Intel & Altera to Build Next-Generation FPGAs on Intel's 14 nm Tri-Gate Technology
Intel has announced that it has reached an agreement with PLD (Programmable logic device) manufacturer Altera to product FPGAs (Field-programmable gate array) on Intel's 14 nm tri-gate transistor technology. According to the announcement, these next generation products will enable breakthrough levels of performance and power efficiencies not otherwise possible and further the company's ability to deliver on the promise of silicon convergence by delivering a more flexible and economical alternative to traditional ASICs and ASSP.
Furthermore, Brian Krzanich, Intel's Chief Operating Officer stated that "We look forward to collaborating with Altera on manufacturing leading-edge FPGAs, leveraging Intel's leadership in process technology and Next-generation products from Altera require the highest performance and most power-efficient technology available, and Intel is well positioned to provide the most advanced offerings."
Apple makes CPU's?.......... Oooohhh.... you mean the chips SAMSUNG produces for Apple.
AMD should pull an end-game action and pour 1B into R&D just to see what happens.
If they create something epic and companies love it - GOOD! 1B well spent.
I'd do that. That's probably why I'm not rich or a CEO. lol.
I do have lots of nice things though..
Investing 1B in improving their CPU/GPU is not going to do much to help when AMD is stuck with Global Foundries being almost two process nodes and 10B$ worth of process/fab investments behind Intel.
Apple makes CPU's?.......... Oooohhh.... you mean the chips SAMSUNG produces for Apple.
Don't forget that AMD have the PS4 and possibly the new Xbox, that's possibly billions for them over the next 10 years (I don't know what it'll be, I'll assume AMD get $40 per console sold, say 150m sold by 2020 that's $6b for them, hopefully they'll still be around by that time).
Because it's a fuckton of money for Intel.
Intel has yet to showcase a working 14nm logic tape-out (a working chip), although unofficial reports says that it does have it in-house already in August last year. Meanwhile, Samsung has showcased a fully functional ARM core on a 14nm tape-out in December last year. The question is why Intel, who is normally very quick to showcase their achievements still haven't showcased a working 14nm logic tape-out, while at the same time (at IDF last year) said that they were ready for 14nm manufacturing.
That it "is not even about consumer hardware" is not the point either.
It is about Intel having its first serious foundry client, something AMD tried but failed at before spinning off their foundry operations under Global.
Altera getting on Intel's 22nm and 14nm process could hurt Xilinx, Cypress and other CPLD/FPGA designers pretty hard if Altera has an exclusivity deal for programmable logic.
While this may not have a direct effect on consumer-level hardware, it will have a fairly direct effect on rapid prototyping and hardware-accelerated simulation rigs. It will also have a significant effect in telecoms and other places that use FPGAs for software-defined radios, network processing, video/signal processing and various other tasks. Reduced competition there due to Altera having a monopoly on Intel's process for FPGAs could eventually translate into higher service costs, higher development costs, etc. (they're the only ticket in town for a given logic density, throughput per watt or other metric enabled by Intel's more advanced process) much the same way AMD failing to keep up with Intel caused Intel to freeze price points.
Domino effect. You may not feel it now but you may still regret it later.
I would guess it to be safe to say that its because of a lack of competition on AMD's part and a serious feather in their cap considering based off what your saying is true then they would already have working proto types a full year in advance and will be able to refine the product to match anything AMD throws at them. currently they still lack a CPU that matches the Upper end of Ivy bridge in terms of performance. Also take into consideration the fact that once 14 nm gets released they really have much more significant challenges as far as transistor density and shrinkage. not to mention cooling those processors as many of you guys know now they tend to run hot.
Well. That's all very good and so on but the world has moved on. AMD is not really a competitor to Intel, ARM processors and the army of fabs producing them, however is. As said, 14nm working ARM processors have already been showcased by Samsung. Both TSMC and GF are in fact getting ready to showcase their own working 14nm ARM processor tape-outs in the coming months.
The only viable reason I can see is that Intel has encountered problems, and are in fact not as far ahead of the pact as they would like us to believe.
They design their own ARM SoC (http://en.wikipedia.org/wiki/Apple_A6), Samsung only provides the foundry.
Why are we discussing AMD against a PLC maker? It's nice for PLC users and industrial clients. It has NOTHING to do with the CPU market. These are very different areas
The A6 was designed in-house, so yeah they don't manufacture it, but they have started designing their own. So if you're going to make that argument against Apple, i could also say that ARM doesn't make CPUs either (they don't, just design) and so bla bla bla.
@Saturnus:
Broadwell isn't in mass production (they don't have their fabs up yet), but considering that Micron was showing off DDR4 RAM at CES, running on a black box they weren't talking about, I'm sure Broadwell is somewhere out there. Though that could be IB-E as well, since that's rumored to support DDR4 too.
Anyway, AMD is still very much competing with Intel, probably not with the i7s but otherwise, it still is. The FX-8350 is a solid competitor to the i5s and their APUs are going to turn things around quick, Bay Trail is going to receive solid competition from Kabini/Tamesh.
You'll have to link to the functional 14nm ARM chip story, because the last i remember hearing about the 14nm Samsung chip was a tape-out somewhere in the middle of last year. That means it's still at least 2 years from production. Intel have no reason to demo Broadwell before Haswell is out, that's common sense. I love the fact that everyone automatically assumes that Intel's having "problems" at the drop of the hat. They've been building 14nm fabs since LAST YEAR. I'm sure they've got their stuff together. They were working on Tri-gate transistors back in the 1990's.
ARM have sort of hit the wall when it comes to efficiency anyway with the A15s, a quad-core A15 will have Haswell proper for competition, Ivy Bridge quad cores (mobile) already idle at 2w, and we know about the Y-series too.
I'm more interested in what Qualcomm manages to do (from the ARM side), they seem to be the actual leaders in efficiency currently, followed closely by Apple's A6.
Another point you'll have to consider is the foundry, currently the only company that can sustain huge volumes of cutting edge processes is Intel, though i believe Samsung is indeed hot on their heels. TSMC has too much work.