, which I think is a useful measurement. If we hit a wall on feature size, this could still allow continually improved performance (and to a lesser extent, performance per watt).
I don't think there is anything they can do. Eventually, they will share the x86 market with AMD about even and maybe even some additional manufacturers. And they will come under increasing pressure from ARM, or at least Intel will. AMD does make ARM chips, even if not at volume at the moment.
It is the typical way giants fail: By sleeping, sleeping, sleeping, until they find that they are not very relevant anymore.
Intel has not been sleeping. They have engaged in one failed project after another. There's a difference.
Definitely not. Having spent enough time in the latter, and living in the former, I can say our weather is infinitely more tolerable.
If cellphone ARM chips can do 1.5-2GHz with little to no heat sinking other than though the phone case/chassis itself ...
They can do that because under typical workload, they are basically doing nothing. A modern x86 chip can do the same under similar conditions. Put some real load on your ARM chip and see how quickly it overheats. Playing some VR games such as Arcslinger on my Pixel/Daydream, I get an overheat message after about 5 minutes of gameplay, and at that point the phone is almost too hot to hold. I'm pretty sure overheating during VR use was the cause for my previous pixel bricking itself.
Yeah, those of us who climbed up from the bottom really do like meritocracy, because we know that people got there on skill, ability, and competence.
I am also one of those people. And I know why so many of us are so bitter about seeing handouts. We struggled, and we want to see other people struggle too. It's some kind of messed up desire for fairness, when really, we should be trying to make sure nobody else has to go through the bullshit we did to succeed. The part you missed in the above formula is actually the largest factor- luck. Get over yourself, asshole.
The use of the x86 instruction set wasn't the big issue with Larrabee. Larrabee would have been a bad idea no matter if it used ARM or MIPS opcodes instead. Using x86 didn't help, but that was just one among many issues of that architecture. The issues of the Larrabee architecture are things such as no fixed function hardware for things such as z-buffering or rasterization, not enough hardware threads to hide the memory latency, memory interface with not that much bandwidth but expensive but not that often usefull cache coherency, etc, SIMD units were not wide enough etc. Sure: Without x86 and with a simpler decoder things would have been slightly better.
Jan
Well except if you read the actual paper (PDF warning) that Moore wrote which created the whole concept, everything seems to be framed in terms of square area and component size.
=Smidge=
Glad you have your own definition of Moore's Law that is different from everybody else.
Intel moved too soon, it was too expensive, there was no transition plan, and they didn't get enough industry partners to buy in. It wasn't the first time nor will it be the last when a major industry player thought they could strong-arm everyone into a new platform only to be shown that they can't.
CPUs are getting faster, not GHz wise but every new generation has better benchmark numbers than the previous. I'd call that "faster".
Only the State obtains its revenue by coercion. - Murray Rothbard
Guys like the poster think that is going to continue forever. But it won't (and hasn't).
Nobody said it will "continue forever". He just said it is not dead yet, and it isn't. Single thread CPU speed has stalled, but transistor density is still increasing.
Why do you have so much invested in insisting that "Moore's Law is dead" anyway? You go to every discussion about this topic and post the same nonsense over and over. You do the same thing in every discussion about AI or machine learning, insisting that AI isn't "real" because it doesn't match what you see in the movies. Maybe you should see a psychiatrist and find out what is driving your weird obsessions.
Funny how you claim to know more than the man who came up with the law. He predicts it will end around 2025. Time will tell https://en.wikipedia.org/wiki/...
Only the State obtains its revenue by coercion. - Murray Rothbard
Intel moved too soon, it was too expensive, there was no transition plan, and they didn't get enough industry partners to buy in. It wasn't the first time nor will it be the last when a major industry player thought they could strong-arm everyone into a new platform only to be shown that they can't.
The problem was that the architecture solved the wrong problem. it should have reduced the ratio computetion power vs transistor. But transistors are (and were) cheap, so it didn't really matter. At the expenses of scalabilty and optimization at runtime, because they tought that the compiler can do the job offline.
But the runtime optimization in the cpu have more information about how the code is executed. For exemple branch prediction is easier when you can see how the conditional branch was executed in the past.
The performance of itanium was never been stellar...
The CEO's love life and work life are too closely integrated, I heard.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
from 1990 Harvard Business Review... https://hbr.org/1990/07/reengi... Equally don't follow what has been successful if you want to be disrupted...
nothing to see here - move along
Now that I've had some time to churn through the memory banks, I recall at the time we were bumping into the 4GB memory limit regularly. We customers said to Intel that we needed 64bit memory addressing in x86. Intel told us that we couldn't have it; I forget all the excuses for why not. Instead they tried to sell us a new platform without software compatibility. Then AMD said we could have it, with backward compatibility, and gave us AMD-64. Intel had to eat crow when they followed up about 2 years later with their own 64 bit CPUs.
Intel dumped ARM (Xscale) over 10 years ago (2006), and it's not clear even with hindsight that it would have been a successful strategy for Intel to use that ARM license. It seems doubtful that an Apple-Intel alliance around Xscale would have been possible given that iPhone's development (2006-2007) likely began when Intel still had Xscale. I can assume it was explored by Apple or Intel, even if only on a whiteboard, but history shows us that Xscale wasn't used by Apple. (probably price, performance, and lack of cell modem integration)
Furthermore I predict ARM's dominance to be on a decline as the consumer industry diversify into more CPU architectures like RISC-V. But sadly I don't think this will translate into more x86 sales for Intel. Ultimately the end user want a very full featured web browser on an inexpensive device (mobile, laptop, or tablet) with a long battery life, and there are lots of ways to reach that kind of high level goal. Full time connectivity to "the cloud" is going to be the marketing mantra for devices for the next decade I believe, and not so much about the gory details of the instruction set architecture. (sometimes I wonder if people aren't stuck in the mid-1980's to mid-2000's PC industry way of thinking about MHz and CPU revision)
“Common sense is not so common.” — Voltaire
Intel is itself a victim of Intel's success. Just like PCs find it difficult to shed the x86 model because of compatibility requirements, even intel can't compete with the x86, newer chips that aren't a part of the x86 line have been flops, they couldn't even get their first 64-bit alternative failed to gain popularity because it wasn't compatible enough with the 32-bit x86, so they got beaten at it by AMD.
And when a product is not a big initial success, Intel loses interest in it and considers it a failure. They could have kept going with their early ARM chips but they dropped those before ARM suddenly became popular. The reason they have all those older CPU products that never got very popular is because Intel has actually tried to break out of its PC stereotype.
On the other hand, this has given enormous room for smaller companies to get more play time and differentiate themselves in the market, which is a good thing overall.
Intel's trying to shoehorn x86 everywhere reminds me of that scene in The Brady Bunch Movie where Mike Brady (played by Gary Cole) keeps designing every project as a clone of his house.
'He who has to break a thing to find out what it is, has left the path of wisdom.' -- Gandalf to Saruman
Historically, Intel was huge for Silicon Valley in the early days. It had the first commercial CPU on a semiconductor chip. Intel was prominent amongst the early semiconductor companies that gave Silicon Valley it's name, founded by two of the original Traitorous 8. Intel kept a big hold when it became the chip used in the PC, moving microcomputers out of the hobbyist realm, thus keeping itself highly influential and relevant even when Silicon Valley started being more about software than silicon.
Yep. Kinda like the ill-fated iAPX 432 (Intel's first stab at a 32-bit x86 CPU). Certain supporting technologies had to mature in tandem, and they finally did four years later when the 386 came out.
If they were smart, Intel would buy up all the Mill Computing IP and base a new architecture off of that. They should think about it whilst they're still sitting on a decent pile of cash.
'He who has to break a thing to find out what it is, has left the path of wisdom.' -- Gandalf to Saruman
full of empty air and wasted space
More space = more powerful convective cooling.
'He who has to break a thing to find out what it is, has left the path of wisdom.' -- Gandalf to Saruman
And there's nothing wrong in trying to be a good corporate citizen - or even to appear to be one.
It's the 'appear to be one' that feels like the big problem in the industry. Trying to 'look good' ends up with less qualified employees *and* being patronizing toward classes of people all at the same time.
Someone who is of a particular minority that is very skilled and has really worked hard finds the position filled by the first minority hire that came along before him, because the organization is biased, but knew it hired to hire 'a' minority and wasn't too picky. This is frankly insulting to that minority group because it presumes the only way to hire is to lower standards.
Similarly with how these companies portray who is responsible for big news items. I was neck deep in a particular project that was in the news and saw a news article proclaiming that 'here are the people who made this possible'. I click to see some people I never had heard of. The company had decided to declare some seemingly random people who had nothing to do with the project at all as responsible. Why? Because it was a better diversity story. The people who were responsible were all white men. This is certainly a problem, but having token minorities to change the optics does little to fix the structural problem, it facilitates denial, and does so in an incredibly expensive and inefficient way.
Of course, this may all be the appropriate strategy in the fullness of time, as otherwise it's a chicken and egg (disenfranchised minorities think it's hopeless due to lack of examples and never try, and it's impossible to get the examples while people are thus discouraged), but it is far from a satisfying situation to watch it first hand, and unclear whether the same biases that prevented any minority hires are being reshaped to treat minority hires as more than a token effort.
XML is like violence. If it doesn't solve the problem, use more.
Intel could have been first in mobile space, except for Microsoft repeatidly leaning on them to stick to making chips for the IBM PC, that would be IBM, 'the PC company' as they referred to them in internal emails:
.. We have entered another round of "partnership" talks with the PC company and mentioned this as an issue, but they claim thay can't fix this for us."
.. yup, it would be crazy to Intel define this the only urgent issue I can think of is defining how it boots, if we let Intel do this in a proprietary way we're screwed."
.. Overall we wil never have the same relationship with IBM that we have with Compaq, Dell and even HP because of their software ambitions .. On their side I mean JAVA and NC."
.. did 2 things that amaze me: They kept the NC specification around despite saying they would not .. They snuck in a server specification. There is some failure in communication"
.. If we don’t get Intel off of Linux internally (the failed EDA project) – we will never get the *cultural* alignment that we want'
March 1994: "IBM has a LOTUS NOTES
Dec 1996: "we have a conference call with them (intel) re NetPC today at 9
Oct 1997: "I have a critical meeting with Intel a week from Wednesday. I want to convince them that they need to stay away from Oracle NCs and work more closely with Microsoft."
Nov 1997: "IBM refused to big anything related to Backoffice. I said they to use their PCs to distribute things against us. I said they are dabbing in NCs in a way we don't like
Nov 1997: "Intel
Nov 2001: "I think we will have to live without a Chinese wall clause for the front end of the compiler
You tall a lot of shit but speak in general terms that are often wrong. For example MANY things go into faster transistors and hence more powerful processors. Feature size is just one of those things.
Moore's law was an overhyped piece of shit from the beginning that was latched onto by the media and people who don't understand microprocessor and semiconductor design and fabrication at the practical level (let along theoretical).
Your statement about transistor counts leading to more powerful processors is so naive it is nearly laughable.
Intel is in deep shit because as the article states they have insisted on being their own fab plant and shoving x86 EVERYWHERE. They haven't innovated a damned thing design wise since Core and that was largely done in Israel. Yes their 10nm being late is an issue but only because they have relied SOLELY on their fabs to keep them competitive for decades. Remember Core is a rehashed P3. That should give you some idea how old it is. The next thing they tried was Netburst and Itanic.. both were horrible ideas. Intel hasn't built a good uArch from scratch in 25 years.
It's that they don't want to do anything else.
They hold all of the keys to x86 (you need a license from them), why would they give that up?
I highly recommend people go read about X86 on Wikipedia, it tells you all you need to know about why Intel is not going to give up on x86. And before anyone says the patents expired, that is true for the original instruction set, however there has been quite a few improvements since the patent expired. SSE, MMX, PAE, virtualization and a whole host of others have come along since then which rely on the older license (including 64bit) so unless you are part of the club, you cannot build a modern x86 processor and since Intel refuses to issue new licenses, the only way to get one is to buy or license it from someone who already has one.
Eagle was earlier than Compaq - until the founder drove his brand-new Ferrari over a cliff.
Intel cannot postpone a crash with a truth that has been in the air for the last 20 years: x86 architecture is a beast of the past.
At once accumulated expertise on it made it win over new designs, but it is not the case anymore.
If you dig at it carefully, you will discover that Intel's "acquisition" of DEC's semiconductor properties was a case of DEC having "Intel over a Barrel" re blatant patent infringement coupled with DEC desiring to exit the chip manufacturing business. At the time, the StrongARM CPU was more capable, and significantly faster than the top-of-the-line Intel CPU. The catch was that Intel had to negotiate rights with ARM re: the StrongARM/XScale processors in order to make them (and that took 'Moore' than a year).
You mean like Microsoft? You have a point.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
I agree. Competition is good. And the only things I have tied to a specific CPU architecture is MS Office (for working on customer documents) and games. Hopefully that will change as well.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
Moore's law is about transistor counts. Adding cores adds transistors.
Moore's law is about cost per transistor whether it is achieved through density or area.
https://www.youtube.com/watch?...
The problem is those cores aren't very useful outside of the datacenter. Home/Gaming computing is still mostly about 1-2 threads and single core IPC has only gone up meager amounts since 2012.
"They can add cores, but it will not matter much. Most workloads are not core-limoted these days."
What? The average user does precisely one thing that's not either trivial or highly parallelizable, and that is gaming. Literally everything else done more than rarely either benefits from multiple cores, or is I/O bound, and no amount of CPU performance will accelerate it.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
EISA then
Wrong again. EISA was a flop. It was actually MCA vs. VLB. And yes, VLB sucked rocks. That's how much people didn't want to use MCA or EISA — both of which had the same pathetic problem. To wit: configuration floppies. While Macs had NuBus and Amigas had Zorro, both with drivers in ROM and complete hardware autoconfiguration*, PCs were still dicking around with floppies for configuration and drivers. You had to put one floppy in when you installed the card, and another one in to load the driver, and keep track of them both...
* The Amiga Zorro bus was superior to NuBus not only because it used a cheap cardedge instead of an expensive and delicate multi-row connector, but also because AmigaDOS was a multitasking microkernel-based system and drivers were simply processes, which could be loaded from ROM as easily as from disk. Most NuBus cards still had drivers, except for most graphics cards.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
I do not care to educate you here, read up on the subject yourself.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
I do not care to educate you here, read up on the subject yourself.
Who's educating who? Games are single-thread-bound, which is why Intel has higher minimum FPS (which is probably the most important criteria; when things are exploding all around you, you probably don't want your system to choke or stutter.) But users can wait for almost all other heavy lifting tasks (compilation, rendering, compression, etc.) except for rare birds like realtime video tricks, which are generally parallelizable.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
"They'll listen to Reason."
Just finished Snow Crash for the first time a couple weeks ago. I can't believe I wanted so long to read it, what a great book.
Wow, what a crazy story. I'd never heard of Eagle before, thanks for sharing, I just pulled up wth wikipedia page on them.
Read the original paper yourself: https://drive.google.com/file/...
But there are alternative definitions of "Moore's law" that are more or less accepted. One is the increase in performance per time period (which wasn't really "created" by Moore but is strongly linked with the law).
Moore's law isn't really about maximum possible transistor density but of maximum _economical_ number of _components_ per device. As process yields have improved and manufacturing streamlined to reduce overall costs die sizes can be increased. In 1965 silicon wafers weren't as good (fault density) as current designs and they were ~2" diameter compared to the 30cm (11.8") used now. Dies can be put closer together as the dicing technology have improved etc.
What hit a wall wasn't some time ago wasn't really Moore's law but other aspects not covered by the law itself (but mentioned in the paper above): heat and the impact of heat itself. Dennard scaling (decreased transistor size improved efficiency) slowed considerably, more computer devices are mobile so heroic cooling of high power devices aren't acceptable, the heroic cooling efforts would in some cases not be enough to keep the chip functioning as heat would build up before the cooler being able to transport it away.
Uber relies on Google Maps, Apple relies on Intel and ARM, Google relies on Sprint/Nextel/AT&T/Comcast/etc...
The part you missed in the above formula is actually the largest factor- luck. Get over yourself, asshole.
So luck was the hardworking part? Really now. Was it luck that got you your skills? Luck that made you decide that you didn't like living dirt poor? Nope. It was the desire not to be in the same place your parents were because you saw how much suffering it caused them.
Luck is about the furthest thing from reality in the world in being an actual factor. Chance on the other hand, now that can make a play in what happens. The arrogance in believing that luck though screams that you don't think things through. Though I fundamentally pity you, because I'm an asshole. And that's all your worth.
Om, nomnomnom...
Yeah so here's the difference between the person you were replying to, myself, and you. In each case when we got knocked down by chance, or because we made a bad choice. We dusted ourselves off, tried to, or figured out what went wrong, and made another choice and learned from it. You on the other hand, got knocked down and started whining "life isn't fair." Something the both of us already learned, and instead of whining, we made the choice to do something else.
Luck is absolutely garbage mumbo-jumbo. There is chance, and chance can be modified by your own actions. You're the perfect example of someone that believes "luck is the deciding factor in life" instead of: Changing one's circumstances, being in the right place or wrong place at the right time, and to think further a head then the 10 minutes to get laid and wondering why you're paying child support for 20 years.
Om, nomnomnom...
When you said you were a "half-asian male" I didn't realise you meant that in total you amounted to about 50% of an asian male. I'm sorry that all those women made the height requirement and that you didn't.
I'm sorry that you failed the basics of grade 9 science, it would likely explain your lot in life. And having to fight against illegals pushing the cost of your janitorial services down.
Om, nomnomnom...
So luck was the hardworking part?
Not sure from where you drew that twisted ass logic to come to that conclusion. I'm pretty sure I said luck was the luck part.
Seriously, how the hell did you make it past grade school?
Was it luck that got you your skills?
Nope. It was luck that allowed me to use them.
Luck that made you decide that you didn't like living dirt poor?
Nope. It was luck that allowed me to use my skills, and pull myself out of poverty.
It was the desire not to be in the same place your parents were because you saw how much suffering it caused them.
Pretty sure that isn't a desire unique to the successful. I think the fact that you say that says a lot more about you than you realize.
I think it's the tortured logic your brain has developed to cope with the cognitive dissonance you feel inside over the knowledge that those of us in the white collar 10%+ percentiles aren't as special as luck has allowed us to position ourselves.
Luck is about the furthest thing from reality in the world in being an actual factor.
An interesting claim... One you haven't actually made any argument to support.
I have, however, already give you a counter-argument to that assertion- so keep trying. You can have all the skill in the world, but it's luck that gets you to where you can use it. Or do you really think all those kids in the third world are just lazy brown colored animals?
The arrogance in believing that luck though screams that you don't think things through.
Another claim you haven't actually backed up with any logic to speak of.
Though I fundamentally pity you, because I'm an asshole.
I find that assholes often pity their betters. I think it's another coping mechanism.
You on the other hand, got knocked down and started whining "life isn't fair."
That's a rather boneheaded interpretation of what I wrote, don't you think?
I'd say it's more accurate that to say that I dusted myself off, made it to the top 6% of income earners and then reflected upon the journey and said, "wow. life really isn't fair."
Luck is absolutely garbage mumbo-jumbo.
Followed with...
There is chance, and chance can be modified by your own actions.
I think it's clear you're not entirely familiar with the words chance and luck.
Also, anything can be modified by your actions, but the hubris displayed here regarding your apparent belief that you can overcome all odds, and that anyone who cannot obviously just didn't have what it takes is mind-blowing. You may be a skilled person, but you're certainly not intelligent.