The x86 Power Myth Busted: In-Depth Clover Trail Power Analysis
anandtech.com
anandtech.com
Now nothing stopping Intel doing ARM chips down the line or indeed doing some hybrid chip. They have many cards they can play if they don't get the sales they want.
I would love the day when you could get a motherboard and the current Intel and ARM chips are pin compatable so you could drop what you like in, wouldn't that be a wonderful dream.
I'm already crossing my fingers that AMD will offer that, as they have announced they are developing an ARM chip and historically are friendlier about socket compatibility.
Not really sure whether it's actually possible though. It also might not be quite as dreamy as you and I are imagining- are drivers CPU architecture dependent?
I admit the AMD/ARM possibilities you mention are certainly very interesting prospects.
And I've never quite understood NVidia's design choices with Tegra 3. All the other ARM SOC vendors used a low power version of the process node they were on for their ARM cores, but NVidia used the standard process version for 4 of their cores and the low power version only for the 5th core. And yet those four cores are clocked only slightly faster than the cores of their competitors, while their 5th core is clocked way lower.
Now, NVidia's GPUs are top notch and while it hasn't been that successful in recent generations of phones I assume that NVidia's existing Windows software compatibility is why it was selected for the Surface.
EDIT: I seem to remember that there were some big issues on the GPU front a couple of years ago with NVidia's design tools and TSMC's 40nm process having something of an impedance mismatch. The big improvements between NVidia's GPU's from that time to their current generation attest to these issues being sorted out.
I still don't get how embedded chipset manufacturers can keep competing with the largest graphics company in the world in the gpu space so easily though. Tegra 4 is supposed to be jumping from 12 up to 72 in the highest end model Kepler cores instead of old 8000 cores. If they manage to get some super efficient high performance graphics in mobile I could see a games revolution around it.
Especially if they put that in the Nexus7#2 next year.
Just for some perspective:
Samsung total revenue: $247 Billion Apple total revenue: $157 Billion Qualcomm total revenue: $19.1 Billion Texas Instruments total revenue: $13.7 Billion NVidia total revenue: $4.00 billion
NVidia is the relatively small fish in the pond here, and if they have relatively more graphics experience the others have relatively more experience with everything else in this space.
Of course, this would have worked a lot better if people cared about WinRT at all, but I'm sure that was the idea.
I don't think Intel needs MS to say that. The lost money on the table in the smartphone-tablet ecosystem is enough incentive!
For a lot of people, a smartphone or tablet is sufficient to meet their computing needs. Especially with ability to connect a TV and keyboard, and a number of small ARM based "mini PCs" starting to arrive.
If I could get a Surface Pro with a Haswell at the listed Surface Pro prices, it would be a no brainer. But with a 3rd-gen Core i5 -- not so enticing.
Next year there will be a 22nm version of Centerton, Avoton, with a 14nm processor following the year after. Centerton, and possibly Avoton, will be out before the first 64-bit ARM chips.
http://www.anandtech.com/show/6509/intel-launches-centerton-...
It looks like Intel is taking the Atom's intended market serious enough to extend their tick-tock cadence to it.
Eitherway it certainly will wake up Intel, took AMD with the XP series chips to get Intel into gear many years ago and they need another kick for momentum.
But on a like for like generation of CPU and nanometer size ARM certianly have more headroom. Intel though certainly have a crazy no lose situation as they have a full ARM liscence so can always make and sell ARM chips.
Intel also have there multi core knights corner CPU's and they will only get better and have many area's of research into feilds that have yet to emerge. I find there whols onchip radio research very tantalising for future chip to chip communicatons with regards to production costs of servers.
Fun times, and as a consumer I'm happier we now have a race on our hands.
"User experience (quantified through high speed cameras mapping frame rates to user survey data)"
With Atom, they've always had the performance advantage. In fact they've had a big advantage over ARM at the time when when Atom came out. It was like several times more powerful than the ARM chip inside iPhone 3G.
But to be competitive, Intel needed to compete on the other 2 levels, as well. So they kept Atom's performance mostly the same, while they tried over the years to push Atom down from a 10W TDP. But while Intel tried to lower the power consumption, ARM's chips were gaining in performance - faster than Intel. So now when Intel has finally reached parity with ARM in energy efficiency, ARM has surpassed it in performance with Cortex A15, and from what I've seen so far, by a lot.
Then there's also the pricing issue. I think they've improved there quite a bit. An Atom chip used to be close to $100. Now it's probably more like $40-$50 (this dual core version). But that's still like at least 50% more than an ARM chip.
So Intel still has a lot of catching up to do in "overall competitiveness", because right now it looks like they've only switched the much bigger performance leadership for "energy efficiency parity", and are losing at the other two. I'd like to see how high they are willing to go with Atom on performance, until they realize that it's going to cannibalize their Core i3 market.
It's easy to dismiss Intel's offerings as not being competitive with state-of-the-art ARM designs, but only a sucker would dismiss Intel and rest on their laurels. Intel's 2nd and 3rd generation chips are just going to be that much better, and ARM could be in serious trouble in as little as a year or two from now. Imagine, for example, an Intel based android powered tablet that can run stock Windows in a VM or emulate windows and can run most windows applications and perhaps even games, while also being able to run all android applications. And do so at a cost/performance/efficiency level equivalent to any ARM based system. That's the sort of thing that should send shivers down the back of Samsung or even Apple.
If I remember correctly, projected volume of ARM CPU's for this year is in the region of 3 billion units.
The smartphone and tablet market is not even near being the majority of ARM CPU sales. So I don't think ARM will despair over Intel's designs just yet.
That roughly matches my memory of previous recent years' reported sales.
If we're going that far, we should imagine an iPad that is also a Mac.
Also, it is not a far off prospect at all, since the Surface Pro tablet will run Windows 8 Pro (on a Core-i5). It'll be able to run almost any windows app, even photoshop or visual studio (or MS SQL for that matter).
It's a nitpick, but surely you don't mean this literally? What if Intel's offerings beat ARM chips on, say, efficiency and price, or efficiency and performance... would you still believe ARM to be in a superior position?
AMD has been competing with Intel on price for the past two decades.
I know which of these two stocks I'd have rather owned over the past ten years:
http://www.google.com/finance?chdnp=1&chdd=1&chds=1&...
The data from this article is the first I've seen that indicates Intel could dominate the high-end mobile chip market in the near future.
Not with mobile trends and ARM uptake trends persisting. ARM speeds are getting more and more competitive. (I would argue, please forgive me, that signs are pointing towards Linux uptake. Intel's high-margins are on consumer processors as well, surely a large percentage of it. It will hurt as mobile continues to displace desktop and as desktop increasingly targets ARM.
Very unlikely. First, because ARM already has chips that are much more efficient, like Cortex A7. Intel can't bring Atom to that level right now. It would take a few years, just like it took for Atom to get from 10W to 2W. By that time A7's successor would probably again be more powerful.
I also doubt it will be cheaper. So far no indication that Intel has managed that compared to ARM - ever.
>or efficiency and performance
That may already exist, in the form of their high-end chips, and in terms of performance/Watt. But their prices are 10x bigger. So pretty irrelevant when it competes to competing with ARM in the same markets.
I dont think anyone should doubt Intel's Enginerring Power to bring Power and Performance Level unseen before to SoC market. The only problem is Margin and Price. Who these should work out are remain to be seen.
Long before Atom even came out a dualcore Core Duo ULV processor with a TDP of 5 was usually twice as fast as the Atom, if you disabled one of the cores on the Core Duo ... In other words about four times the performance at a much lower power consumption. I'm writing this on a 7 year old laptop sporting such a processor... Yes it is slow, but compared to an atom it the performance is just hysterical.
The Atom is, and has always been, a pathetic joke in all regards except for price. And the performance is so abysmal that the only reason it ever sold was because of shortsithed consumers that can't look passed the price tag. Which of course took everyone by surprise, and intel obviously haven't dared to improve on Atom too much since it's launch for this very reason.
It just isn't that simple.
Really? From TFA:
"""For us, the power advantage made a lot of sense. We've already proven that Intel's Atom core is faster than ARM's Cortex A9 (even four of them under Windows RT). Combine that with the fact that NVIDIA's Tegra 3 features four Cortex A9s on TSMC's 40nm G process and you get a recipe for worse battery life, all else being equal."""
>The untold story of Intel's desktop (and notebook) CPU dominance after 2006 has nothing to do with novel new approaches to chip design or spending billions on keeping its army of fabs up to date. While both of those are critical components to the formula
If those components are "critical to the formula" then you cannot say that "it has nothing to do" with them.
Critical means: it has absolutely something to do with them too.