Too little, too late: Intel's legacy is eroding Intel CEO Pat Gelsinger's carefully assembled house of cards is collapsing around him. And it's not really that surprising when you look at the hand he's been dealt. For those that haven't been following Intel product roadmap that closely, here's a quick recap of where things stand today. The company is stuck on an aging 10nm …
Honestly, I would think that, if your company has passed the billion dollar mark at Wall Street, you shouldn't be taking handouts any more.
You're a big boy now, you can manage on your own.
The Government needs to keep its funds for, oh, say, education, for example. Or healthcare.
Things that don't make money, but are sorely needed.
Are you kidding... the sick bring in vastly more income to the system. A doctor who signs up a new Diabetic Patient can afford a much larger yacht and a built-in swimming pool. Stupid unhealthy people are an endless tap to the gravy tankard.
"I'm lame I can't clean my gutters...."
"I would gasp for air just to get to the sweets shop."
"Do you sell neuropathy sox?"
"My wheelchair battery is dead!"
The many hidden income streams of God bless the sick...
Ignoring the albatross of the healthcare system, how is Intel encouraging education? They do nothing but pick and choose the best they can find from wherever they find them, and do nothing to help train more engineers in the US.
In point of fact, they would be far happier to higher some people from elsewhere and put them on work-dependent visas so they can't complain. Which is what they do.
Agreed, but we are talking about for profit companies, There are many healthcare, educational and correctional institutions (prisons) in the US that are for profit companies. They really shouldn't be or at least be capped on how much they are allowed to take out of the system. I would rather see money go to assist those that make products vs. the last big bailout we had of companies like AIG. I would argue that the loss of Intel would be worse for the US than the loss of a company like AIG. What does "too big to fail" really mean for purely financial organizations when nearly identical products and services can be had from another that was "too smart to fail". From CBO.gov "The final net subsidy cost of TARP assistance to AIG was $15 billion." But hey, what's a few billion between friends?
By "at Wall Street", do you mean market capitalization, I.E. how much your entire company is worth when investors apply a multiple of what you actually have? If they value your company at a billion, you have much less than a billion in reserves and profits. Meanwhile, do you know how expensive semiconductor manufacturing plants are? Having a few million to throw at the problem won't get you there. Companies are certainly greedy and ask for and accept money they don't need or deserve, but your scale underestimates the costs of manufacturing at scale and overestimates what a billion in someone else's imagination can buy you.
I think you should state X amount of revenue or X amount of profit.
Stock prices are meaningless when a company with 25m revenue was valued at over 300B. No doubt it's crashing now, but it was worth more then goldman sachs at one point of time.
https://abcnews.go.com/Business/21000-surge-amtd-digital-latest-meme-stock-craze/story?id=87923597
And I think there are also other stocks which has shown such weird movements (and am not including gamestock from a year back).
You're kidding, right?
This is America, where profits get support to be privatized but losses become..."national security issues". LOL.
Multi-billion dollar industries that get 'subsidies'? Get in line behind Big Agra, Big Pharma, Big Oil, automotive, banking and just about any other big industry that either has the pockets to find and lobbyist-support their legal tax loopholes, or conversely simply keep their hands out while they cry "Woe is Me!" because their stock dividends aren't as high as Wall Street [hallucinates] them to be.
Welcome to the 1890's Gilded Age. We knew you liked it [back] here.
https://www.glassdoor.com/Reviews/Employee-Review-Intel-Corporation-RVW28224843.htm
"Prople getting hired and making decisions that are not qualified to make decisions"
"If you hire the right people to do the job you do not need diversity you will get a well rounded workforce"
A 300 million dollar budget for diversity and quotas implemented everywhere (I know this from a former employee). I think this started in 2015 or so, maybe this is one reason for the downfall of the company?
Conversely, there are plenty of other tech companies with robust diversity programs who are going gangbusters (Salesforce and Google leap immediately to mind). Maybe the cranky old white dudes should quit their bitching about having to work next to people not entirely like themselves and get back to work.
Or maybe something entirely else is at fault.
It is strange that people who say "Maybe the best person should get the role, and not a funded quota program" are deliberately misinterpreted as being cranky old white men. If you have to distort what someone's saying by that much to fit your emotional understanding of the situation, maybe just don't.
@Robert Grant: the point of diversity programs is usually to counter implicit bias, i.e. the problem of hiring exactly the same sort of person each time with the assumption that only, e.g. white males who came from a certain kind of engineering school are capable of doing the work in question. Diversity programs are ideally meant to open up the candidate pool to qualified candidates who wouldn't otherwise be hired because they don't tick the right boxes in regard to conforming with the hiring manager or interviewing staff.
As with all things, these programs can of course go awry, so I would not want to assert that they're perfect.
The error you're making here is that assuming that people will hire the best candidate for the job based entirely on impartial judgement of the person's skills and abilities. Study after study has shown that this is simply not what people do- the same CV with an Indian-sounding or European-sounding name often get quite different responses. Interviewers will ignore a more skilled candidate who doesn't look or talk like them - there's no shortage of data on this.
The reason companies are willing to invest in diversity programmes is because they tend to balance out some of that bias and mean they have a better chance of *actually* finding the best candidates. Any wider benefits to society are entirely co-incidental; a more diverse workforce is good news for shareholder value.
I have had to deal wit this problem in the past.
At $COMPANY in the deep south in the 90s, the director of operations (and a lot of the local males) considered females to be inferior; I am sure you know the type. That is not my view as I have met many women in engineering and other professions who were clearly superior to many, if not all, of their male colleagues.
There was an opening for a supervisory position in the repair department which was a bit top heavy with females (low pay and all that) so I talked to the CEO (my direct boss at the time) and he agreed I could completely anonymise any applications.
Given that I was guiding the technical operation, I was rather well placed to know the strengths (and the areas in which they struggled) of each potential candidate.
The applications were all handed to me directly and I then made new applications with the titles Candidate A, Candidate B and so forth. The operations director was livid because he could not see the names but we had done an end run around that.
Each candidate interviewed with an outside expert in management (some management skills were necessary) and that was written up with scoring.
I did the technical scoring. It had been agreed that the candidate with the best overall score (somewhat weighted as both technical and management skills were required - the solution was to multiply the scores). There were some other things such as attitudes to others and so forth - the last thing we wanted was a psychopath [1].
Now the entire process was explained to the candidates individually who all agreed it was fair and above board.
This objective method yielded a very clear winner who happened to be female. I am still amazed we managed to keep it all secret.
The operations director grumbled but there really wasn't anything he could do about it.
One of the male crew (who really wasn't that good) said he wouldn't work for a woman so he was told to not let the door hit his ass on the way out.
I am certain that had that process not been followed one of the male candidates would have been chosen.
The person chosen was the best choice from all the candidates, which is as it should be.
[1]. Certain psychopathic traits are actually a good thing. The good psychopath's guide to success is both fascinating and a pretty good read although a lot of it is in the vernacular.
There are women who are very qualified to do various engineering jobs.. I have a friend who left the job she had at my company to become, on the face of it, a tech support bod in one of the skyscrapers around Canary Wharf. She never gave real details of her job, but she was important enough for her company that when their Edinburgh office had a problem with their computer system, she was flown on one of the company's private jets to Edinburgh, driven to the office, had to fix the problem, and was flown back on the same private jet the same day.
Haven't spoken to her in years, but last I heard she had a very senior role in one of the online banking development teams for one of the big 4 banks.
Gelsinger was raised on family farms by his parents, June and Paul Gelsinger, in rural Robesonia, in an Amish and Mennonite part of Pennsylvania.[6][7] As a teenager, he received a high score on a Lincoln Tech electronics technology test, winning an early-admission scholarship. He then skipped his final year at Conrad Weiser High School and left home at 16 for college. There, he earned the remainder of high school credits for graduation and worked at WFMZ-TV Channel 69 as a technician,[8] while obtaining an associate’s degree from Lincoln Tech in West Orange, New Jersey,[9] in 1979. He moved to Silicon Valley to work at Intel as a quality-control technician, at age 18 in 1979.[7][10][6] While at Intel, he earned a bachelor's degree in electrical engineering, graduating magna cum laude[11] from Santa Clara University in 1983, then a master's degree in electrical engineering and computer science from Stanford University in 1985. In 2021 Gelsinger was inducted into Indiana Wesleyan University’s Society of World Changers. While speaking on campus he received an honorary doctor of science degree and a bronze bust of Gelsinger was placed in the university’s library rotunda. In 2022 he was awarded an honorary Doctorate of Engineering from The Ohio State University. An Intel 486DX2 microprocessor using the 80486 architecture designed by Gelsinger in the 1980s. He spent much of his career with the company in Oregon,[13] where he maintains a home.[14] In 1987, he co-authored his first book about programming the 80386 microprocessor.[15][1] Gelsinger was the lead architect of the 4th generation 80486 processor[1] introduced in 1989.[9] At age 32, he was named the youngest vice president in Intel's history.[7] Mentored by Intel CEO Andrew Grove, Gelsinger became the company's CTO in 2001, leading key technology developments, including Wi-Fi, USB, Intel Core and Intel Xeon processors, and 14 chip projects.[2][16] He launched the Intel Developer Forum conference as a counterpart to Microsoft's WinHEC.
Even if he has picked up an MBA somewhere (?), Gelsinger clearly has an accomplished engineering
background.
Pat Gelsinger has a AA from Lincoln Tech, a BSEE from Santa Clara University (go Broncos!) and a MSEE from Stanford, so decidedly not Ivy League. Lisa Su has done amazing work at AMD, but she's lived in the US since she was three and her degrees are all from MIT.
What does 'woke' mean in this context?
Those on the right in American life use it to mean everything bad in American life today and I'm never actually sure what it means.
Is the poster saying that Intel is too diverse? I would that very hard to believe.
How lately anytime any business in the US hits even the slightest speedbump, the Trumpsters will say it is because they are "too woke". If Intel's profits are down "too woke". If a Marvel movie fails to earn $1 billion "too woke". Lather, rinse and repeat.
Fox News spoon feeds them the latest white man outrage and they parrot it all over the internet like a bird vomiting lunch for its hatchlings. A couple years ago everything was BLM or Antifa's fault, but they need to keep the outrage fresh so they can have a ready supply of IQ 45 stooges willing to commit election fraud in Trump's name by being a fake elector, or attempt to storm the US capital and hang the Vice President.
I wonder what the summer 2024 outrage will be, any bets?
Intel is not hitting a speedbump, it has already hit the brick wall. Its latest venture to become an extension of the US Gov welfare office isn't going to save it. FAB is just a five-year road into the abyss. Intel already failed at 10nm FAB yet they are the Politician's favourite to lead the way...
Who is going to bring their fabless designs to Intel? Qualcomm? And who is going to buy these FAB chips that Intel makes? The US gov will only sanction them so the largest buyers, ASIA won't touch these chips and they will rot like bananas in an unrefrigerated train car in Needles California.
The way you win in the marketplace is to make better products, not to strap on the Chuck Schumer Tom Cotton GOV feedbag. One of the many Gov subsidized intel FAB plants will be built in Ohio.... Ever been to OHIO? It's Liverpool in a rust storm. It's a long long way from Santa Clara.
Capitalism is a great system, but this blind faith you have in the private initiative is beyond absurd. The reason why Taiwan has today the most advanced microchip FABs in the world is because the Taiwanese Government decided many years ago that it would be a national priority. That included lots of government funding in many different areas.
The reason why China will most probably lead (or be one of the leaders in) microchip design and manufacturing in the next 5-10 years is because of its Government focusing on it as a national initiative. That includes funding many different areas, from industrial espionage to education, and obviously subsidies to any company that tries to compete in that area.
To become a technology leader you need a lot of money. The US has great universities (that receive lots of Gov. funding, btw) and that helps us with the knowledge leadership, but the manufacturing process is also extremely complex and you cannot learn it on a whiteboard. You learn to make things by trying to make them. And that is very expensive and risky for the private initiative.
I do not know if Intel will succeed, I do not know if the funding from the CHIPS program will help the US regain the manufacturing knowledge and experience to be a leader in the area, but I am pretty sure that w/o government funding it will not happen.
We definitely need more diversity.
But diversity of *thought* rather than random crap like gender and skin color.
Lets be honest, if a pregnant, dark skinned, transgender person is making the same old shite business and technical decisions as some old sleazy white guys who should have retired decades ago, this is basically just as useless.
And annoyingly, those who are of the same mindset are more likely to be hired because they "fit" within the company profile. Looking different is fairly irrelevant.
Let me guess, you're white, male and middle aged? Anyone who looks different knows it makes a massive difference to most interactions. I just need to wear different clothes to get treated differently, I can't really imagine what it's like if your skin colour is the thing causing people to be dickish.
But you keep feeding yourself this line bud, that's why you got the job, because you were the best candidate!
Intel has been dropping the ball since the Itanium days and perhaps even before, so you can place your diversity quota ideas back where they came from.
Your former employee friend is probably a former employee for good reason who can't accept it and blames others for their failure.
I think we should stop benchmarking Intel (and other products) and instead giving them participation trophies.
What if someone says that Intel CPU performs worse than AMD? They should be more inclusive and that not every CPU is built to perform and some CPUs are born disabled.
What if someone says that Intel CPU consumes too much energy? Well. First of all we should think of CPU positivity. Some CPUs are bigger and need more energy than others and we shouldn't engage in TDP shaming.
So what's the solution? Intel needs more subsidies to discover its true self. Maybe it is not a company that makes CPUs after all but it has been pressurised by the society to be that way and they are clearly unhappy! Maybe Intel always wanted to build wooden boats or craft güiros?
I still have the review of Intel 11thGen chips in the ear... "Waste of Sand." I see the difference every day. My company machine is Intel 11870 (or so), which can be quite fast for about 10 seconds. After that it is only loud. As in LOUD. Most of the other tech here is AMD Ryzen 3000 or 5000. It is usually faster or way, waaay faster, and this sustainably without making much noise. A bit frustrating for even slightly compute-heavy tasks.
"Intel has managed to squander its lead over rivals and its massive reserves of talent on over-ambitious projects that have time and again arrived late to the party"
How come the i7-7500u launched Q3'16 is 2 cores 4 threads, TDP 15W, when the i5-8250u launched Q3'17 is 4 cores 8 threads, TDP 15W?
How come the press has not loudly applauded such a leap forward in a such short period of time?
No they are not. The i5 is better than the i7. They are comparable in single thread ratings but "obviously" the i5 is 2 times better than the i7 in multi threads rating.
https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i7-7500U+%40+2.70GHz&id=2863
https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-8250U+%40+1.60GHz&id=3042
This is an obvious example to show that Intel has been twiddling its thumbs for years profiting from the cash cow.
Because they improve. Everyone improves to some degree, and Intel has produced better products recently than they had before. The question isn't "has Intel exceeded its processors from last year"? If the answer is no, they have really big problems. The questions are "Has Intel's improvement over last year's designs exceeded AMD's [or a similar competitor] improvement in the same time" and "Is Intel's product better or worse than a comparable AMD one".
The way you define "better" changes the answer to this question, and many have their opinions on the subject. You can consider performance, power usage, and cost to decide what variables you value and how you weight them. In addition, the example you show is from processors that are five years old. This was a nice improvement, but were the subsequent generations similarly improved? Was the power usage really the same between these (one thing that both Intel and AMD have been doing to annoy me is using power consumption numbers and then dynamically consuming so that the actual consumption is something very different than the reported number).
In many ways, Intel's CPUs are fine. They're still reasonably fast and work in a lot of things. I have no problem using one. I similarly have no problem using someone else's, and when it comes time to get a CPU for something, I'll let the performance, power usage, and price govern my decision. Intel needs to continue keeping those values up if they want to stay in the running.
Intel tried to leverage is market share to paper over their early troubles and push substandard offerings and marketing BS at inflated prices. It's early troubles then became a full blown dumpster fire as instead of hedging their process difficulties and doubling down on the added functionality in the extra tock cycles, they rested on their laurels waiting on the performance bump of a new process node that did not materialize.
In addition they gave up on many of their white box options, were plagues by microcode and side channel issues, and shipped motherboard, network, and gpu drivers with major issues.
That's a lot of "not fine". Someone with half their head missing is fine.
The "i7-7500u" is an i7 by marketing only for its actual i3 tech inside. A dual core with hyper-threading, giving it the i7 label so those typical sales droids and some manager levels can display the "hey, I got an i7 laptop, I am important" at meetings. It is like putting a Porsche logo on a WV beetle for these CPUs.
Then AMD came along, and i7 had to kick that fake-label trick one notch up to put is i5 tech behind that i7 label for the next CPU generation.
To me you sound exactly like those people falling for fake labels instead of checking the actual.
From Sandy bridge (2000) to Skylake (7000) intel knew they had no real competition - so took advantage of this. Each gen was only a small improvement and higher core counts were only available on their stupidly expensive HEDT and server chips. Ryzen was a kick to the proverbials - and when AMD also started to release good laptop chips intel responded late and in some cases with desperation (e.g. back porting the newer architecture onto the 14nm process).
This is the sort of article usually written about Premier League football managers who're struggling in an attempt to get them sacked. It's easy to be an armchair critic...
The thing is : the analysis in this piece is absolutely correct. So the question is, what do you do when nobody wants what you've got to offer anymore? And the answer is : not have the hubris to continue regardless.
To be fair, I think Gelsinger absolutely understands that, and has made some fairly radical decisions:
- The tilt to fabbing, and to fabbing for any ISA is a smart move, especially since not every chip has to be produced at the smallest node. He absolutely knows that the way geopolitics looks at this moment : the US government is never going to give up on an American Tech giant.
- Helping companies with getting their designs to fruition seems to be a good way to use the talent at Intel too. And
- Starting to design using RISC-V, especially, is nothing short of revolutionary, but offers Intel the potential to take the battle right back to Arm, and give them a foothold in emerging markets in particular.
Impossible to say definitively, but
a) as I understand it, Intel's design and the fabbing units are completely separate entities and
b) Intel would face punitive costs if they were found guilty of infringing anyone else's intellectual property, not to mention the damage it would do to their reputation : they simply couldn't afford to do it.
I'd guess that, given the current chip shortage, companies would be pleased that there was a way to get chips produced.
I believe there are 2 or 3 RISC-V design companies currently working with Intel to get their designs fabbed.
They are just behind TSMC/Sammy/Micron on the most advanced process nodes. The need to keep their fabs ramped up, and they should be able to push the older nodes at full tilt. Plenty of the semi's industry is produced on bigger processes.
The other factor to consider is that while the process pipeline is long, they may already be sitting on major breakthroughs they are in the process of rolling out, or they may spend the money to license process technologies they are currently missing.
Any way it hashes out, it would be months before any demo's see the light of day, making it likely and turnaround for Intel is two or more years out. The other players should keep their foot on the accelerator in the meantime.
INTEL FAILED AT 10nm. It will be five years before they get one chip off any FAB line if ever. And the competition, the real competition is way ahead in Korea, and Taiwan and Mainland China. China has 7nm Fab now. YOU ARE BEHIND in an area which is labour intensive and operates at very low margins. This is work that should have been outsourced. But somewhere in the insanity of Tiny Doll Hands and; Union Joe Biden, the US has come to believe that the division of labour MAKES YOU DEPENDENT ON OTHERS... That is called protectionism and IT NEVER WORKS.
Involved in this colossal scheme is the idea that the US should control all tech sales on earth including ALSM machines made in Europe,
The US and UK took their best shot at Huawei and SMIC and lost. So if the Tiny Doll Hands, the self-proclaimed stable genius flopped along with Union Joe Biden, how do you expect the grander plan in this Nancy Kerrigan Kneecapping strategy will work? It will FAIL.
The US has done its level best to suppress Japan, Korea and Germany with its standing armies and massive military presence. The criminal acts of US Military personnel continue. All of these countries have growing anti-American sentiments. With the great and absurd push by the US to force other nations to subrogage their technology and trade to support some notion of US dominance, is self-defeating.
This post has been deleted by its author
"intel engineers looking over their designs" - ain't gonna happen, Intel is too big and too compartmentalized for designs to leak from the fab to engineering. And even if Intel took the route of stealing other people's designs, that would mean they could never come up with anything new, and they would always be at least 2 years behind the competition. Because that's how at least long it takes to get a chip design to tapeout of A0.
Just like Sun Microsystems.
Hire anyone and everyone, regardless of qualifications; tolerate continual delivery slips and product quality issues, and trade on your reputation. Remember "No one ever got fired for buying IBM"?
People forget that no tree grows to the sky.
Chipzilla is the got fat, dumb, and happy many years ago
AKA 'The Motorola strategy'..
(I worked at MotRot ECIG many years ago for a couple of years - they took the idea of internal competition to an insane level - both within the MotRot group and within each company. We wanted to use the Motorola Computer version of the Mac - not only was their quoted price to us *more* expensive than an Apple, it was more expensive than they would sell to a non-Motorola company. So we stuck with Apple.. ECIG made cellular basestations - they were, in many ways, the first to mass market successfully and, for a time, all the sales guys could just phone up the mobile companies and ask them how many they wanted this month. Then the mobile companies started to not buy - Ericsson had started to sell compatible cellsites for less money and more features and, due to us having not bothered to innovate because 'we owned the market' we had nothing to compete with them. )
There's a reason why a search engine look for 'Motorola CIG' mostly finds links from before the millenium!
They've sort of been here before at least related to AMD(early Opteron era etc), don't forget the years spent pitching Itanium and that funky RDRAM(unrelated to Itanium). Then Intel turned things around(around Xeon 5500 I think?) and not long after AMD made some huge mistakes and they went downhill(not long after Opteron 6000).
Now Intel has done some huge mistakes and AMD has turned things around. Though I think even under the best conditions AMD lacks the fab capacity(yes I know they don't have their own fabs anymore) to consume a large part of the market, they just can't make enough chips to do it. Not that they can't make a bunch of money doing what they are doing already.
Intel has tons of money, and tons of resources and huge market and mind share. Seems a big mistake to dream they are out for good, just another cycle.
It takes a long time for the market to turn around, think of how long it took the clearly superior Opteron to make any kind of dent in Intel's data centre share. Pre-Gelsinger Intel had lost focus on its core business and even though current CPU Intel products are broadly competitive with AMD, I think AMD has lost its (undeserved) reputation as "cheap substitute" to Intel.
So going forward I think there's going to be a lot more balance in the x86 duopoly, maybe more like the Xilinx/Altera FPGA duopoly where the lead swaps from generation to generation and there is a roughly 40/60 market share split. Which is a very good thing for customers. And the smaller player has enough revenue to keep competitive, and the big one doesn't make so much money that it gets complacent.
"Too little, too late: Intel's legacy is eroding
Can Pat stop the rot – or is the chip giant facing a Kodak moment?
This is an easy one.
Intel was founded by engineers, its success was built by and depended on engineering (coupled with a very healthy dose of shrewd marketing)...and then the bean-counters took over.
Kodak moment.
Intel is a big ship. It takes time to turn. We aren't going to see the real effects of Gelsinger's leadership for a couple years, good or bad. If you don't work there anyway. The real key is when the team internalizes "what is it, exactly, that we do here?" And then the good stuff happens, and that pipeline takes years to show up on the retail shelf. He could be leading the ship onto the rocks. Or not. But we won't know for quite a while.
So. Looking to upgrade my kit after a long hiatus and not buying any of the stories the old school pundits are pushing about nanometers not being important. Tomorrow will come when tomorrow comes. Today - rather, this Fall - the choice is clear.
Lot's of companies have been killed by success:
(1) Intel had a huge and maybe continuing success with x86. Remember the really nice Motorola 68000?
(2) The UK motorcycle industry "retreated up market" for years....then got killed by the Kawasaki Z1 and the Honda 750.
(3) Once upon a time there was a European communications industry. Now we've got Huawei!
(4) Once upon a time Nokia in Finland ruled the cell phone roost. Now we've got Apple (but "Made in China") and Samsung
....and so on. Success is often the harbinger of a longer term failure!
The US created the semiconductor industry (cf. the invention of the transistor, Bell Labs; the invention of the IC, Texas Instruments).
The doors to the US Patent Office were thrown open to the world in the 1990s.
There is no more US semiconductor industry, nor the hundreds of thousands of jobs---directly and indirectly---it created. All gone, and we'll never get that back...
"The haft of the arrow had been feathered with one of the eagle's own plumes.
We often give our enemies the means of our own destruction.”
― Aesop, The Eagle and the Arrow
"under condition to deactivate half of those advertised features"... RDMA ? Using VSwitch instead of a Team in Windows Server (Blue screen-Boot loop in quite some cases, where you have to know how to recover)? VMQ-weirdness reappearing every few month depending on which moon phase was active when the driver was released? The list of things to NOT activate is long. And this is only for Intel networking cards for Servers, since this, besides CPU, is usually the only other Intel component in the "mass of servers" segment.
As for the laptop segment: Don't try to use your Intel graphics beyond the office-requirement level...
I was happy when Intel announced that they were dropping Optane for home users. I never saw any benefit from it and computer manufactures would have been smarter to increase RAM instead of putting int Optane. Optane was a real pain in the ass to try to do a backup or restore using a third party backup program booted from USB or CD because you had to put in the Optane driver to access the drive. Eventually I did discover that Optane could be disabled, first in Windows and then in BIOS and if done in that order, you didn't have to reinstall Windows. I always had a deep distrust of Optane because, without the Optane driver, you could not access data on the hard drive. Questions I never knew the answer to were things like if the Optane module failed, could you retrieve data from the drive using another computer? Could the Optane module be replaced and would you then be able to boot the system or do you have to reinstall Windows? Is there more than a 1 second difference in bootup time? I never really could see a difference in the laptops I worked on where I disabled Optane.
Perhaps Optane made sense in a server environment, but I have no experience with such a configuration.
Personally,Optane seems somewhat reminiscent of the RAMBUS DRAM technology. Yes, RAMBUS DRAM computers were faster than others of the time, but the monetary cost was high. As I recall, that technology also lost out to other technologies after only a couple years.
RAMBUS performed well, but was expensive compared to DDR. Couple that with the (very) thirsty, expensive and inefficient P4 it's advantages were neutered.
I played around with a dual processor RAMBUS system for a while. It was otherwise destined for the e-waste box. It could be tolerably made to run Windows 7, over a decade after the hardware launched. Whereas I would not fancy putting a K7 onto Win7 to be usable.
But that was also a multi-thousand pound workstation in it's day whereas the K7 could comfortably be had sub 800 quid.
Optane, by comparison, is a product I have never adequately saw a need for, nor the marketing people able to explain why I needed it. I can think of many, many downsides in security-land for a persistent memory system. Need boot speed? Get an NVMe. Need networking persistence? Have node redundancy.
High proportions of people have upgraded in recent years. That means sales across the board are likely to drop for a while. CPUs are usually good for 5 to 7 years. If I were Gelsinger, I'd be looking at how do I get my product lineup "right" for the next upgrade cycle. In about 3 years time. Their marketing people should have better data than I on the subject.
On 1st gen gfx cards; nobody expected anything from them. It's the 2nd or 3rd gen which will determine their fate - get them right and Intel stand to do well. Get them wrong and it's an expensive write-off. They also need to be aggressively shipping examples to developers to iron out snags early.
CPU-wise Intel have been on the backfoot for multiple years. e-Cores aren't a terrible idea, with the predictable software support catch up. The performance and power consumption of the core itself has to come next. High powered ARM derivatives would differentiate Intel from AMD while satisfying goals of both reducing power consumption and improving work done per clock.
But besides hardware, Intel also needs to learn how to live within its means in terms of cost control while still producing that next competitive item for the next upgrade cycle. Running multiple year losses in the hope of one good year later to make good isn't sustainable, especially if the track record says that the good year is looking even more unlikely.
Basically, Intel needs to get back in bed with the rest of the supply chain or it's current downward trend will turn continue.
Intel needs to get back in bed with the rest of the supply chain
Intel grew so big because (in part) they made a really good strategic partnership with Microsoft (hence the 'Wintel' abbreviation).
And now? Microsoft is fast diversifying off to ARM (and potentially RISC-V) and their dependency and strategic alliance with Intel is fast disappearing. Adding the fact that AMD chips are faster, cheaper and *available* means that Intel is going to have to do something special to survive in it's current form.
I moved to Hillsboro, Oregon a long ago to work for Intel. After 3 months, Intel announced their quarterly results. Record profits, and in order to insure an even more profitable next quarter, they were firing 1,800 employees. Including yours truly (I'm an electrical engineer if you were wondering).
Well, we all have our priorities, no?
Earlier in "Comments", it was stated, "...the US government is never going to give up on an American Tech giant.".
Forget the "Tech" part. There is absolutely nothing about the fact that Intel manufactures processors for computing machines which makes them 'sacrosanct', or untouchable, or worthy of saving under any set of circumstances.
So; the "...the US government is never going to give up on an American Tech giant.", eh?
What, exactly and precisely, makes Intel any different than all those 'giants' which the US government did 'give up on', whatever that particular phrase is intended to mean.
Because of the context, "give up on" appears to mean, "allow to fail".
What about The US Government's "giving up on" the likes of---for an extremely, very extremely short list---Eastern Airlines, Burlington Northern, K-Mart, Studebaker, TWA, Motorola Semiconductor...and, of course, KODAK?
Intel simply does not get a 'free pass' because it was, once upon a time, a brilliant innovator with a very large portfolio of diverse semiconductor products, and certainly not because it is now one of the very last of US manufacturers of an extremely myopic range of semiconductors. Its 'catalog' of semiconductor offerings is limited to an very narrow range of what is needed by the world's designers of electronics and consumers of semiconductors.
The electronics world runs, for the most part, on plebeian components such as diodes, transistors, and 'dumb' integrated circuits; and not on sophisticated computer processors.
Kodak is a great example. Kodak did not believe digital media and cameras were the future.
They stuck doggedly to film media when the rest of the world jumped to CCD; and low and behold, they are no longer a relevant player.
Intel would be bought out before it died completely, it's facilities and patents back catalogue alone have enormous value. Certain elements of Intel do very well. Arguably, the company is ripe for a break up
But yes, there is no free pass. A semiconductor firm doing badly is not the difference of tens of millions of homeless or not as the crash of the banks could have been.
What Intel is, is one of a very long line of formerly successful USian businesses that have failed to evolve; leading to decline. In turn, if the US and it's institutions do not evolve, the country will lose out. Propping intel up is not a solution to that.
Intel's Sapphire Rapids Xeon Scalable processors are without a doubt its most ambitious chips ever
Oh, I think there's some doubt. How is Yet Another x86 Chip more ambitious than, say, Itanium? I'm not claiming Itanium was good, but it was more ambitious than just another generation of the same damn architecture. And the iAPX 432 was significantly more ambitious than Itanium. Even the i860 was probably more ambitious than Sapphire Rapids.
I take your point that Sapphire Rapids is trying to cram a lot of improvements into this x86 generation, but it's still the same old ISA and the new features are known quantities. Sapphire Rapids is Intel playing catch-up, not taking a chance on something new.
Intel is just lucky that Apple isn't going to sell its ARM chips to anyone else. That gives them a year or two breathing space until AMD squeezes them at the server end and Qualcomm or whoever can do it squeezes them from the other side with ARM processors. Note that on a high level, ARM and x86 are very, very compatible.
Apple might actually sell more CPUs (inside Macs, iPhones and iPads) and probably more cores than Intel.
Intel's best hope is unfortunately China invading Taiwan so that the US invests in local chip manufacturing, though Samsung might still beat them even then.
Intel is the new RIM. RIM had a wildly successful phone that some people preferred to the iPhone for a while, but then the glass slab won and Blackberry couldn't compete. x86 is still an enormous market, but, and I didn't expect to type this a few years ago, AMD has the advantage because it doesn't have its own fabs.
15 years ago, or maybe even just 10, Intel's vertical integration still made sense because it had enough manufacturing volume that made sense to invest in the best fabs, but Intel missed the smartphone market and the smartphone chip fabs had the volume (and perf/watt pressure) to make smaller nodes happen. AMD has been able to surf on the wave of the smartphone chip fabs.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
