Slashdot Mirror
Intel Confirms Major Layoff: 12,000 Worldwide, 11 Percent of Workforce (ieee.org)
Tekla Perry writes: It's all about the cloud and the Internet of Things, says Intel explaining the planned layoffs, which will affect some 12,000 employees. Intel CEO Brian Krzanich promises in an email today to employees, that the "transition" will be handled with the "utmost dignity and respect." According to IEEE Spectrum, "Intel Corp. today announced that it would cut some 12,000 jobs -- that's 11 percent of its total workforce -- by mid-2017, with the majority of those affected getting the bad news within the next two months. In a press release, the company said the 'restructuring initiative' would 'accelerate its evolution from a PC company to one that powers the cloud and the billions of smart, connected computing devices,' and that the company would be increasing its investments in 'data center, IoT memory, and connectivity businesses.'"
Cloud = Their core market (desktop+server CPUs) is in a deep consolidation phase where future purchases will be made by a relatively few number of large cloud players and total unit volumes will be drastically lower.
Internet of Things = Intel is being forced to chase razor-thin margins just to have a new market to soak up their excess semiconductor production capacity.
$300 Million because Diversity(TM) http://fortune.com/2015/01/12/...
"Now, I doubt any of you would prefer a rolled up newspaper as a weapon against a dictator or a criminal intruder."
I can see it if it's the team and support making their stupid Atom CPUs. They've utterly failed as a line to do anything like break into the phone business, meanwhile their main "Core" CPU line has been squeezed down enough with the Y series to get into fanless tablets. Meaning there's neither rhyme nor reason to keep the Atom line around at all.
Intel has to restructure. It's too bad that they couldn't move some of those employees over to the growing parts of the business. Intel's business is kind of a niche one. Where will those laid off employees go? AMD? Motorola? And that's assuming that they need people.
And when Yahoo! starts their fire sale, there will be re-orgs there as well as layoffs and it will flood the market with even more tech people.
There are some bad times coming to Silicon Valley.
It's not that we want people to lose their jobs, but maybe what Silicon Valley really needs is another DotCom-type crash.
I mean, what positive things have Silicon Valley as a whole really accomplished over the past 10 years?
Well, they've managed to make advertising and the collection of personal data far more invasive and pervasive than it was before.
They've given us "social media", which is really just another way of delivering targeted, inane advertising, and harvesting personal info.
Anything new showing some potential, like virtual reality, is quickly being hijacked as yet another method to, you guessed it, deliver advertisements and collect personal data!
The programming languages they've created, like Go and especially Rust, are nothing impressive. We're still using C and/or C++ for any and all real work, and will be for some time.
Databases have taken a big step backward with all of the NoSQL hype they generated in Silicon Valley.
Web browsers today are worse than they were a decade ago. Just look at how badly Firefox has regressed. Its UI went from being really usable to being awkward to use thanks to Australis, lots of good functionality was removed, lots of dumb functionality was added in, and its performance still causes problems for lots of people. Chrome isn't impressive either.
The price of everything, and especially of housing and rent, has been distorted beyond belief in San Francisco and the surrounding areas, causing untold headaches for long-time residents.
Silicon Valley has a lot of potential, but so much of it has been wasted these past 10 to 15 years. Maybe another economic reset is just what that region, and the technology industry in general, needs.
So you can buy a cheaper PC and get more overall processing done.
Unless you need that processing done locally for one reason or another, or else the prices being charged by "the cloud" were exorbitant or required subscriptions that essentially cause you to pay for unused cycles.
I'm not yet convinced this is actually the reasoning behind this trend. I suspect PCs are "fast enough" for the majority of the market, and the minority that requires faster PCs is going to end up paying more or possibly be starved.
The companies that compete with Intel are Global Foundries, TSMC, Toshiba, Micron, Samsung, Texas Instruments, etc..
..and we havent even gotten to RAM chips yet.
Intel is a manufacturer. ARM is not.
The manufacture of flash chips alone dwarfs any impact that ARM could possibly have.
Inside my desktop:
1 CPU
1 GPU
16 flash chips
Inside a smart phone:
1 CPU
1 GPU
1 ASIC
4 or more flash chips.
Picking up the pattern here?
If Intel is laying off so many, its because some of their fabs arent running 24/7. Intel has 16 fabs and maybe 3 of those (the most advanced) actually make CPU's. The rest bang out flash, ram, etc. Eventually the current "advanced" fabs wont be so advanced and will be making flash, ram, etc..
Intel can also manufacturer ARM designs (and has so in the past) so that cannot be it. The fact that they are not producing ARM designs while simultaneously they are laying off workers.. that doesnt tell a 'beaten by ARM' story like you imagine. It instead tells a 'beaten by the other foundries' story.
"His name was James Damore."
He said "growth days" and is 100% correct. Investors want exponential revenue growth even if the company loses money and will be dead in 10 years. Investors are stupid.
RISC already won, long ago. X86 has been RISC for a while. Legacy x86 instructions are translated into core RISC instruction and the later is (re)scheduled and run. No direct access to the RISC core is available,
Not surprising, as it's not guaranteed to remain the same from processor to processor, and probably doesn't remain the same from processor to processor.
Not only that, but they are lobbying for more H1B's while hiring ~2000/year
Custom electronics and digital signage for your business: www.evcircuits.com
Yep. I just love to confound compiler optimization enthusiasts by demonstrating a benchmark for my "inefficient" x86 code that runs rings around their "more efficient" compiler code. My dumb compiler and/or hand written assembly code is simpler and nowhere near as "clever" as GCC or Clang output, but it somehow runs faster?! It's because I know x86 opcode is getting translated by microcode into RISC instructions, and so I use a small subset of instructions with low translation overhead rather than try to be clever and do a bunch of optimizations to produce tricky "efficient" oppcode. What's more efficient isn't reduced number of instructions but easier to translate instructions. Even with no optimizations enabled, a compiler will do peep-hole optimizations and reorder instructions to save a register access, ignoring that the more complex instruction actually gets processed by a bunch more register moves in microcode. Yep, that LEA (load effective address) which most compilers use to squeeze an add and multiply into one instruction? It can be slower than issuing the individual operation opcodes and takes up about the same number of bytes.
You're speaking as if the foundries provide Intel a unique competitive market position. In terms of the x86 market they certainly do (at least they used to) - the mobile market is another story. Those vertically-integrated foundries can be a capital nightmare. The arrangement works well when growth is strong but become a multi-billion dollar albatross when it doesn't. This is why nearly every company has become fabless; the contract fab model works better for all market participants, for the fabless companies because they avoid the capital investment and the fabs who get better utilization of their resources when the semiconductor product mix changes.
As for the forest, you pointed out in your previous message how many more chips get put into a computer vs a mobile device. That's fine except for the fact that PC shipments for 2016 will probably be around 270 million, compared to a projected 1.5 billion smartphones. Add to that tablets, internet of things, microwaves, automobiles, etc.. etc.. etc..
"I think there is a world market for maybe five computers." Thomas Watson, president of IBM 1943
Amazon, Azure, iCloud, and a few others
READY.
PRINT ""+-0