Having a monopoly on x86 chips and charging eyewatering prices really does pay off – Intel CEO Intel today revealed that its first-quarter earnings and sales were more or less as forecast, sending its stock down in after-hours trading. Analysts had anticipated a Q1 2017 earnings-per-share figure of $0.65, non-GAAP, in line with Chipzilla's guidance. The Santa Clara, California-based processor giant reported EPS of $0.66 …
Any mobile offering, and the statement that when Quantum Computing arrives Intel will be there to shoehorn old timey 1980s x86 architecture into the future "q86!"
Meanwhile, ARM is coming to a data center near you, and I don't know about you, but I am only buying new ARM systems and not looking at replacing my two main Intel boxen, yet. But those will go ARM as well, I suspect. Never mind AMD, here come the ARMs! My data centers at workplaces are all x86, but then they used to scoff at Linux, and that's all I do in the data center nowadays. Things are going to get interesting.
x86, how long can this go on? I guess don't fix what's not broken, but surely we are going to advance to QC and leave x86 as an emulation on better hardware.
This post has been deleted by its author
ARM remains as stillborn in datacenters as intel is on mobile.
For my newest home server I got a xeon e3-1240L v5. Quad core 8 thread 2ghz 25W TDP. With 16g of ECC ram(soon 32gb) 6 HD 2 SSD and low power (19W) nvidia video card, and 4 80mm fans, the whole system(70W) uses less power than the triple core amd cpu alone that my home server was using before (90W).
I just love this cpu. But holy shit is it hard to find. Took 2 months to get one. Seems very strange for a current gen xeon to not be sold by anybody. After waiting a month from one vendor to buy a 1235L v5 for about 300, i found the 1240L v5 from dell for 500. Still took dell a good 3 weeks to deliver it. I left the other order for the 1235 on for an extra 2 weeks the vendor kept saying the distributor will have it in stock next week.. thinking I would use it to build another box. But I canceled it in the end.
It rips in encoding blu ray too, easily twice as fast as my dual socket 6 core (12 total cores) opteron workstation. Mainly because the video encoder doesn't scale to 12 cpus on blu ray for whatever reason.
Main thing is i wanted a chip/system that could easily survive the 100 degree summmer that is coming.. my amd system lasted fine last year but the chassis is just too tight. I don't want to risk another hot year on the old system(4 or 5 yrs old). New system has great filtered airflow front to back. And the cpu cooler is built for a 90w("normal" chip.
Server side i am interested to see how AMD's new 64? core chip comes out. My work servers run 22 core xeons today and I'd love to have more cores (still have about a dozen dual socket 16 core(ea) opterons in use) vsphere loves cpu cores.
You are a strange customer. You have the patience to trawl through Intel documentation for the CPU that meets your personal requirements, and the determination to track one down and buy it. Intel's marketing team hate you. You are supposed to look at the reviews of Intel's finest and assume when you buy the chip that is actually in stock that you are getting the same thing.
Perhaps if you order 1000 CPUs per month Intel will treat you a little better. Can someone who buys on that scale please tell us: does Intel make high performance low power chips in quantity, or do they hunt for a few exceptional parts to send out to reviewers?
"To be fair, if you're counting your boxes in ones and twos you're not exactly Intel's target market."
To be fair, if anon's counting his own boxes in ones and twos, but his name is Mark Zuckerberg, then he probably "is" Intel's target market. That's the way life works, our own personal experiences reflect in corporate decisions, good/bad as that is.
(Another Anon - not the same Anon as above)
I guess someone forgot the memo that ARM don't make any processors and that Intel are an ARM licensee.
If somehow the entire enterprise market did move to ARM design, Intel will make them.
To put things in perspective, we run a relatively small data centre (few thousand racks) and we spent more yesterday on processors than you personally will in an entire lifetime.
I didn't see a reference in the quoted portion of his comments to demand for the x86 architecture, or anything else that might be construed as a reference to the advantages of not having competition - until recently, with Ryzen - for that ISA, such as a comment on synergy with the Windows operating system.
Brian said strong prices boosted sales - ie: they charged high prices and yet people still bought their stuff - netting them a lot of cash.
Which is easy when you've got a monopoly over data center and desktop compute.
It's a flippant headline for effect: it's all explained in the story.
C.
IMO the higher ASP is because people are buying fewer low-end processors. This could be because the useful life of a machine is much longer, so the cost of the processor is amortised over a longer period.
I've also observed that the marketing of i3, i5, i7 variants of the same silicon has been very effective - family and friends tend to opt for an i7 because it is "the best" and the i7 system is "not much more expensive". They don't have the technical insight to evaluate the cheaper options, even though IMO what they really need is "anything that isn't an Atom".
INTC issued a forecast. The market expected them to beat the forecast, and that expectation was reflected in the stock price. INTC has been rising for the past couple of weeks for sure. When the actual results were announced and the market expectations were not realized the stock price dropped.
Some people made a bet too many and lost - bought a bit too high, no one is willing to buy INTC at those prices once new information became available. Not absurd at all.
[Disclaimer: the following is from memory, from the times when I had a markets-related job. I have not tried to find historical data on the 'net today.]
As a more extreme example, back in the late 90ies INTC announced a rather significant increase in sales, etc. However, the market had expected even better performance and INTC prices had been pumped up before the announcement. As a result, something like $91B was wiped off INTC price in a day. To put the matter in perspective, at the time that number was higher than all the market capitalization of all listed American companies except 29 biggest ones[*].
[*] Anyone who finds this absurd should look at the current list of companies by market cap. The first 5 are Apple, Alphabet, Microsoft, Amazon, Facebook, with ~$8.2T between them. Only one of those existed in any significant bulk back in the 90ies. Times change - sigh...
Intel's had plenty of competitors over the years.
Remember Cyrix? Via? Now AMD are still around but they don't own anything near a decent portion of the market. Not because they couldn't, but because they haven't. AMD, to me, has always seemed one generation down. Even my completely non-techy bosses specify "proper Intel" (meaning not mobile or low-end i3 chips, and not AMD) because of their years of dealing with things even if they are removed from the end hardware. I can't say that I'll be the IT manager to disagree and change their spec (even though I have the power to do so).
In fact, Intel's biggest threat now is ARM. Their only ally against ARM is really Microsoft. While Windows doesn't "work the same" on ARM, Intel can maintain their position. But if Windows falls out of favour, or if the PC truly does start to die or being just a web device, Intel could be in real trouble very quickly.
I bought an RPi 3 the other day. Have you seen the speed of that thing, for a tiny 5v, 2A = 10W device? Your phone is ARM, even if it's Apple. Your tablet is almost certainly ARM if you paid less than a grand for it.
It's not a huge leap to imagine that in a decade or so, we'll be using Office 365 (harder to kill off) on mobile devices and non-Windows machines (Chromebooks, etc.) and the x86 will be the reserve of, say, gamers.
Why would ARM be a direct threat to Intel?
ARM don't produce chips themselves, and Intel are licensed to manufacture ARM based chips already.
Even if ARM do take off in the server and/or desktop space, it could well still have an Intel badge on it, and you can bet the budget holders when given the choice of purchasing and 'Intel ARM' chip, or some other brand, would likely stick with Intel simply as they already have an existing relationship with them for their existing x86 estate.
The one thing it could bring is some better competition, but even then, Intel probably have the most advance FABs around. So could likely produce some decent chips, if they put the same effort into ARM as they do currently with x86.
Given that the immense cash flow generated by their x86 monopoly has allowed Intel to build the biggest and best fabs around - so that they can make newer, shinier chips to get people to upgrade more often - if ARM suddenly took over, Intel would still be positioned to make the best ARM chips that money could buy.
So their dominance in processors could well outlive the dominance of the x86 ISA, even if the engine that created that dominance would be gone. And, of course, the x86 architecture is protected by patents, not by copyrights, so, given that AVX-1024 or 128-bit addressing may not really be needed for general-purpose computing, even if the Wintel duopoly remains in effect, someday it will be possible for anyone to implement the x86 architecture commercially.
Protected by their effective monopoly, Intel has developed the most expensive manufacturing process in the world. That is fine for their CPUs, but a barrier to entry into any competitive market. If it ever got out that a Pi3 is fast enough for a call centre terminal, Intel would find their monopoly being restricted to data centres. The big data centre operators have been looking at licensing ARM and getting their own design manufactured. I think an ARM data centre would be big enough news that we'd know. I think Intel giving big discounts to Googazonple Book because they have a viable data centre ARM would be so secret that we would not find out for years.
[ARM are already in data centres - spinning disks and SSDs are often ARM or a heavily mutated descendent of the 8051. Switch builders who selected ARM of MIPS are busily celebrating that they didn't pick Intel.]
I don't believe Intel has a process advantage any longer. Intel's upcoming 10nm process looks to be at best on par with TSMC's 7nm process that will be ramping a few months after Intel. TSMC will introduce their 5nm process in late 2019 / early 2020 while Intel is still on 10nm, and have a clear lead at that point.
I also don't believe that Intel could "quite easily make the best ARM chips money can buy". They were unable to beat ARM SoCs in the mobile market with x86. If they can't do better using the ISA they know best why would they do better using an ISA they have no experience with?
If you are suggesting they could make the best server-class ARM chips money can buy, I'm sure that's true, but that's because no one has been willing to spend the considerable resources required to design max performance ARM CPUs - because there's no market for them. It isn't clear that Intel could do any better job of designing a server class ARM CPU than Apple could, for instance, but Apple has no more incentive to spend money on such an effort than Intel does.
The day of the one atom wide transistor is coming.
Then everyone will be on a level playing field and it will be down to how many layers of active devices you can lay on top of each other, how thin those layers will be and how well you can extract the heat from them.
We won't live to see that. A single atom transistor has an effective gate width of anywhere from 0.1 nm to 0.5 nm (depending on the atom) Even if Moore's Law kept plugging away with a new sqrt(2) scaling every two years, that's at least 15 years away. And as problems continue to mount and expenses continue to rise, we won't be doing a new generation every couple years.
Though maybe we'll get there sooner from bottom up manufacturing since it will become unaffordable to keep making them top down long before we reach single atoms. That is, we'll be able to make them in the lab, but not economically manufacture them on a large scale.
Come to think of it, I thought I read somewhere that some university researchers had made a single atom transistor in the lab a few years ago.
I think most of us can expect to live another 15 years, barring the D pressing the wrong button at the wrong time.
"And as problems continue to mount and expenses continue to rise, we won't be doing a new generation every couple years."
TBH we can't do that now.
"Though maybe we'll get there sooner from bottom up manufacturing since it will become unaffordable to keep making them top down long before we reach single atoms. "
I'd tend to agree with you. but of course that will require Intel to totally re-jig its mfg plants and the PHB's really don't like that idea. They will keep squeezing the next few nm out of the process before they swallow the idea of additive mfg applied to semiconductors and TBH no system exists that can give both the feature size and the production speed of lithography.
If that were to change, or come other company came up with a way to do that.....
I remember hearing this exact same quote 15 years ago. Its like they can never stop being that company that makes bucketloads of money by making processors. I think the media ought to pick up on this and start reporting, "Intel fails to deliver the cloud for the 60th consecutive quarter and plans to make it a record 61"
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
