Print 82 comment(s) - last by jharper12.. on Jun 5 at 8:50 PM

Cheap small notebooks and netbooks, like the HP dm1Z have powered Fusion to massive sales, taking rival Intel Corp. by shock.  (Source: Netbook Live)

AMD's Fusion is a mid-market powerhouse that Intel has no direct answer to in terms of price and performance. Atom is too weak; Sandy Bridge is too expensive.  (Source: Funimation/Toei)
Company sold 5 million chips since the Fusion platform's launch at the start of 2011

As if Intel Corp. (INTC) needed any more bad news after all the grief ARM Holdings plc (ARMH) is giving it, it appears that Advanced Micro Devices, Inc. (AMD) may be closing the gap in notebook and netbook sales.

Late last month, Microsoft's Bill Koefoed, general manager of investor relations, reported "a 40 percent decline in Netbooks" in Q1 2011.  In Q1 2010, netbooks sold about 10 million units, so this indicates that about 6 million netbooks sold in January, February, and April.  If that pace were sustained, through May around 10 million netbooks would have sold.

AMD's Raymond Dumbeck announced this week that AMD sold 5 million "Fusion" chips in the first five months of the year.  AMD's Fusion chips went into a mix of 10-, 11-, and 12-inch notebooks.  Given that the line between a "netbook" and "ultra-portable PC"/"small PC" lies somewhere in 11-12 inch range it seems like a fair guess that A.MD sold about 3 million "netbooks" -- or about a third of the total market (~10 million units)

That's incredible news for AMD which traditionally had little sales success in the mobile sector and virtually no sales in the netbook sector.

The flip side of the AMD victory is that Intel's ultra-mobile Atom processor appears to have bled a great deal of market share in a very short time.  Atom has suffered as AMD reportedly undercut it with Fusion and delivered a superior on-chip GPU compared to Intel's solution.

Intel's only saving grace has been that the rush to Fusion has taken AMD somewhat as surprised.  Mr. Dumbeck comments that the company is currently sold out of its existing stocks.  That shortage slowed sales from approximately 3.9 million in the first quarter to a mere 1.1 million the next two months.

That brief reprieve for Intel may be short-lived, though.  AMD is reportedly working to increase shipment of existing Fusion processors, as well as widening the channel to include new models.

AMD's current Fusion CPUs fall into the Brazos family, which feature the company's new mobile Bobcat core.  They are split between Ontario, which features lower power consumption and lower clock speeds, and Zacate, which features slightly higher power requirements and clock speeds.  Both Brazos flavors are built on a 40 nm process.

The company is preparing to ship a new type of Fusion "advanced processing unit" (APU) codenamed Llano.  Llano packs a beefier K10 core (found in the Phenom II and Athlon II processor lines), but a die shrink to 32 nm should help to mitigate the power bump, slightly.  The processors are expected to come in dual core and quad core varieties and fall between 25 watts and 100 watts in power consumption.

Intel doesn't exactly have a new processor that can compete at the same market point as Llano.  Atom will be much weaker than Llano.  Sandy Bridge will be much more powerful, but will also be more expensive.  The lack of a mid-market CPU offering from Intel should help Llano see strong sales when it begins shipping early next month.

Brazos has clearly exceeded AMD's wildest expectations.  With support from Dell Inc. (DELL), Sony Corp. (TYO:6758), Acer Inc. (TPE:2353), Hewlett-Packard Company (HPQ), and Lenovo Group Ltd. (HKG:0992) AMD's Fusion project is seriously threatening Intel's netbook and notebook processor offerings.

With 28 nm enhanced Bobcat 1-4 core models and 32 nm 2-4 core Bulldozer based models due out next year, AMD clearly hopes to continue to apply the performance pressure on Intel's Atom line and pricing pressure on Intel's Sandy Bridge line.

As with its success in the GPU market, AMD's turn-around owes to the company shifting its focus to the consumer budget and mid-market sectors, versus the traditional enthusiast race that it has waged versus competitors in the GPU and CPU markets.  A company focusing on the majority of consumers, versus a select few?  What a novel thought!  But it sure appears to be working for AMD.

Comments     Threshold

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

By ncage on 5/30/2011 3:48:13 PM , Rating: -1
Intel REALLY needs to aquire Nvidia (assuming it will be OK'd by the SEC). No only would it allow them to counter the threat of fusion and future AMD CPU/Graphic solutions but it also would give them a leg up in the mobile sector with Tegra. Yes it would be extremely expensive but apparently intel isn't able to design their own high end video solution (larrabee)

RE: Nvidia
By kensiko on 5/30/2011 4:08:33 PM , Rating: 3
Sure it would help Intel, but it would kill AMD !!!

No competition, no low prices for us, consumers. This must not happen !

RE: Nvidia
By YashBudini on 5/30/11, Rating: 0
RE: Nvidia
By YashBudini on 5/30/11, Rating: 0
RE: Nvidia
By SPOOFE on 5/30/11, Rating: -1
RE: Nvidia
By Targon on 5/30/2011 10:05:02 PM , Rating: 5
This is also a false conclusion. The difficulty in designing a CPU that is fairly competitive is so high, it would be virtually impossible for a new player to enter that particular market. The key is all in the compatibility that people demand, so that existing software will work on the new chip.

Basically, if AMD, which has a TON of experience in this area, can't compete, we will NOT see a competitor enter the market. Most start-up companies come in with millions to spend on the initial R&D, compared with the $100 BILLION that would be needed to even make an attempt at getting into the x86 market. Chip fabrication ability is where AMD has had the greatest difficulty, since a 32nm version of the Phenom 2 would have allowed AMD to release a faster chip before now, and going from 32nm down to 22nm is a huge increase in fab complexity. Yes, AMD split off the fab business, but who do you think is going to make processors for AMD, some other company that also can't compete with Intel?

RE: Nvidia
By SPOOFE on 5/30/11, Rating: -1
RE: Nvidia
By p3ngwin on 5/31/2011 12:45:14 AM , Rating: 2
Basically, if AMD, which has a TON of experience in this area, can't compete, we will NOT see a competitor enter the market

also a false conclusion.

An example would be Nvidia and their recent entry into the ARM processor market.

Nvidia had zero experience with ARM architectures, yet since Nvidia's first ARM chips in 2009 they have been first with dual-core ARM chips and even first with quad-core.

They beat all other ARM licensees who have had as much as a decade each head-start on Nvidia. Qualcomm is one of the most successful ARM licensees, with their ARM chips in the majority of most smart/phones for over a decade.

Nvidia beat them to multicore within 2 years of starting from scratch. then did it again with quad-core. Nvidia is demonstrating working silicone of it's Kal-El quad-core ARM-based processors at this year's COMPUTEX (currently running), yet where are the similar chips from the other ARM licensees ?

Nvidia is also aggressively (as is their style) perusing their project "Denver". an ARM based processor for more powerful markets such as laptops, desktops and even servers. no other ARM licensee is even thinking of putting ARM in those markets.

If AMD were to die tomorrow, somebody would happily step in and buy up the juicy patents and enter the processor market, or infuse an existing processor company with even more resources.

RE: Nvidia
By mvs on 5/31/2011 4:01:46 AM , Rating: 2
I disagree. Examine some of the market parallels. Look at how many commercial airplane manufacturers there are. Or how many railroad locomotive manufacturers. Or how many space launch vehicles. Intel vs. AMD is fairly unique.
Nvidia had zero experience with ARM architectures
Nvidia didn't just start yesterday. Check out
posted Feb. 11, 2008 it talks about the APX 2500. Basically the start of the Tegra line.

There is also a difference between being an ARM licensee and having an x86 license [which doesn't exist, AMD being the lone exception].

If you read
the exclusivity the x86 market has enjoyed appears to be coming to an end. Windows 8 will run on other architectures w/ Microsoft's blessing and support.
If AMD were to die tomorrow, somebody would happily step in and buy up the juicy patents
yeah, even somebody like Intel.

RE: Nvidia
By encia on 5/31/2011 7:30:36 PM , Rating: 2
If you read

"Update for Windows 7 Server Beta for Itanium-based Systems"

Since Windows NT Itanium line is EOL, ARM version can replace this non-X86 Windows NT line.

RE: Nvidia
By justjc on 5/31/2011 3:43:10 PM , Rating: 2
Why would it kill AMD?

Intel and Nvidia have had cross licensing deals, since 2004, giving Intel access to all relevant patents and licenses. If they had to merge with Nvidia it would primarily result in a clash of cultures that will cost Intel an enormous amount of losses in both resources, valued employees and lost research. If anything AMD would probably gain from Intel making such a move.

RE: Nvidia
By YashBudini on 5/30/2011 4:32:45 PM , Rating: 4
Intel REALLY needs to aquire Nvidia (assuming it will be OK'd by the SEC).

That's really funny. When was the last time any merger was voted down in a corporatocracy?

The SEC rubberstamps everything, the same way the FCC rubberstamps everything Rupert Murdoch and Clear Channel ask for. They are no longer regulators, they merely collect fees.

Spineless bastards.

RE: Nvidia
By Jeffk464 on 5/30/2011 4:55:22 PM , Rating: 1
Isn't nvidia a foreign company? I don't think they give a flying fart what our congress does.

RE: Nvidia
By OAKside24 on 5/30/2011 5:54:54 PM , Rating: 2
"Nvidia is an American global technology company ... based in Santa Clara, California." ( I wasn't sure, either.

RE: Nvidia
By YashBudini on 5/30/2011 6:41:59 PM , Rating: 3
Global means, "We'll screw any country, anytime, even our home base. We're omnivores, you can bet on it."

RE: Nvidia
By Targon on 5/31/2011 8:43:15 AM , Rating: 2
This is a very different situation than most. There are only TWO players in the CPU industry right now, and there ARE three major players in the graphics industry. No regulator would be stupid enough to allow going from three players down to two in the graphics market, but they would allow going from four to three.

RE: Nvidia
By jabber on 5/30/2011 4:35:17 PM , Rating: 2
Isnt Tegra just a rebadged ARM chip? Cant see that helping Intel.

The fact is 90% of the worlds computing market no longer needs a CPU costing much over $100.

Intel doesnt have much of value in that space. They are slowly becoming irrelevant to most of the market. It's only their mass marketing that keeps them going really.

If AMD started doing a few TV spots it would really put the pressure on.

RE: Nvidia
By Jeffk464 on 5/30/2011 4:57:26 PM , Rating: 2
Yup,I think cpu power will take a backseat to apu power. Most people browse the web, do productivity, and play multimedia.

RE: Nvidia
By SPOOFE on 5/31/2011 1:16:13 AM , Rating: 2
I think we're just going to see a greater and greater divide between "good enough" chips and "powerhouse" chips. We see the early aspects of this trend; I just think "high end" and "low end" are going to wind up with a fairly large gap between them.

RE: Nvidia
By Targon on 5/31/2011 9:02:53 AM , Rating: 2
I disagree here, there will always be people who want something a bit more than the "basic" system, without going up into the very high end. This is why you can buy all these different processors at different speeds, core counts, etc.

Some people will want a six core processor, even now, because they do so many things at once and leave all of their applications open ALL the time. Others will be satisfied with a quad-core processor. EVERYONE will see the advantage of quad-core over dual-core, but not everyone wants to spend a LOT more.

On the AMD front, there will be a fairly small group that won't go with a APU and instead will want a pure CPU with discrete video card, and that is the 10 percent crowd. Now, that is the key to what AMD is doing.

Right now, you have AMD selling CPU and chipsets. If AMD moves the GPU from the chipset to the CPU and adjusts the costs, the CPU+chipset cost may end up being the same, but it also encourages allowing people to upgrade the CPU more often. How many of us have just done a CPU or video card swap without wanting to replace the motherboard? If replacing the CPU also replaced and upgraded the graphics, that could be seen as something more people are willing to do, in the same way that swapping a video card is seen as an acceptable thing today.

There really are four tiers of CPUs out there, the extreme high end, the enthusiast, the mid-range, and then the low end. AMD has pretty much dropped out of the top two tiers on the CPU front due to not having been able to move to a 32nm process soon enough. So, for your mid-range CPU and low end, you have a LOT of room for differences. Dual-core vs. quad, and we will probably see hex-core entering that mid-range in the next year. Do you want a 3.8GHz quad-core, or a 2.8GHz dual-core? We WILL see more of a CPU performance difference between the CPU-only chips/systems and the APU market, and that is where things will get very interesting going forward.

With AMD pushing the APU, the overall performance between AMD and Intel will be a bit more balanced(with AMD having the better graphics and Intel POSSIBLY having the better CPU at the same price). As the complexity of the GPU side goes up, heat and expertise will come into play, and AMD may be able to increase performance a bit faster than Intel. This may hit Intel hard enough to change its approach, and their focus on the CPU may decrease to the point where AMD can stay competitive.

Then again, if Intel is found to be doing more anti-competitive stuff, they may be forced to divide up the company to split off its fab capabilities, which would really help level the playing field. If Intel didn't have the fab development advantage it does, it might not be nearly as far ahead at this point.

RE: Nvidia
By Jeffk464 on 5/30/2011 4:49:12 PM , Rating: 4
AMD has to prove they can get llano to the market first. This is years in the making.

RE: Nvidia
By wiak on 5/30/2011 9:08:21 PM , Rating: 2
amd llano dosnt need to come to market first
it will beat intel anyway mainly due to the ATI discrete derivered graphics in LIano in its price range

and for nvidia, amd fusion has x86, tegra does not

RE: Nvidia
By SPOOFE on 5/30/2011 10:29:24 PM , Rating: 2
it will beat intel anyway mainly due to the ATI discrete derivered graphics in LIano

Yay Llano! Now I'm getting 12 FPS in modern games instead of 9! Weeeee!

RE: Nvidia
By Philippine Mango on 5/30/2011 11:10:52 PM , Rating: 2
I don't think you're giving Fusion graphics enough credit.. They're significantly better than intel graphics and in your scenario it'd be more like 4fps with intel and 16fps with the AMD Fusion.

RE: Nvidia
By encia on 5/30/2011 11:29:57 PM , Rating: 2
400 stream processor Llano runs Aliens Vs Predator DX11 just fine.

RE: Nvidia
By SPOOFE on 5/31/2011 1:23:21 AM , Rating: 2
Does it? I did a quick search and found several sites ASSERTING that it does, but no actual evidence and no detail about settings other than "DX11 features".

RE: Nvidia
By encia on 5/31/2011 8:50:46 AM , Rating: 2
RE: Nvidia
By encia on 5/31/2011 8:53:30 AM , Rating: 2
"Can anyone tell me what MobileMe is supposed to do?... So why the f*** doesn't it do that?" -- Steve Jobs

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki