Slashdot Mirror
Intel's 14-nm Broadwell CPU Primed For Slim Tablets
crookedvulture writes Intel's next-gen Broadwell processor has entered production, and we now know a lot more about what it entails. The chip is built using 14-nm process technology, enabling it to squeeze into half the power envelope and half the physical footprint of last year's Haswell processors. Even the thickness of the CPU package has been reduced to better fit inside slim tablets. There are new power-saving measures, too, including a duty cycle control mechanism that shuts down sections of the chip during some clock cycles. The onboard GPU has also been upgraded with more functional units and hardware-assisted H.265 decoding for 4K video. Intel expects the initial Broadwell variant, otherwise known as the Core M, to slip into tablets as thin as the iPad Air. We can expect to see the first systems on shelves in time for the holidays.
Because what I was missing from a tablet was 4K movies!
Bye!
You just haven't seen a movie the way the director intended, until you've seen in on a 10 inch tablet in 800ppi at an airport. Now, how do I get this 160 gig movie on there.
I am MUCH more interested in Broadwell DESKTOP chips. I'm using a Haswell Xeon E3-1245v3 in a server now, and it speedsteps all the way from 3.4 GHz down to 100 MHz under light load. Ivy Bridge only stepped down to 800 MHz, and Sandy Bridge only stepped down to 1.6 GHz.
That was my first thought. What does a tablet need 4K compatibility for!?
Though I guess technically rather than having a 50" tablet, it might allow someone to use the tablet as a media device to the TV.
I used my Samsung phone for example in a pinch when both my Netflix, and my media computer was on the fritz.
However that said, they better start offering some much larger storage configurations if they plan on people carting around a bunch of movies that don't look like garbage on 4k.
Not yet but soon, When intel make this chip, Apple could do the same with the Ipad as they did with macs. This time it's from Arm to Intel instead of PPC to intel
http://chimpbox.us
Given that the covalent radius of silicon is 111 picometers, that comes to a channel that's 63 silicon atoms across.
And I thought 65nm (~300 silicon atoms across) was impressive five years ago.
The transistor budget may still be scaling according to Moore's law, but that's failing to translate into real-world speed increases. The 5% increase in single-core IPC is weak sauce. And an annoying number of apps don't scale to multiple processors, or scale badly (Amdahl's law is unforgiving...)
You can add more cores, add more compute units to your GPU, or add DSP (Broadwell) or FPGA (Xeon), but that has an ever decreasing marginal impact on real-world speed.
We're probably stuck in a "5% IPC increase per tick/tock" world until they eventually shift off silicon onto Something Else (III-V semiconductors or something more exotic like graphene)
Going from one third-party chip to another is fine and dandy, but why the hell would Apple -- especially Apple! -- dump their own design?
"[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz
Power is governed by change of states per second. It varies by the voltage, but by the square of the current. There's only so much saving from reducing voltage, too, as you run into thermal issues and electron tunnelling errors.
You are much, much better off by saying "bugger that for a lark", exploiting tunnelling to the limit, switching to a lower resistance interconnect, cooling the silicon below 0'C and ramping up clock speeds. And switching to 128-bit logic and implementing BLAS and FFT in silicon.
True, your tablet will now look like a cross between Chernobyl, a fridge-freezer, and the entire of engineering on the NCC-1701D Enterprise, but it will now actually have the power to play those 4K movies without lag, freeze or loss of resolution.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
As I have said for years it'll be easier for AMD and Intel to scale down than it will be for ARM to scale up because both companies already have such a high IPC. Between the new Jaguar cores at AMD and Broadwell and the latest Atom chips at Intel it is looking like by second quarter of next year the only thing ARM will have going for it is price. I predict ARM will end up in devices where price is king, your ultra cheap tablets, STBs, and industrial applications while X86 ends up taking a large chunk of the mid and high range.
ACs don't waste your time replying, your posts are never seen by me.
They've done it before--ADB.
One issue, though, is how long would Apple maintain both? Yeah, Apple dumped 68K for PowerPC, but they also stopped developing PowerPC. When Apple switched to Intel, they stopped doing things for PowerPC.
They could probably switch the iPad Air to Intel. But what about the iPhone? Apple--and 3rd party developers--would probably end up supporting both ARM and Intel for several more years in iOS.
They have the tech to make the best low power processor(s) on the planet...if only they would make an ARM one!
4wdloop
I have a Thinkpad 8 and a Miix 2 8. The Thinkpad 8 is a desktop replacement. I use bluetooth for keyboard and mouse, run an HDMI monitor, and stick power through the USB. It works well, but not perfectly. I'll upgrade to a good broadwell or cherrytrail. Anyway, the future looks awesome.
... that is 12 nm slimmer than it might otherwise be without this new technology.
They did, at one point. They bought the rights to StrongARM and sold it for some time, then abandoned it completely.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
These innovations make me want to buy Intel stock.
StrongARM was before iDevices...
I mean a 14nm ARM SoC for mobile devices, in fact why wouldn't they now when they push with x86 into this market space?
Can't be profit margins and they do not care about Windows anymore there, do they?
4wdloop