Global Foundries has 1st customer

223 answers Last reply
More about global foundries customer
  1. considering AMD owns half of global foundries, this is good news! ;p (ok less than half. but still!)
  2. this whole fabless things seems to be like a giant pot for multiple design companies to compete against larger corps like intel. I think they are going to need a lot more than just one partner on board for this to work, hopefully when fab 2 is finished they will have lots of customers.


    wonder if there are any possibilities of issues with this pot approach, like a regular pot approach and one or two players take too much.
  3. i think the move to concentrate on design and leave production to gf was a good move, and its not only for small companies, heard nvidia is interested, lets just hope their 40nm node is better than tsmc.
  4. is there a link to explain this whole ranking thing, i have 331 posts and 674 points, thl above me has 1191 posts and 1160 points, yet we are both apprentice???
  5. They'll probably start 40nm early next year. Apparently ATI is suppose to have a certain part of their GPUs made by Global Foundries as per the agreement with AMD. Just what part that is seems to be anyone's guess right now.
  6. xaira, hijacking a thread is unacceptable.

    start your own.
  7. i dont wanna start a whole thread, i just wanna no if anyone has a link, and no one has to stop talking about GF having their first customer, i never said explain to me on this thread!!!
  8. So 22nm is set for..... 2012? Thats if they can make it on 45nm that long..... Thats 3 years for 45nm. Thats a long time really. At least AMDs yields will be pretty good for 45nm. but skipping 32nm might have some drawbacks.

    So if thats that then Intel will have 28nm in 2011 but hmmm.....

    Ok not sure about this. Will it be 22nm for GPUs only then? Or what? because the process for GPUs generally can't be used for CPUs unless they are both the same style...

    Guess its more confusing than it looks.
  9. Well... AMD won't be competitive until 2012 then, right?
  10. well if nvidia, ati, DAS storage controller ppl (adapetec?) that needs high performance, and others are willing to pool their money in, from the high end to low end, then this would work.

    /** pure BS/thinking on my part**/
    Intel has to continuously phase out or sell off their old stuff when it no longer suits them, whereas other fabs has more flexibility. For example, old 90nm nodes? no problem, here are some high performance networking designs/chip for some of that talking doll crap? that could be used and put the 90nm nodes to use.
    /** pure BS/thinking on my part**/

    thats how I think about this, although someone else correct me if I'm wrong.

    also, ranking atm is pure fudgy imo, I stop worrying about it until some noob call me the noob with a apprentice ranking lol
  11. I hate the new rankings :/
  12. amdfangirl said:
    I hate the new rankings :/


    I'm not fond of it myself.

    Hey thanks Badtrip for this interesting read. I am glad GF is here, it should allow smaller companies to thrive and that will drive competition in all markets.
  13. Having ATI is more than AMD ever did with their fabs doing their cpus alone, besides anyone else
    Thats an instant profit for AMD, as part of those monies spent rolls back to them thru GF, the only 1 missing out is TSMC.
    Tried splainin this awhile back, almost got run out heheh
  14. Yep, it's just free money for AMD. At first I thought it was bad that they had to start up GloFo but in the end it's gonna be very apparent that they did the right thing, the only thing that made sense.

    I actually think Nvidia will end up working with GloFo too, because history has shown that it's never good to have all your eggs in one basket.
  15. o man TSMC is gona be pissed, especially if GF flys under the flag of the most innovative company bringing out the best CPU or something, once AMD gets their performance crown back lol
  16. TSMC haven't actually lost anything, because ST Micros made their own chips up till now. It's kind of like an emerging market, the businesses who could happily make their own chips at 65nm or above simply cannot afford the newer tech.

    AMD couldn't afford to go lower than 45nm, that's why Global Foundries was created. Every other company that reaches that stage (and we're talking $billions in development costs and fabs now) will be forced to go with either GF or TSMC. It's a big growth area for them both.
  17. From Overclockers:

    Quote:
    Analysis: Globalfoundries needs to grow

    TSMC Chairman Predicts Tough Victory Over GlobalFoundries

    AMD's spinoff seeks non-AMD customers - no surprise here, although I think the deep-pockets fueled by oil money will keep Global Foundries going when others might fold. The more GF wins new business, the less AMD will get preferential treatment - could GF ultimately bite the hand that feeds it?


    There have been other reports stating GF is trying to distance itself from AMD, esp. as AMD's share drops to the minimum 25% ownership required by the x86 license.
  18. GF can distance themselves all they want, AMD is always going to own 25% of it. In the end, the more successful GF is, the more money AMD will make. It's win-win, there is no way AMD can lose out of this.
  19. jennyh said:
    GF can distance themselves all they want, AMD is always going to own 25% of it. In the end, the more successful GF is, the more money AMD will make. It's win-win, there is no way AMD can lose out of this.


    This another one of your famous "guarantees"?? :D

    AMD currently has no spare cash to pump into GF, so as GF spends more $$ to upgrade & build, eventually they will have to spend more than the remaining ~10% ownership margin left for AMD (currently around 35% IIRC). I guess AMD could try to borrow the $$ somewhere (pretty hard to do when your credit rating is somewhere between really bad & abysmal), but the idea was for AMD to shift the fab costs to somebody else.

    AMD's last & best hope would be to renegotiate the x86 license, PDQ, to get rid of the 25% ownership clause...
  20. We don't know the exact details of the deal that was struck though do we. I mean, GF have already started on 22nm development, so either AMD paid their share on that or a deal was made when the company was created.

    Assuming GF get to 22nm first, AMD will be on 22nm before intel and that will bring in more cash than they've ever had, which in turn will help GF to continue with smaller and smaller fabs. Both companies are forever linked to each others success now, GF have made it clear that they are out to dominate chip production and they sure as hell have the money behind them.

    You can be sure that AMD have no intention of ever giving up the 25% minimum. Why would they? The more money GF makes, the more AMD make. With GF ahead on nanometer tech, AMD will be ahead on cpu power which means more people will want their cpu's instead of intels = more money.
  21. OK, besides rumor, who can bring more business to GF than AMD? And, thats including ATI as well, and any chipset business thats still going?
  22. Sure, theyll dunp their largest by far business oartner/buying customer for lessor trails, uh huh, and I just saw Baron wearing a i7 T shirt too
  23. jennyh said:
    We don't know the exact details of the deal that was struck though do we. I mean, GF have already started on 22nm development, so either AMD paid their share on that or a deal was made when the company was created.

    Assuming GF get to 22nm first, AMD will be on 22nm before intel and that will bring in more cash than they've ever had, which in turn will help GF to continue with smaller and smaller fabs. Both companies are forever linked to each others success now, GF have made it clear that they are out to dominate chip production and they sure as hell have the money behind them.

    You can be sure that AMD have no intention of ever giving up the 25% minimum. Why would they? The more money GF makes, the more AMD make. With GF ahead on nanometer tech, AMD will be ahead on cpu power which means more people will want their cpu's instead of intels = more money.


    We do know that ATIC (Abu Dhabi) forced AMD to renegotiate the deal last December, after AMD's stock value plummeted, leaving them with just 35% ownership instead of the original 50/50. Nominally AMD retains 50% of the voting rights, but I'm pretty sure they tend to vote however ATIC tells them to. The 50% is just subterfuge to get around the x86 license requirements.

    The trouble with your prognosis is that you forget Intel is also working hard to stay 1 yr ahead in process. Intel learned its lesson with Prescott - they don't stand still waiting for the competition to catch up or surpass them. They had a 4-quarter lead on 45nm, will likely maintain that on 32nm and unless some disaster occurs, I'd bet on them staying in the lead at 22nm.

    AMD is not in the driver's seat when it comes to anything external. Financial necessity forces them to be reactive, not proactive. Maybe AMD would like to retain some ownership in GF, maybe not. What I'm fairly confident about is that, should GF need a cash infusion, they won't be getting it from AMD. Ya can't squeeze blood from a dried-up turnip, as the saying goes :D.
  24. JAYDEEJOHN said:
    OK, besides rumor, who can bring more business to GF than AMD? And, thats including ATI as well, and any chipset business thats still going?


    Heh, well there's a distinction between cash business and credit, so I'd imagine GF would like to line up some customers who can actually fork over some bux to pay for the fabbed product, instead of IOU's :D. Besides, I imagine GF would like to diversify their customer base, just in case their biggest customer winds up BK. :sol:
  25. What happens if GF get to 22nm first? 22nm AMD/ATI fusion vs 32nm intel fusion, thats almost unthinkable for intel but it's what AMD and GF are aiming at.
  26. JAYDEEJOHN said:
    Sure, theyll dunp their largest by far business oartner/buying customer for lessor trails, uh huh, and I just saw Baron wearing a i7 T shirt too


    Well that would be an improvement over the silver leisure suit he was last seen wearing :D.

    We'll see how GF does wooing nVidia - if unsuccessful, you can probably blame the ATI div. of AMD. Or else Jen-Hsun accidently opened that can 'o whoopas on himself [:thegreatgrapeape]
  27. Quote:
    Intel ain't getting overtaken anytime soon in process terms I wouldn't imagine.


    If they are, I'll get AFG to design me a new hairdoo! :whistle:
  28. You be surprised.

    http://www.custompc.co.uk/news/604668/ibm-and-amd-create-first-22nm-sram-cell.html

    Nothing from intel as yet. As a comparison, Intel demonstrated a 32nm cpu in September 2007. It seems much more likely that the IBM alliance is ahead on 22nm.
  29. Per Intel's externally published plans, they will ship 22nm CPUs in 2011. If GloFo starts manufacturing 22nm in 2012 as discussed in the article, they will still be a year behind.

    Unless you are privy to special insider information from both companies, I'd not count on GloFo making it to 22nm shipped CPUs before Intel.
  30. Sure intel put it on their roadmap, but where is the demo chip? Intel are not shy about bragging over being first, they were happy to do it with 32nm (IBM incidentally showed theirs 1 month later).

    I'm sure they have a 22nm chip somewhere, but if they had it before August last year we'd have heard about it. Intel don't miss opportunities to tell us how amazing they are so where is that working 22nm chip?
  31. Quote:
    I'm pretty sure intel already demonstrated sub 32nm silicon not too long ago.


    The point here is, they weren't first. IBM and AMD were the first to show actual working 22nm, not intel.

    There are 2 possibilities, one says intel were beaten to working 22nm, the other says they were first but forgot to tell us about it. I know what I'm more likely to believe.
  32. Intel don't consider 5-10 SRAM cells on a die newsworthy. They typically wait until they're able to replicate millions before they call it "soup".

    When to expect it? Don't know, and wouldn't tell you if I did. I like my job.
  33. I see GF being the biggest GPU fab in the near future. They have the built in business of ATI and you can bet Nvidia would love to find someone who can actually manufacturer what they want without the 30% or less yields on the 40nm process. Should be interesting.
  34. Problem here is, Intel is facing a tougher competition than theyve ever faced. Thats fact. The GF partners could do a buy off of Intel. Thats fact, without intecession. So, at this point, anything is more possible coming from them, whereas Intel has many internal solutions to solve, as its their products that need to hit their tick tock, and AMD no longer has to worry about that
  35. As an example, wheres the 32nm igp? And why isnt everything 1 die? Why a 45nm igp? Thats wasnt on their roadmap
  36. That's always been on the roadmap. The only thing that changed was cancellation of the 45nm CPU+IGP to pull in the 32nm+45nm version.
  37. So, wheres the 32? It makes more sense doesnt it? Like by alot?. They arent cuttin the mustard accross the board, thats my point. Is it easier using the 45nm than a 32nm igp? Will the power be the same? The costs the same?
    All I hear about is Intels presicion, but this example throws it a bit off. Im sure theres alot more, and especially as we head into these smaller nodes.
    I see it as once again, Intel leaving their video/rendering processing behind, and putting their cpus first. They can no longer aford to do that. Itll be 1 chip, but this is a test, and first out I can understand, but it still shows they put the igp last
  38. I have lots of comments, but all are NDA. You can consider yourself the winner of the argument, if you like. :)
  39. It isnt that. I understand. I also understand its not worth it for Intel to switch to 32 in their igps, but, you have to admit, itd be nice.
    Of course theyll have to soon anyways. Its just that, if somethings left out....
    Dont mean torattle your cage. You above most must know my allegiances are aligned more towards the gpu side of things, and wed all benefit from Intel moving in this direction, which again, I know they are, but having it done this way, its hard to seperate from their past history
  40. lol, this epeen race has a winner!

    Seriously, declaring NDA in a anonymous forum topic so it looks like you are some badass insider is just a joke. Guess we'll see if you are a troll or not in the coming year.
  41. It's not comments we're looking for it's proof of intels 22nm. We have IBM's and AMD's proof, we have comments on IBM's process. So far what do we have from intel except their roadmap?

    You realise that if intel hadn't released that roadmap, everyone would be asking 'why no 22nm in 2011?'. They *had* to release that roadmap even if they don't have the capability.

    In order to stick to that roadmap, and assuming we see 32nm by Q4 this year, intel had better be demonstrating working 22nm silicon before this September is finished.
  42. He knows... a few things.
  43. Is that the traditional timing for an announcement for a new node is Sept?
  44. I'm just going on what happened with the 32nm demo. Sept 07 - Q4 09 release would be the same as Sept 09 - Q4 11 release for 22nm.

    Unless, ofc, intel are throwing an awful lot more cash at moving it forward this time. They are already having to throw more cash at it just in terms of research remember. It's no coincidence we're down to talking about IBM and intel now, really who else could afford it? Last I checked, IBM are a lot bigger than intel are and there can only be one winner of a money fight there.
  45. JAYDEEJOHN said:
    He knows... a few things.


    Wait, wait, is he an actual insider on this? Could this mean that intel is going to open up another can of hurt real soon?
  46. He already says he doesn't know when to expect it so working in intel's bribery department probably doesn't count as being an insider in this. :D

    Going on past history, if intel are to stick to that 2011 roadmap they should be showing us a working 22nm cpu very soon. Not just a few SRAM cells, a working 22nm cpu. If they do that, they have sure been quieter than usual about it, and kudos to them. Until that point, for all we know intel cannot get their 22nm to work, not even with a few SRAM cells.
  47. No, their plans arent that devastating. Its where theyre doing it. Introducing low powered, HT enabled dual cores with a igp thats maybe comparable for now to the competition, in the average Joe price range...
  48. yeah imagine a few years from now, Joe Random comes in,

    I can't play Crysis 2 Teh way it's meant to be played on Intel edition on my intel box! What do I do? I hear you need a better GPU. What is a GPU?

    12 pages of pain and suffering about what is a discreet GPU and why it costs that much later.

    someone says, ah well Intel's new drop-in processor has a better IGP, and can run the said game, just go buy the newest intel processor and do a drop in replacement for around 100 dollars and he gets to play the game.


    I think that would be what Intel wants lol. a IGP upgrade that profits them and not mobo maunf and makes people associate intel name with performance with games and other main stream stuff again.
  49. jennyh said:
    You be surprised.

    http://www.custompc.co.uk/news/604668/ibm-and-amd-create-first-22nm-sram-cell.html

    Nothing from intel as yet. As a comparison, Intel demonstrated a 32nm cpu in September 2007. It seems much more likely that the IBM alliance is ahead on 22nm.


    Since you like CD, Fuad, etc here's a little article from the latter 6 months ago :D:

    Quote:
    Intel working hard on 22nm CPUs
    Written by Fuad Abazovic
    Monday, 02 February 2009 10:28

    Getting ready for the future

    Intel engineers are currently designing the entire 22nm process generation of CPUs and even if this sounds like the very distant future, we can expect the first 22nm CPUs probably at some point in 2011. This is of course the current plan that Intel can achieve if it doesn’t end up with some major issues that can delay things.

    In 2009, Intel should be launching Westmere, a 32nm CPU, which can be seen as a refined Nehalem derivative. Even so, in 2010 Intel plans to launch a new 32nm architecture codenamed Sandy Bridge.

    In 2011, the 32nm Sandy Bridge will get its 22nm "tick" revision rumored to be known as Ivy Bridge, but at this point the name remains unconfirmed. In 2012, Intel should have a brand new 22nm CPU architecture. We have confirmed that Intel engineers are already working on 22nm chips and as you can imagine, they should be loaded with more cache, more cores, and should end up faster than anything we have today.
Ask a new question

Read More

CPUs Product