Page 39 of 46 FirstFirst ... 293738394041 ... LastLast
Results 381 to 390 of 460

Thread: Entire GCN Lineup ? HD 7000 series Specs and Price

  1. #381
    Senior Member
    Join Date
    Feb 2010
    Location
    Planet Earth, Solar System, Milky Way
    Posts
    3,465
    Quote Originally Posted by distinctively View Post
    Do you mean the next gen 8xxx series?
    Yes, I mean that 8000 series, if they launch in February- March 2013, it should be much more than a simple respin given that at that point in time 20 nm will not be ready to launch in 5-6 months.

    Add 14-15 months (like you claim between generations) and that would put 9000 series somewhere around mid 2014. Almost ready for 20 nm. Yes, it is possible that 9000 series will be on 28 nm but I find it highly unlikely given that you will see almost negligible performancs improvements between 7000>8000>9000.
    So, they should increase the 14-15 months gap if 20 nm is not ready and launch 9000 on 20 nm in late 2014 as the worst possible case.

    Quote Originally Posted by esrever View Post
    TSMC is set to have 20nm by the Q3 next year and 14nm by end of 2014
    This is a very good wish but it is so overly optimistic. Yes, they have to given where Intel will be then but...
    One year and one quarter between 20 nm and 14 nm sounds impossible.

  2. #382
    Quote Originally Posted by distinctively View Post
    Tough call there. It could happen but there was 15 months between 4870 and 5870. If we are guessing right, there's only going to be 9 or 10 months between the newest generations.
    Perhaps, but we don't really know that for sure. Plus, RV790 was probably a bigger effort than "Tahiti.B" would be, since it's mostly just applying know-how from Pitcairn to an existing design, unlike RV790 which added a decoupling capacitor ring and all that.

    Finally, I think the reason why the HD 7990 hasn't been released yet is that AMD was waiting for Tahiti.B, which makes sense because you need really power-efficient chips for dual-GPU solutions.

  3. #383
    Senior Member
    Join Date
    Feb 2010
    Location
    Planet Earth, Solar System, Milky Way
    Posts
    3,465
    Tahiti.B That is only one revision later, like A2 to A3 kind of stuff... isn't it? Do they need a B type of revision?

    Don't forget that in the case of 4890 you also had nvidia releasing GTX 285 after GTX 280 and also GTX 275 as a direct counter-measure against 4890.
    Now, I don't see where this GTX 685 would be...

  4. #384
    Quote Originally Posted by GeorgiD View Post
    Tahiti.B That is only one revision later, like A2 to A3 kind of stuff... isn't it? Do they need a B type of revision?

    Don't forget that in the case of 4890 you also had nvidia releasing GTX 285 after GTX 280 and also GTX 275 as a direct counter-measure against 4890.
    Now, I don't see where this GTX 685 would be...
    Tahiti.B is just what I'm calling it, it doesn't have an official name. In fact its very existence isn't official.

    If AMD wants to improve speed and power (most notably leakage) to get to Pitcairn's level, yes, they need a silicon respin, i.e. a B0 stepping.

  5. #385
    Senior Member
    Join Date
    Sep 2009
    Posts
    3,021
    Quote Originally Posted by Guild View Post
    Sea Islands also shows up as 2013 in the roadmap, so it's not going to release in 2012.
    I stand corrected. I had really just assumed, incorrectly, that 8xxx would launch for Christmas season which is the usual time to aim for.
    When people tell me "You're going to regret that in the morning" I sleep in till noon because I'm a problem solver!

  6. #386
    Senior Member
    Join Date
    Sep 2009
    Posts
    3,021
    Quote Originally Posted by GeorgiD View Post
    Tahiti.B That is only one revision later, like A2 to A3 kind of stuff... isn't it? Do they need a B type of revision?

    Don't forget that in the case of 4890 you also had nvidia releasing GTX 285 after GTX 280 and also GTX 275 as a direct counter-measure against 4890.
    Now, I don't see where this GTX 685 would be...
    The 4890 wasn't really a re-spin as much as it was a small improvement over one particular area of the chip. I really don't remember the details other than there was a reduction in interference allowing better clock speeds. The 285 was an optical shrink over the 280.

    I have no idea if AMD has the ability to make an improvement on Tahiti like that which was done with the 4870>>>4890. Tweaking with more spins probably won't create much better clock speeds. As Charlie said, AMD could up the bin for an faster 7970 but that would just take that market away from their OEMs. I'd tend to think that AMD may encourage OEMs to provide that boost.

    Another thing to remember is that Nvidia stated that they would have 28nm shortages for the rest of the year. Tahiti's sales are not really going to be hindered by the 680 very much because of the shortages. It may not be cost effective to make many changes to Tahiti. The only real downfall here for Tahiti is the image of the performance crown.
    When people tell me "You're going to regret that in the morning" I sleep in till noon because I'm a problem solver!

  7. #387
    >intel 4004
    Join Date
    Jun 2009
    Posts
    11,281
    Tahiti doesn't need a B revision, or a minor revision.
    NVIDIA is not about shortages, no one yelled about shortages or yield issues for TSMC 28HP except NVIDIA.

  8. #388
    Administrator
    Join Date
    Feb 2009
    Posts
    7,533
    Quote Originally Posted by Alexko View Post
    AMD has all but officially confirmed that they were making a silicon respin of Tahiti, so a new card would make sense, just like the HD 4890 for RV790.
    If they respin it, it will be for die area, not for speed. I would guess they can lop 50mm^2 off of the total.

    -Charlie

  9. #389
    Quote Originally Posted by Alexko View Post
    Tahiti.B is just what I'm calling it, it doesn't have an official name. In fact its very existence isn't official.

    If AMD wants to improve speed and power (most notably leakage) to get to Pitcairn's level, yes, they need a silicon respin, i.e. a B0 stepping.

    Ok, I thought leakage was more a function of process than of design. What can they change on the design that would significantly affect leakage, and yet have about the same chip performance?

    Quote Originally Posted by Charlie
    If they respin it, it will be for die area, not for speed. I would guess they can lop 50mm^2 off of the total.
    Are there open spaces on the current layout? Are parts of the chip not being used that can be removed?
    "Today's scientists have substituted mathematics for experiments, and they wander off through equation after equation, and eventually build a structure which has no relation to reality." - Nikola Tesla

  10. #390
    Senior Member
    Join Date
    Sep 2009
    Posts
    3,021
    Quote Originally Posted by Thrillseeker View Post
    Are there open spaces on the current layout?
    Sort of. It's transistor layout. Kinda like Tetris.
    When people tell me "You're going to regret that in the morning" I sleep in till noon because I'm a problem solver!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
WordPress Appliance - Powered by TurnKey Linux