General Garcia is dead now, but there are other Garcias - Elbert Hubbard
|
|
|
Intel should get out of the mobile market
ABI RESEARCH ANALYST Jeff Orr has penned a report critical of Intel's chances in the chip market for mobile gadgets.
He said Chipzilla isn't going to beat firms like Qualcomm, Texas Instruments and others that use ARM processor designs.
ARM-based ultra-mobile devices will surpass x86-based devices by 2013 which is a reversal from this past year, he said. At the moment 90 per cent of ultra-mobile devices are x86-based.
Orr said that 2010 will be pivotal for building momentum behind non-x86 solutions and gaining adoption. He thinks that the hard work is already done with ARM pushing to ensure that popular software, browsers and operating systems worked on its instruction set.
Android, which runs on ARM, being jacked under the bonnets of a variety of devices, and making sure Adobe Flash runs on ARM-based chips is what will help the company gain favour, Orr said.
He thinks that ARM has always had an advantage in mobile applications because chips based on the ARM instruction set were designed to sip power rather than glug it. This means a longer battery life and presumably a smaller battery and end-user device.
However the biggest hurdle it had was that most of the software people want to use is designed to run on x86 chips.
But in the last two years, thanks to the efforts to port software to its instruction set, and the overall movement of applications to the web, ARM has whittled away Intel's advantage, he said. µ
Share this:
Share
Tell me hoohoo, which Atom uses 10W? Do you mean the platform?
I have this Atom based Mobile Internet Device: http://www.umpcportal.com/products/Viliv/S5
It uses 4-4.5W as a whole system when the device is in use playing video.
If you get that in a smartphone size, and bring even lower power components, along with CPU advancements, it'll get much lower.
Those looking at current Atom Netbooks and MIDs/UMPCs and concluding it can't compete with ARMs in Smartphones have ABSOLUTELY ZERO clue what they are talking about.
You can't run an MID or cell phone or Jesus phone with a CPU that consumes energy at a rate of several watts if you want long batter life.
Atom AFAIK is still in the roughly 10 watt range. This is fine for a netbook. It is not fine for a cell/MID/etc.
Non-x86 chips will be hobbled until they work out a way to take an existing x86 binary file and morph it into one using another instruction set (this could be done fairly invisibly to the user and the converted file cached). I thought Transmeta figured this out years ago but I've not seen it being adopted much.
Intel is right on track for taking down ARM processors as the most power efficient and widely used mobile chips.
I don't know what MirekCz is going on about but ARM netbooks is what I was referencing.
DavidC1 has it right. It's just a matter of time before intel has the mobile market in their hands if they don't screw it up.
The analysts are never right in predicting these things.
The big difference in power consumption between current Atoms and smartphone ARM will be solved with Moorestown. In fact, Moorestown is a specific, MID and Smartphone optimized Atom.
The CPU/GPU/Memory Controller/I/O controller aren't the only reasons for higher power. You can optimize by using low power interfaces, special BIOSes, optimized Operating Systems, and integrating multiple motherboard components into single, bigger chip.
The difference in battery life between ARM smartphones and Moorestown smartphones will reach 1.5x with similar hardware configs. If they get true SoC with 32nm Medfield, they WILL be power competitive. And it'll have performance to boot.
That will shut up the analysts.
This article is obtuse. 'Intel has 90% of the market', but Intel should leave the market.
I dunno, it just doesn't get much dumber.
Well, Intel know how to do things well.
And they have economy of scale.
Intel is a company typical of the USA:
They do it wrong until they do it right.
This analyst does not have inside Intel insight.
Of course nobody does. But what this means is, that he cannot factor in what he does not know.
On the other side, everything regarding ARM is well known, due the basics of their business.
ARM and Intel have pros and cons.
ARM will continue to have more designs because the market is growing in total, the same holds true for Intel.
Morphing software to another platform is a major headache. It's even an issue to port software from Desktop-Windows to Windows CE / Windows Mobile.
Whether there will be a winner, I do not know.
What I know is, the one who counts out Intel already has lost his head.
The analyst is a bit stupid.
He is like a lightfooted person who tries to predict an earthquake by analyzing the historical weather forecast data.
Maybe that analyst needed to make headlines.
A clever analyst would be more cautious in what he predicts and what he is basing his predictions on.
Whatever, Intel probably did have a hearty laugh about his analytical skills.
If I have to read the word sip next to the word power one more time, I may lose my lunch.
The argument boils down to cost vs compatibility. The lower cost of ARM - in both direct silicon cost and power is compelling, but compatibility with x86 is something to be reckoned with - consider iAMD64 backward compatibility for both 32-bit and 16-bit modes in hardware (although M$ couldn't support 16-bit in S/W, that would have been much too complicated).
Web based stuff - Java in particular - should help blur the lines for compatibility, although the Vole is doing its part to keep x86 in the mix with limited portability.
I think what most people fail to grasp, is that many ARM designs are fabricated by a foundry on a 65nm generic process with none of the secret Hi K sause that Intel has implemented for 45nm and the rev 2 sauce for 32nm. Intel is rapidly moving all its SoC desings onto the new 32nm node(with a version tweaked specifically for SoC's) where the power requirements are dropping off the cliff. As Intel continues its relentless Tick/Tock model and moves further ahead of the foundries, the power requirements for ARM -v- Atom will reach parity and x86 compatability and probably even price as Atoms are fabbed at the 22nm node will become a huge Intel advantage.
At the moment 90 per cent of ultra-mobile devices are x86-based.
What's an ultra-mobile device? Phone? Small tablet? Haven't seen one atom in those devices available in shops. It's ARM all the way.
Yes, compatibilty with x86 is a bit of a problem, but if you use 1/10th the power to do the same work you're WAY ahead and as stuffs move to internet with everything it's not a problem anymore.
ARM is also getting to Atom on the performace front. Tegra 2 and similar devices will knife Atom to death in anything smaller then netbook.
He is stating what anyone who reads a small amount of tech news already knew. That ARM make chips that consume less power than Intel, so the ARM chips run for longer.
Great for battery life, and cheaper batteries mean lower cost of manufacture, or longer battery life.
I think both ARM and Intel will move forward with their designs.
Intel will have analysed the ARM design and see where Intel can make improvements.
It will be Cat and Mouse as usual. And with the Mobile Market set to be grow globally I expect that Intel and ARM will be in it for a long time to come.
The Atom is fantastic for what it is. I don't know why people keep saying ARM is where it's at and ARM this and ARM that. I've not really seen an ARM chip do as well as a main netbook processor as the atom is.
Losing x86 compatibility kills it for quite a lot of people. The next line of main laptop processors should be even more battery friendly.
So, it's a load of garbage. ARM makers are pissed at how well the ATOM chip is doing and pinetrail appears to be even more powerful.
Some of these analysts really suck at analyzing the information properly.