"Global Foundries will bash up Intel's process lead" says AMD
 

http://tgdaily.com/images/stories/45...dries/atic.jpg

"The chief technology officer of Advanced Micro Devices (AMD) has claimed that the spun-off division that makes its microprocessors will outpace Intel in the foreseeable f...."

http://www.tgdaily.com/business-and-...s-process-lead

Does this mean Bulldozer will come with 22nm instead of 32nm?? Amm probably not but who knows.. :p:

Seems like a lot of hot air, but if it can help AMD play some catchup with Intel over the next few years, it seems like a good deal to me. We need innovation like Conroe coming from AMD, and while a process shrink might not put them there, I would think it would help.

Anything is possible but it seems a little funny to see the executives of the tech companies huffing and puffing so strongly as of late, like they've had a redbull party.

Well I hope they can catch up. GF is getting some big customers on board so maybe this statement isn't being overly optimistic. I just hope they can get both Nvidia and ATI as accounts so new GPUs won't have shortages and won't be delayed months. TSMC is just a pile of :banana::banana::banana::banana: right now.

I won't believe it until I see a screenshot, until then this is just another puff of hot air.

People are being very liberal with what was actually said. heh

Yet people are trying to spin it into:

Hmmm i might see them bringing smaller process before intel to the mainstram market, but i doubt they will beat them in performance. If they want a broad spectrum of consumers, they need to have a flexible process to cover most needs. And a wide spread process that covers many needs never can beat a specialized process that only aimed at "doing one thing".

TSMC is potentially going to lose a lot of clients to GF, etc. But i don't really see GF competing with Intel's process lead. Sorry.

Why not?

The arabs are REALLY going to invest a LOT of money on the fabs?
It's not easy nor cheap to beat Intel at process tech.

they spend more money on their process and take pride in it. their egos are too big not to have the best process. they will probably launch before years end so they can keep up with moores law and stick it another slide show.
http://www.youtube.com/watch?v=R4XyL...layer_embedded

While I agree in principle, the Arabs have enough OIL money to bury Intel. It's definitely going to be interesting the next few years.

I also wonder about their GPU division. Will GF be able to meet expectations for 32nm?

While i agree they have a lot of money, they are investors, not money throwers.

Although i think that the moment amd gets it act together and catch up some market share, say 30-35%, Intel is going to be much more strained to be competitive, because they will have a bigger manufacturing burden than amd & global foundries.

Maybe not in the inmediate future, but think about 16nm and below, were the cost of putting up an state of the art fab is going to be increased ten-fold compared to say, 45nm; in this case global foundries will have quite some years in the market and quite some customers, they will have much more flexibility and possibility of putting up a factory like that and getting away with the fixed costs of maintaining a fab like that.

Ironically, JohnJohn posted right above me because I was going to say I was having a deja vu courtesy of a certain can of whoop-a s s.

TMSC already manufactures in 40nm, that is actually not really better than Intel's 45nm, but is an smaller number. It is planing to jump to 28nm soon, definitely before intel's 22nm. GlobalFoundries can't stay behind TMSC, and that means that they will be ahead of intel in some way on the 28nm.

the problem is that fabrication is the one of the least profitable business there is. back when the newest fabrication processes doubled clockspeed and were cheap to make it was a different story. intel invested $7 billion for its 32nm fab in arizona and their 45nm fab was 4 billion. thats going to be a horrible return on investment if you can only use it for 2 years.

Intel a year ago had roughly 79% of the market share a year ago, so I HIGHLY doubt they're going to lose another 9-14% because of their FAB not being able to turn out a smaller process. Consumers don't give a rats a$$ about die size.

edit: consumers = average Joe.

http://247wallst.com/2009/06/10/amd-...ntel-amd-intc/


I doubt you'll see any "large gain" in market share by AMD unless they roll out a new chip architecture that gives them a performance boost above Intel like we saw with Intel's release of the Core 2 Duo.

Until then AMD will remain as "a cost effective alternative" to buying an Intel CPU.

In other words.... I could go out and buy a Japanese or American car with great performance... but I'd rather pay a little more for a nice German car which has performance, styling, AND features. Oh yeah.... higher brand name recognition too. :toast:

Intel makes a lot more than desktop and server (and mobile) CPU's my friend.

They're not gonna spend tens of billions of dollars in a short period of time in order to simply have a better manufacturing process than Intel. Its really quite unnecessary. I know they have something like a Trillion dollars, thats great. But they have many various investment plans. And GF is definitely one of their important investments, but they also have to make sure they dont burn bridges within the western business society. I mean, really... they could buy up many technology companies, but at what cost? and for what purpose?

What purpose would it serve for ATIC to invest that much money for such a goal? Their goal is to create a world class fab, they are taking the slow and steady approach. Once they get more customers, then they can begin to invest more in upgrading the processes and frankly, it doesn't make sense right now to be throwing down a lot of money in this kind of an economic environment when companies may still be reluctant.

Well, when you think about it, both firms are shipping CPU's at 45nm. Not that much farther to go till they hit the limit. So, kinda too late?

You can use it for a lot more then 2 years. Not every customer needs the top process. The older process machinery could keep cranking out chips until there is no more demand for that process. They might upgrade their process every two years, but the machinery gets a lot more use then that. Even Intel doesn't build every chip it makes on the top process.

If you follow Intel's cadence, the leading edge process technology supplies processors (CPUs) for ~ 2 years, until the new one comes out. Chipsets/embedded and other stuff gets done on node - 1.

I think the x58 is still actually made on 65 nm technology, for example.

Overall, I think it makes sense, as building out capacity for a new technology depreciates most of it's costs up front, after a few years, most of the depreciation has hit, and it is much cheaper hence the lower priced chips get put on the least costly technology.

EDIT: Yes, the x58 is on 65 nm ...

http://www.insidehw.com/Reviews/CPU/...el-DX58SO.html

arab money = oil money .. i doubt intel has the same kind of spending power as someone financed by everyone on earth ;)

ATIC is not the whole oil industry.

And JohnJohn already explained it nicely: