The NEXT LAST R600 Rumours & Speculation Thread

Discussion in 'Pre-release GPU Speculation' started by Geo, Mar 1, 2007.

Thread Status:
Not open for further replies.
  1. Creig

    Newcomer

    Joined:
    Nov 20, 2006
    Messages:
    57
    Likes Received:
    1
    Please stop quoting Inq/Fuad.



    Just... please.
     
  2. INKster

    Veteran

    Joined:
    Apr 30, 2006
    Messages:
    2,110
    Likes Received:
    30
    Location:
    Io, lava pit number 12
    Well, since they started the rumor in the first place, it's only fitting they should be the ones to bury it. ;)
     
  3. Kaotik

    Kaotik Drunk Member
    Legend

    Joined:
    Apr 16, 2003
    Messages:
    8,184
    Likes Received:
    1,841
    Location:
    Finland
    It has been buried by other sources as well, even if it wasn't in such big manner as another news "article
    but anyway, wasn't it k-hardware or something that said first it would be 65nm, not inquirer?
     
  4. Razor1

    Veteran

    Joined:
    Jul 24, 2004
    Messages:
    4,232
    Likes Received:
    749
    Location:
    NY, NY
    Heh at first I thought you were talking about a possible Inq future posting :wink:
     
  5. INKster

    Veteran

    Joined:
    Apr 30, 2006
    Messages:
    2,110
    Likes Received:
    30
    Location:
    Io, lava pit number 12
    K-hardware's google-auto-translation was very ambiguous, to say the least.
    So i'm willing to give them the benefit of the doubt. ;)
     
  6. Doomtrooper

    Veteran

    Joined:
    Feb 6, 2002
    Messages:
    3,328
    Likes Received:
    0
    Location:
    Ontario, Canada
    Oh yes still alive ;)
    Agreed, although I personally wish some of these demos had a screensaver mode. I also believe that both the IHV heavyweights are going to have a hard time 'impressing the consumer' just do to technological equality.

    Some of the leaked benchmarks mean nothing.
     
  7. overclocked_enthusiasm

    Regular

    Joined:
    Apr 26, 2004
    Messages:
    424
    Likes Received:
    3
    Location:
    United States
    My head hurts...so WTF is going on with R600?

    1. R600 is 80 nm and they waited to launch it with the rest of the R600 family and/or early Barcelona production samples?

    2. R600 is actually R650 and is 65 nm as R600 at 80 nm was just unworkable and all the AIB partners are in the dark?

    3. R600 is going to be targeted into a lower price point (I am assuming because of yield issues) and R650 at 65 nm will then compete against the G80 flagship refresh in the fall leaving the enthusiast segment to Nvidia unopposed until then?

    4. Fuad and Inq are so utterly and hopelessly confused that none of the above is true.


    I am putting my money on #1 with a slight chance of #2 and sprinkled with some #4. It will be funny to see what other explanations we get in the next month or so.
     
  8. Kaotik

    Kaotik Drunk Member
    Legend

    Joined:
    Apr 16, 2003
    Messages:
    8,184
    Likes Received:
    1,841
    Location:
    Finland
    It's not 65nm, #1 + mysterious reason why they originally pushed it forward so much and then got the idea about "all-in-one" launch
     
  9. Geo

    Geo Mostly Harmless
    Legend

    Joined:
    Apr 22, 2002
    Messages:
    9,116
    Likes Received:
    213
    Location:
    Uffda-land
    Looking at Tridam's report, it's not real clear just how many R600 skus they are launching with, so arguably they may have still rushed it. Personally, I scratched the liklihood of it being a clearly dominant part when we got a look at G80. Then when 512-bit became clear I upped that to being possibly dominant in some narrow selected scenarios (some of which might be quite tasty, however). That's still pretty much where I am. I've just never believed in putting that level of expectations on a product in advance. Competency and competitiveness we all have a right to expect; greatness should be judged after you've actually seen it.
     
  10. R300King!

    Newcomer

    Joined:
    Aug 4, 2002
    Messages:
    231
    Likes Received:
    5
    I say #2 & #3 with the first part of #4 ;)

    Fuad's doin his disinformation campaign. Not lying but telling us R600 is 80nm and not 65nm. Yes, this we all know. If you change from 80nm to 65nm more than likely you'd need to do a whole redesign and that chip would be called something else...most likely R650 or R680.
    But I think we'll see some R650s and the OEMs will see the R600s first. When 65nm yields for the R650 improve then we'll see wider availability for everyone and the OEMs will get the R650s too.
     
  11. icecold1983

    Banned

    Joined:
    Aug 4, 2006
    Messages:
    649
    Likes Received:
    4
    why would g80 being a great product prevent ati from launching an even better one? seems they aimed higher.
     
  12. Geo

    Geo Mostly Harmless
    Legend

    Joined:
    Apr 22, 2002
    Messages:
    9,116
    Likes Received:
    213
    Location:
    Uffda-land
    Depends on your frame of reference, I suppose. A product that is as good as a great product is a great product, isn't it? I suspect that's not what other people mean, however. Someone stated that a product that was 20% faster than a great product would be disappointing, for instance. What would you call a great product that is as much greater than another great product was greater than whatever you were comparing it to? Umm, "really great"?

    Yes there are limitations, and different priorities. We've seen, for instance, that HDMI sound seems to be in R600 silicon. I haven't heard it's in G80s. There's been indications that UVD may be in there too, so that'd be another chunk of silicon of unknown size (tho also unknown comparison to PV). There is no evidence they went the NVIO route and sucked that chunk of silicon out of the main die. Doing 512-bit must have cost them some mc control logic that didn't go elsewhere.
     
  13. Razor1

    Veteran

    Joined:
    Jul 24, 2004
    Messages:
    4,232
    Likes Received:
    749
    Location:
    NY, NY
    heh that first paragraph gave me a headache :wink:
     
  14. Geo

    Geo Mostly Harmless
    Legend

    Joined:
    Apr 22, 2002
    Messages:
    9,116
    Likes Received:
    213
    Location:
    Uffda-land
    Those last two sentences! :lol:
     
  15. Aerows

    Regular Newcomer

    Joined:
    Nov 19, 2002
    Messages:
    317
    Likes Received:
    6
    Me too. I'm beginning to wonder how much wood could a woodchuck chuck if a wood chuck could chuck wood, and that just isn't right when I'm on a forum about 3D hardware :grin: .
     
    Razor1 likes this.
  16. Geo

    Geo Mostly Harmless
    Legend

    Joined:
    Apr 22, 2002
    Messages:
    9,116
    Likes Received:
    213
    Location:
    Uffda-land
    This is a difficult question. I'd like to know more about the number of woodchucking units the particular woodchuck you have in mind has, whether they are tied on a 1-1 basis to his wood culling units or are disassociated to hide latencies, and if his woodchucking and culling scheduling units are robust, or 1st generation, or what.
     
  17. icecold1983

    Banned

    Joined:
    Aug 4, 2006
    Messages:
    649
    Likes Received:
    4
    geo would u agree that nvidia usually doesnt push the envelope and instead focuses on great margins? its usually ati who pushes the envelope of performance/clocks. nv30 excluded.

    if a product launches half a year or more after another product at a higher price(its very unlikely that r600 will be cheaper than g80 at launch), that product has to be substantially better not to be a disappointment. otherwise u waited 6 months to pay more for something u had over half a year ago.

    heres an interesting what if situation:

    if we all knew back when g80 launched that r600 would be 5% faster, how many people would bother waiting for r600? i would guess only the biggest ati supporters.
     
    #977 icecold1983, Mar 21, 2007
    Last edited by a moderator: Mar 21, 2007
  18. Geo

    Geo Mostly Harmless
    Legend

    Joined:
    Apr 22, 2002
    Messages:
    9,116
    Likes Received:
    213
    Location:
    Uffda-land
    I think you're making assumptions on prices that may or may not be accurate. There are indications otherwise. Look at Richard's statement and recognize what broader means. Look at what Kombatant was quoted as saying at Rage3D.

    I think Nvidia was fairly scarred from NV30, and claimed to learn some lessons. One of the major ones was "go wider rather than faster". I think it's not clear just yet how it works out vs R600, given clock domains and such. How do you square "not pushing the envelope" with 1.35GHz shaders? That's roughly, what? 60% or so faster than any piece of any previous GPU, from any manufacturer?
     
  19. icecold1983

    Banned

    Joined:
    Aug 4, 2006
    Messages:
    649
    Likes Received:
    4
    well im making assumptions on the price yes, but i will be shocked if when r600 launches, the street price isnt higher than g80s street price.

    when g80 runs the close to the same temp as r580, and barely uses more power its hard to believe that they couldnt have done more with the card if they wanted to.

    edit - just checked pricewatch, 8800gtx as low as 543, if r600 can be had that cheaply at launch, awesome. but im not expecting it. it also wouldnt be a stretch to imagine that g80s price will go down when r600 ships.
     
  20. INKster

    Veteran

    Joined:
    Apr 30, 2006
    Messages:
    2,110
    Likes Received:
    30
    Location:
    Io, lava pit number 12
    Yes, but since they went all scalar, it's likely that clock speed was planned because scalar made it easier to achieve.
    So, we have to frame that 1.35GHz figure in the architecture of a specific part of the GPU, not necessarily in the need for speed vs the competition of the moment.
    I mean, it was about efficiency vs brute force, not exactly what the comparatively huge 1350MHz number might have made us suppose at first glance.

    Well, before 7800 GTX 512 came along, even the regular 7800 GTX'es were very difficult to overclock from the standard 430MHz up to ~500MHz with stock cooling.
    Then, suddenly, the 7800 GTX 512 showed up with a 550MHz core, still built on the old 110nm half node, and sporting a redesigned cooler that was actually in use on the earlier QuadroFX 4500.
     
    #980 INKster, Mar 21, 2007
    Last edited by a moderator: Mar 21, 2007
Loading...
Thread Status:
Not open for further replies.

Share This Page

  • About Us

    Beyond3D has been around for over a decade and prides itself on being the best place on the web for in-depth, technically-driven discussion and analysis of 3D graphics hardware. If you love pixels and transistors, you've come to the right place!

    Beyond3D is proudly published by GPU Tools Ltd.
Loading...