backtop


Print 57 comment(s) - last by Pirks.. on Jan 19 at 10:34 AM


Intel's Paul Otellini reported record earnings and in a conference call denied that ARM posed a threat to Intel's notebook CPU business, despite the fact that ARM would offer customers greater battery life on identical performance.  (Source: VentureBeat)

At the same time Intel is confident it can make headway in the tablets market with likely inferior products that have yet to be delivered to market.  (Source: Hi-Tech Russia)
Is Intel overconfident about ARM? Maybe, but it doesn't think so

Intel Corp. Chief Executive Paul Otellini was pleased to report strong earnings [report], with Intel making a fourth quarter net income of $3.4B USD (up 15 percent from last quarter) on revenue of $11.5B USD (up 3 percent from last quarter).  Intel's quarterly gross margin reached a record 67.5 percent, and the company reported an operating income of $4.3B USD.

For the year, Intel reported an incredible 67 percent rise in net income, which soared to $11.7B USD on net revenue of $43.6B USD (up 24 percent on a year-to-year basis).  The gross margin raised 10 percent for the year, and the operating income jumped 79 percent to reach $15.9.

Those strong results were overshadowed somewhat by Microsoft's announcement that it would offer full Windows support for the ARM architecture.  ARM is a superior architecture to Intel's championed x86 architecture in several ways.  It has more registers, so it eschews Intel's costly register renaming.  And it has fewer instructions, leading to more power efficient execution.  Thus an ARM laptop CPU could accomplish the same tasks while using less battery life.

Still, Intel's CEO Otellini claims his company isn't worried.  In a earnings conference call [audio, transcript] he comments on Microsoft's decision, stating:

In fact, in big Windows it had support for Alpha, PowerPC, MIPS and at one point ARM on the Vista program that they dropped. So this is nothing really new from that perspective. The plus for Intel is that, as they unify their operating systems, we now have the ability for the first time, one to have design from scratch, touch enabled operating system for tablets that runs on Intel that we don’t have today.

Secondly, we have the ability to put our lowest power Intel processors running Windows 8 or next generation Windows into phones, because of the same OS stack and I look at that as an upside opportunity for us. On the downside there is a potential given that Office runs on this products for – there is some creep up coming into, let’s say PC space. I am skeptical of that for two reasons. One, that space has a different set of power performance requirements where Intel is exceptionally good. Secondly, users of those machines expect legacy support in terms of software and peripherals that has to all be enabled from scratch for those devices.

Intel seems to be blaming Microsoft for its tablet delays.  However, Microsoft has showed off working Windows touch-tablets, where as Intel has yet to deliver tablet-geared chips (Atom-based "Oak Trail" and "Moorsetown").

Intel also expects the issue of legacy support to prevent ARM from making a larger splash in the Windows marketplace.

Intel is clearly feeling the heat from ARM.  That is ironic, given that it used to produce ARM CPUs, but chose to divest itself of those holdings. Intel acquired its ARM offerings in the 90s from its purchase of Digital Equipment Company (DEC).  At the time it took responsibility for the design and production the company's ARM-based "StrongARM" processors for mobile devices. In 2000 it transitioned to a newly named line of ARM CPUs called XScale. 

But in 2006 it sold its XScale mobile processor unit to Marvell. An XScale processor is found in the Blackberry Torch, among other devices.  To this day Intel and Marvell still co-own some XScale (ARM based) processor lines -- but only network processors, embedded processors and their ilk.  Intel firmly passed away its rights to mobile ARM designs.

There is no question that ARM represents a more power efficient architecture.  Of course that matters little in the desktop space.  In the server space it's mildly important, but GPU computing currently offers a far greater threat than ARM.  

Where the real trouble starts is in the laptop and tablet space.  ARM already rules the world of tablets, and Intel is unlikely to deliver a true competitor in terms of battery life in this space (hence its many delays). 

For laptops, Intel may currently reign supreme with its Atom-based chips and Core i-Series processors, but it faces a significant threat.  If Microsoft makes good on its promise of full in-Windows hardware support for ARM-based platforms, the story becomes the same as tablets -- ARM will be able to beat Intel's offerings on power, while offering similar performance.  The performance gap will largely be nullified by coming ARM chips, such as chips based on the eight-core A15 architecture revision

At the end of the day two things are sure.  First, Intel is doing great in the present tense.  It is recording record profits post-recession and enjoys a healthy lead in global CPU shipments.  

Second, though, is that ARM is serious threat to Intel's bottom line and growth opportunities.  ARM is unlikely to "kill" Intel's CPU business anytime soon (as we outlined, there are few advantages of ARM for desktops), but it may cut its sales.  



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

By DanNeely on 1/14/2011 10:39:29 AM , Rating: 3
Apple's biggest problems with PowerPC were cost and supply related. PPC chips were produced in significantly lower numbers which meant a higher per chip R&D cost. This was much more of a problem for Apple's $1-2k computers than for IBM's $$huge high end servers and mainframes.

Apple also had supply problems with getting enough chips at times. While these were partially self inflicted (Apple consistently lowballed estimates of how many chips they'd need resulting in IBM not having sufficient capacity) they caused chronic problems for the company. The x86 market is orders of magnitudes larger, and Intel (or AMD) can easily handle 10 or 20% jumps in demand from a company of Apples size becuase they're a much smaller fraction of the total market. For PPC Apple was at least the number 2 consumer of standard CPUs; while game consoles rivaled them in total demand the chips they used are customized and not mutually compatible. Because selling extra CPUs after they're no longer state of the art can still be done at a profit Intel/AMD don't have to worry about trying to tune their production capacity as tightly to expected short term demand. If IBM made significantly more PPC chips than Apple used they'd have a much bigger problem since they're nowhere near as easy to unload at a slight discount.


By kattanna on 1/14/2011 10:53:45 AM , Rating: 2
spot on!

and with their much lower total volume, there simply wasnt as much money for said R&D for both chip design and manufacturing processes.

as such, they were being left behind and there wasnt much motorola could do about it. at the time G5 chips were having issues with getting to and working in the 1Ghz range, while intel was in the 2Ghz+ range. and if though they were somewhat comparable in performance, they were losing out badly on the marketing front. which as we all know means a lot to a business.


By nafhan on 1/14/2011 11:25:43 AM , Rating: 2
The other big problem was power consumption and therefore lack of a suitable mobile part. The switch was made around the time Intel started selling the first Core processors. I was actually under the impression that this was the main reason for the switch...


By DanNeely on 1/14/2011 1:48:36 PM , Rating: 2
I hadn't considered it, but that was probably a factor too. High end servers don't put much priority on energy efficiency so any R&D costs on that end would end up almost entirely on Apple's shoulders.


By PrezWeezy on 1/14/2011 2:19:29 PM , Rating: 2
quote:
High end servers don't put much priority on energy efficiency


Not true. When you start talking about hundreds of CPU's the power efficiency starts to become a real issue. Not only do you have to pay for that power usage, which can be very expensive, you also have to keep the room cool, which can also be very expensive. Then when the power goes out you have to have a generator and battery solution which can supply those CPU's so you don't have down time. All of that costs thousands of dollars, so power efficiency matters more to them than it even does in the mobile and desktop space.


By nafhan on 1/14/2011 4:27:39 PM , Rating: 2
He's actually sort of right. Server CPU efficiency, especially back then, is generally more oriented towards efficiency at high utilization. PC CPU's need to be efficient across the spectrum, and especially at low utilization where they spend a majority of their time.


By HoosierEngineer5 on 1/14/2011 6:09:01 PM , Rating: 2
It amazes me that Intel has managed to drag this architecture along, kicking and screaming, for better than a quarter century. But, I suppose if you throw that much money at a single problem, you might even be able to get pigs to fly.

In fact, it seems to me that the biggest problem Intel is fighting is power dissipation (some cores run faster if others are disabled). Going to a more efficient architecture may help to reduce the problem.

One issue with the PowerPC is the 'kitchen sink' syndrome. There are so many peripherals and configurations that it is very expensive to develop with.


By Taft12 on 1/14/2011 9:02:41 PM , Rating: 2
Intel had A LOT of help from Microsoft dragging the architecture along for the past couple decades.


By encia on 1/18/2011 5:21:56 AM , Rating: 2
Microsoft joined ACE to replace X86 based PC.


"We shipped it on Saturday. Then on Sunday, we rested." -- Steve Jobs on the iPad launch














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki