As the heat from AMD reduces
So will the innovation at Intel slow down. We could all see it coming.
Intel is going to leave a massive new multibillion-dollar chip factory empty in Arizona as the slump in the PC market continues. Intel told local paper the Arizona Republic that the factory known as Fab 42 would remain closed for the foreseeable future, but it would be used at some point. “It will be used for future …
It's less that there's been no major developments, but more that a few years ago PCs reached the power level that is sufficient for most users. A 3-5 year old PC will be plenty powerful enough for 80-90% of people, who just want web/email rather than something like cutting edge gaming, video editting or similar more demanding applications.
Hence some new "killer feature" was needed, where the tablets and phones (and mixtures thereof) stepped in with their mobility and portability (and "all in one" nature with something you're probably carrying anyway). So PCs have gone from "need to upgrade to do more things" to "need to upgrade as the old one's broken" which is always a much longer cycle.
The problem is of course that the business model and the driving requirements for a desktop (or even laptop) PC is different from a tablet one. So it's less Intel vs AMD than the "old school" (including both of them) vs the newer custom-designed stuff like ARM based kit. Intel are trying the new market with the old mindset, which doesn't fit and so isn't working (look at the business model behind the new Quark processor - technically good but absolutely sunkin its Edison format by the price and what's not on the die that you also need to bolt onto it). They're getting there, but it's a slow evolution without guarantees that they'll get there alive.
I have one of these at home, from 2003 IIRC. It's a 4GB P4EE 3.4.
The SATA-I disks sucked big time and WD P-ATA disks were way faster btw.
Putting an SSD in has improved things quite a bit. Though the SATA-I interface limits the benefits, it is still a good step up from spinning storage.
No need to upgrade in the foreseeable future, unless something breaks.
(And yes I have a tablet and a smart phone and a MacAir for mobile and other uses - my home PC is used 90+% for work activity where I earn my money to pay for the other toys)
Like you say, you don't need a more powerful PC and you don't want another five 200W PC's in your home. But you might want twenty 0.1W devices controlling your light, heat and kitchen appliances. Microsoft have dropped the ball on this progress hence the shuttered factory, maybe Google can pick up the baton (Intel have an ARM license)
Intel has been nickel and dime ing customers forever. They have a crappy processor so next year, they add a little more speed. The following year, they add another core. The year after that, well, you get the point. Where the heck is my 8 core PC? That would be a hit for business. Instead, you try and peddle broken SX chips at a ridiculous price, or you give us some stupid nomenclature that has no meaning i3, i5, i7. i3 is dual core but so is i5, but only the top i5 might be a quad core, or maybe an i7. Yeah, confusing enough to say, "I think I will try AMD instead. Sorry, I'm not buying your sales pitch...
Multi-Core didnt give me as big as boost as everyone else so am very miffed by the statement, in fact I was rather disappointed after upgrading to a Core i5 the performance felt like that of my old AMD XP3500 single core processor (Comparing overal performance in pc games of the time which obviously still were single threaded).
As I kept the storage and graphics it wasn't until moving to one of those SSD Accelerator products I noticed the difference. The only product to make a difference to me is the SSD.
They've not left a multibillion fab unused. They built the building and then didn't put the fab into it.
Once you've actually built the whole damn thing, Si ingot makers, slicers, etchers and all the rest there's absolutely no way at all that you don't run the thing. For the running costs are entirely spit in relation to the capital costs of having got that far.
And it's all that gear that is "the fab", not the building itself.
Sure, Intel spent good money on the building but it's not multi-billions, no way.
One way of checking this is that if they had they would have already made a release to the SEC stating that they were going to take a multibillion write off of the cost of that plant. Which, at least as far as I know, they haven't. The costs of the walls and ceiling aren't sufficiently large to be relevant to Intel which is why they've not had to announce such a write off.
The above is all speculation BTW but it is how I read this story.
Damn right ... Its a $1.5B box and it needs about $3B worth of equipment.
Intel's cost of capital is about 1 percent. This is costing them $15M a year which is peanuts.
And its NOT a depreciating asset.
It also gives them an 18 month head-start over the next guy who decides to build a 450mm Fab.
Everyone posting so far is right -- desktops have finally reached the point where most people won't notice the speed bump afforded by a new processor. One person mentioned SATA's availability as the turning point...I think the point was actually when multicore came out. When you have one or more cores available just to run the background stuff that a Windows machine needs (like bloated AV software,) even a low end system becomes usable.
Tablets are going to kill a lot of the PC market, but I think desktops and laptops still have their place in business. Even typing emails on the average phone or tablet is a much less productive experience than the same task on a PC. The bloodbath is coming simply because the 3 year maintenance cycle on PCs is over. I do end user computing stuff for my job and regularly deal with HP, Lenovo, Dell, etc. There is a HUGE amount of overhead in their business PC divisions...VPs of nothing, endless revolving-door account managers, project coordinators, analysts, marketing guys...... That overhead was supported by BigCo, MassiveMart and OmniGloboCorp buying 100,000 new PCs each, every 3 years.
My job in end user computing leads me to believe that corporate types still use desktops/laptops to do actual work. An exec might view sales dashboards and stuff on a tablet, but execs generally don't create content themselves.
The engineering laptops here at <global defence company> are still on a 3 year refresh cycle but upgrades for a lot of the other types are indeed begining to slow. I know that the company sells off its old boxes to get back a good deal of the money that went into them but if the market as a whole collapses that will no longer be as lucrative... adding further pressure to delay upgrades.
How do we know that there is not an underground entrance into the building, and inside is an Earth version of Graystone Industries, and that there are not a bunch of Zoes running around shooting self-balancing, semi-sentient "i-wanna-live" targets dodging armor-piercing 20mm bullets?
Alternatively, don't forget that the NSA wants to build a quantum crypto-cracking machine. Would this facility be able to provide chips for that task? Can't masses and masses of chips be rackless, running in a fluid-like bath rather than in space-wasting chassis-farms?
Maybe Fab 42 is a cryptic nod to "Fab for Consumers and CONSUMERS -- public consumers needing new computers, and content/crypto consumers for the NSA", hahahaha. Just slipstream Section 31/NSA chips in with consumer chips, and pray that the NSA chips don't get mixed in and found out (whether or not the distinction is merely an electronic activation code/switch) -- some might end up in a hostile country.
Maybe they're building chips to calculate time travel?
Can we all just agree,
1. Update cycles have slowed in business
2. Consumers like phones and tablets more than desktops
3. Workstations (gamer boxes) still need to speed up, but that is a small market.
4. You think your, and your friends opinion about windows 8.x is important.
5. This is the year of Linux (mint rocks IMHO)
And thereafter have an interesting conversation devoid of these anecdotes and topics.
It might have been interesting 3 years ago, but seriously, talk about flogging a dead horse.
Biting the hand that feeds IT © 1998–2020