1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Nvidia GT300 Yields under 3%

Discussion in 'NVIDIA' started by Asylum, Sep 15, 2009.

  1. Asylum

    Asylum

    Joined:
    Jul 18, 2008
    Messages:
    2,806 (1.23/day)
    Thanks Received:
    679
    Location:
    South Carolina
    Has any one seen this article?
    What do you think. True or marketing gimick from ATI?



    Nvidia GT300 yields are under 2%
    Just pathetic
    by Charlie Demerjian
    September 15, 2009


    THE SAGA of Nvidia's GT300 chip is a sad one that just took a turn for the painful when we heard about first silicon yields. Nvidia's execution has gone from bad to absent with low single digit yields.

    A few weeks ago, we said that Nvidia was expecting first silicon back at the end of the week, the exact date was supposed to be Friday the 4th plus or minus a bit. The first bit of external evidence we saw that it happened was on the Northwood blog (translated here) and it was a day early, so props to NV for that. That lined up exactly with what we are told, but the number of good parts was off.

    The translation, as we read it, says there were nine good samples that came back from TSMC from the first hot lot. That is below what several experts told us to expect, but in the ballpark. When we dug further, we got similar numbers, but they were so abysmal that we didn't believe it. Further digging confirmed the numbers again and again.

    Before we go there though, lets talk about what a good die is in this case. When you get first silicon back, it almost always has bugs and problems. First silicon is meant to find those bugs and problems, so they can be fixed in succeeding steppings.

    By 'good', we mean chips that have no process induced errors, and function as the engineers hoped they would. In other words not bug free, but no more errors than there were in the design. 'Good' in this sense might never power on, just that the things that came out of the oven were what was expected, no more, no less.

    Several experts in semiconductor engineering, some who have overseen similar chips, were asked a couple of loaded questions: What is good yield for first silicon? What is good yield for a complex chip on a relatively new process? The answers ranged from a high of 50% to a low of 20% with a bunch of others clustered in the 30% range. Let's just call it one-third, plus or minus some.

    The first hot lot of GT300s have 104 die candidates per wafer, with four wafers in the pod Nvidia got back a week and a half ago. There is another pod of four due back any day now, and that's it for the hot lots.

    How many worked out of the (4 x 104) 416 candidates? Try 7. Yes, Northwood was hopelessly optimistic - Nvidia got only 7 chips back. Let me repeat that, out of 416 tries, it got 7 'good' chips back from the fab. Oh how it must yearn for the low estimate of 20%, talk about botched execution. To save you from having to find a calculator, that is (7 / 416 = .01682), rounded up, 1.7% yield.

    Nvidia couldn't even hit 2%, an order of magnitude worse than the most pessimistic estimate. Ouch. No, just sad. So sad that Nvidia doesn't deserve mocking, things have gone from funny to pathetic.

    At this point, unless there's a massive gain in yields on the second hot lot, there might not be enough chips to do a proper bring up and debug. This stunningly bad yield could delay the introduction of the chip, adding to the current pain and bleak roadmap. If there aren't enough 'good' parts from the second hot lot, that might require running another set, adding weeks to the total. Q1? Maybe not.

    It is going to be very interesting to see what Nvidia shows off at 'Not Nvision' in a couple of weeks. Will it give the parts to the engineers to work on, or show them off as a PR stunt? We will know soon enough. In any case, the yields as they stand are sub-2%, and the status of the GT300 is far worse than we had ever imagined.S|A
     
  2. phanbuey

    phanbuey

    Joined:
    Nov 13, 2007
    Messages:
    5,205 (2.05/day)
    Thanks Received:
    975
    Location:
    Miami
    yeah I figured something like this would happen... Alot of the early speculation turned out to be true then. I think its true. NV is certainly moving as fast as it can, and yet there is nothing but silence from their end.

    Too bad, it means the graphics wars could come to an end, and we will be seeing $600-$800 cards again, but this time from ATI. bleh.

    Hopefully they get their @$$ in gear and show up to play.
     
  3. [Ion]

    [Ion] WCG Team Assistant

    Joined:
    Sep 15, 2009
    Messages:
    11,813 (6.35/day)
    Thanks Received:
    10,951
    Location:
    North Carolina, United States
    I won't believe it until I read it elsewhere, Charlie's articles are usually complete BS
     
    tkpenalty says thanks.
    Crunching for Team TPU
  4. Paulieg

    Paulieg The Mad Moderator Staff Member

    Joined:
    Feb 19, 2007
    Messages:
    11,914 (4.25/day)
    Thanks Received:
    2,979
    Location:
    Wherever I can find the iron.
    You believe this, really? C'mon, on just one article? I think experience has shown that we shouldn't believe any of this crap this early in the game.
     
    tkpenalty, trt740 and phanbuey say thanks.
  5. pantherx12

    pantherx12 New Member

    Joined:
    Jan 2, 2009
    Messages:
    9,714 (4.59/day)
    Thanks Received:
    1,699
    Location:
    ENGLAND-LAND-LAND
    It's just to bias for me to believe it.
     
  6. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    39,783 (13.20/day)
    Thanks Received:
    14,163
    I don't believe or disbelieve this story. It's quite obvious there are bad yeilds since we've seen nothing from Nvidia. 3% does sound a bit unbelievable though. I hope things get better.
     
  7. InTeL-iNsIdE Guest

    the same Charlie Demerjian who used to blog at the Inq and now blogs over at www.semiaccurate.com ? thos 2 websites say it all :wtf:
     
  8. SneakySnake

    SneakySnake New Member

    Joined:
    Sep 15, 2009
    Messages:
    15 (0.01/day)
    Thanks Received:
    2
    Location:
    Waterloo, Ontario, Canada
    Well Nvidia could do the same thing that ATI did in the 4xxx/GT200 battle. Nvidia's 260 and 280 were dominating until the 4850 and 70 came out of nowhere. People thought that ATI was on their knees on the last GPU battle
     
  9. El Fiendo

    El Fiendo

    Joined:
    Aug 22, 2008
    Messages:
    2,305 (1.02/day)
    Thanks Received:
    1,164
    Location:
    Edmonton, Alberta
    Again, Charlie Demerjian. You can trust anything he says about nVidia about as far as I can fly with my arms. Also, note that the only places reporting it are simply quoting him. No one else mentions anything about this that isn't using his words.

    In fact, what's more odd is that the inquirer, the website Charlie posts his articles, doesn't have this as an article.

    Can you give us a source OP?
     
    phanbuey says thanks.
    Crunching for Team TPU
  10. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    39,783 (13.20/day)
    Thanks Received:
    14,163
    www.semiaccurate.com
     
  11. El Fiendo

    El Fiendo

    Joined:
    Aug 22, 2008
    Messages:
    2,305 (1.02/day)
    Thanks Received:
    1,164
    Location:
    Edmonton, Alberta
    Ok, does he usually post for them? Because I always thought he was an Inquirer journalist. The Inquirer has nothing on this.
     
    Crunching for Team TPU
  12. phanbuey

    phanbuey

    Joined:
    Nov 13, 2007
    Messages:
    5,205 (2.05/day)
    Thanks Received:
    975
    Location:
    Miami
    I just noticed its from the same guy that has been beating his own "nvidia can't make 40nm" drum...

    so that definitely makes it less believable.

    That being said, this article *could* be true, and there is some circumstantial evidence that it is... but for right now its a bunch of FUD from a super-biased anti-nvidia blogger.
     
  13. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    39,783 (13.20/day)
    Thanks Received:
    14,163
    He is biased, but not necessarily incorrect. I still haven't seen anything on any Nvidia 40nm parts. To put it more truthfully it should be said that TMSC can't make Nvidia 40nm GPU's.
     
  14. hat

    hat Maximum Overclocker

    Joined:
    Nov 20, 2006
    Messages:
    16,932 (5.86/day)
    Thanks Received:
    2,066
    Location:
    Ohio
    Yields under 3%? That's highly unbelieveable.
     
    Crunching for Team TPU
  15. El Fiendo

    El Fiendo

    Joined:
    Aug 22, 2008
    Messages:
    2,305 (1.02/day)
    Thanks Received:
    1,164
    Location:
    Edmonton, Alberta
    Just noticed its his website. Guess that would explain it, lol. At any rate, its Charlie Demerjian. This man spends his time bashing the green company. He's notorious for it. When I see more independent (not to mention less biased) journalists report this, I'll believe it.
     
    Crunching for Team TPU
  16. SNiiPE_DoGG New Member

    Joined:
    Apr 2, 2009
    Messages:
    582 (0.29/day)
    Thanks Received:
    135
    in this article he is more specific than he ususally is ..... but at the same time I think he is making a bunch of crap up.....
     
  17. HalfAHertz

    HalfAHertz

    Joined:
    May 4, 2009
    Messages:
    1,892 (0.95/day)
    Thanks Received:
    380
    Location:
    Singapore
    That's really hard to believe. He states, that TSMC produces 104 dies out of 1 waffer. If I'm not mistaken, they used 300mm waffers. Dunno if my mad math skillz are correct but that would leave us with a total area of ~7050mm^2 per waffer and a die area of over 650mm^2. That's supposedly twice the die area of the HD5870 and highly unlikely case...
     
  18. Jstn7477

    Jstn7477

    Joined:
    Aug 30, 2009
    Messages:
    3,878 (2.07/day)
    Thanks Received:
    1,560
    Location:
    Sarasota, Florida, USA
    Charlie quit working for the INQ a few months ago. He's now part of Semiaccurate.com. Since Charlie has always bashed NVIDIA, you probably shouldn't rely on the contents of this article unless there are other credible sources that match his information.
     
    Crunching for Team TPU More than 25k PPD
  19. W1zzard

    W1zzard Administrator Staff Member

    Joined:
    May 14, 2004
    Messages:
    14,944 (3.92/day)
    Thanks Received:
    11,734
    any time you read the words "yield" and "%" together in an article on the internet, click the X, no matter the source. Yes, not even press releases from AMD, Intel, NVIDIA, TSMC or God himself. Besides, that number has no meaning for you.

    If you buy some coke do you ask how many plants died because the DEA burnt them down? No, you ask how much it costs and what it does for you (and how you can get more of it)
     
    Last edited: Sep 15, 2009
  20. newtekie1

    newtekie1 Semi-Retired Folder

    Joined:
    Nov 22, 2005
    Messages:
    20,027 (6.15/day)
    Thanks Received:
    6,091
    1.) I doubt this article has any truth to it after seeing the author.
    2.) Sounds more like a fab issue than nVidia's.
    3.) TSMC is struggling with 40nm, that is no suprise.
    4.) What were the yields on RV740? I'm guess just as bad, though I can't find any hard numbers on it. Even after it went retail, TSMC continued to have yeild issues. To the point that rumors were going around that ATi was going to discontinue the HD4770 due to terrible yields.
     
    Crunching for Team TPU 50 Million points folded for TPU
  21. Flyordie

    Flyordie New Member

    Joined:
    Oct 26, 2008
    Messages:
    1,870 (0.86/day)
    Thanks Received:
    247

    As much as you all may rip Charlie.. he was right on the defective GPUs from Nvidia. That was a big mess... He gave numbers and normally he doesn't give numbers unless he had them confirmed by another source. He usually uses words like "very poor yields" or "mostly defective" if they were actually unconfirmed.

    Im not saying take this story to heart, just don't dis-credit it yet. GT300 is a VERY VERY big GPU... Just like the GTX 280, it will have crap yields at the beginning and work its way better. What he may be saying is out of the full dies, only 7 were fully functional. Disable some of the shaders or w/e and you never know, those might be usable dies for something else...
     
  22. AsRock

    AsRock TPU addict

    Joined:
    Jun 23, 2007
    Messages:
    10,978 (4.10/day)
    Thanks Received:
    1,710
    Location:
    US
    "cough" B*llsh*t
     
  23. jagd

    Joined:
    Jul 10, 2009
    Messages:
    457 (0.24/day)
    Thanks Received:
    89
    Location:
    TR
    Everyone is correct and what it is said by charlie until today .TSMC has problems ,nvidia no experience with 40nm and their parts will be at market at december or Q1 2010 if everything go smooth .Now telling yields are very poor nothing surprise. Ati used rv740 a pipe cleaner for 40nm production but it was a experimental gpu not a flagship gpu ,if nvdia will have same problems with rv740 they have big problems at their hand.
    Time will tell truth but remember 2% is for first silikon and not final silikon if nvdia needs tweak it will cost time to them ,for some info http://www.semiaccurate.com/2009/08/18/nvidia-takes-huge-risk/
     
  24. tigger

    tigger I'm the only one

    Joined:
    Mar 20, 2006
    Messages:
    10,183 (3.25/day)
    Thanks Received:
    1,399
  25. hat

    hat Maximum Overclocker

    Joined:
    Nov 20, 2006
    Messages:
    16,932 (5.86/day)
    Thanks Received:
    2,066
    Location:
    Ohio
    What about techpowerup? :nutkick:
     
    Crunching for Team TPU

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page