TechSpot

ATI has "something up our sleeve" for the 59xx series

By red1776
Nov 1, 2009
  1. koorosh

    koorosh TS Rookie

    Come on man! They hardly can get performance out of two gpus there's no need to go three IMO:haha:

    Waiting for nvidia launch... The competition would get really HOT:evil:
     
  2. fimbles

    fimbles TS Evangelist Posts: 1,160   +198

  3. red1776

    red1776 Omnipotent Ruler of the Universe Topic Starter Posts: 5,219   +157

    oh i don't know....fastest single GPU on the planet that rivals the dual GPU GTX 295...i think that's pretty good.
     
  4. koorosh

    koorosh TS Rookie

    Oh man why everyone misunderstands me here... I wasn't teasing just kidding.
    Not to mention that I'm not a GREEN guy. (and a RED guy neither)

    Honestly this time the triple gpu setup doesn't have optimal output on both sides.
    And about 58xx they are pretty awesome;) but I appreciate competition.
     
  5. red1776

    red1776 Omnipotent Ruler of the Universe Topic Starter Posts: 5,219   +157

    LOL koorosh, I wasn't being hostile, just giving you the needle back :rolleyes::D
     
  6. JieMan

    JieMan TS Rookie Posts: 63


    I doubt that seeing as PhysX is proprietary to NVIDIA and has been disabled in latest drivers with the use of an ATI gpu as its display hardware.

    Also there are two types of NVIDIA PhysX , GPU accelerated and CPU implemented ( like Havok ) , just because a game uses PhysX it doesn't mean its hardware accelerated on the GPU.

    AMD had an opportunity to develop PhysX whith NVIDIA a while back and implement it into the 4000 and 5000 series gpu's but they refused the offer to focus on the development of there own parallel computing environment ATI Stream.

    The major GPU companies are currently being pushed to the open-cl standard ( including Intel with Larrabee) for hardware accelerated physics and parallel computing so it wont be long before we see both NVIDIA PhysX ( CUDA ) and ATI Stream fade off into the sunset ( for most games anyway). This will bring the dawn of a better age for gaming and GPGPU computing where proprietary solutions are gone and both companies are on the same standard for the majority of games and software.

    I believe they will continue to develop there own solutions for a while so they can have exclusivity for future game titles with hardware accelerated physics and apps exclusively taking advantage of CUDA or STREAM.

    Good ehh? So they are using there flagship to compete with the competitors soon to be second tier gpu's ? I guess we'll know when Fermi comes to market .. who knows it could be crap but its looking good so far , might be why AMD is pulling things out of there sleeve for its 5000 series so soon some even before hitting the shelves 59xx for instance.
     
  7. red1776

    red1776 Omnipotent Ruler of the Universe Topic Starter Posts: 5,219   +157

    nice try jie, the comparison was a single gpu solution..compared to a dual gpu. and yes because thats the best Nvidia has out there right now. would you like them to compare it to the as of yet 'alleged' Fermi? that would be rather tough to benchmark, dont you think? and GPGPU ? folks are purchasing graphics cards for gaming at this point, not for GPGPU capabilities.
     
  8. JieMan

    JieMan TS Rookie Posts: 63

    Right a next gen single GPU vs a last gen dual gpu from the competitor not something for AMD to gloat about, Should they bench it against the Gtx295? sure but that's like benching against there own 4000 series it doesn't mean squat. It should be able to surpass the Gtx295 quite easily as its a newer generation.

    "Alleged" it might be but what NVIDIA is doing with the Fermi architecture is a breakthrough and an advancement for everyone in the GPU industry. Make no mistake Fermi is real and AMD is holding there pants!

    So what about GPGPU ? Well you might be doing alot more with a GPU than just playing games.
    Even tho this would benefit games most directly with hardware accelerated physics , much like NVIDIA PhysX on a GPU.
    But hopefully with OpenCl so everyone can benefit no matter who made the video card.

    The thing about the Fermi architecture is that it is not just designed to play games is it also going to take GPGPU to the next level with the advancements NVIDIA has made with support for CUDA, C++, Direct Compute, DirectX 11, Fortran, OpenCL, OpenGL 3.1 and OpenGL 3.2

    The big thing here is native support for C++ something thats ready to take the industry by storm . Look at Intel with the "Alleged" Larrabee GPU they are doing the same thing.

    I don't think you understand how fast this is going to spread "like a wildfire", its already being utilized in CUDA apps for converting codecs ( video/movies) or HD playback from ATI and NVIDIA, its going into the Adobe Flash player as well for Ion based netbooks and nettops. Anti virus makers are starting to notice how fast a parallel computing environment can be for them as well. Its uses are wide for big and small tasks.

    I think that AMD should have something up there sleeve related to this as well if they don't they might lose out seeing as how there GPU gaming performance is awesome I think they are concerned about the other end of the GPU's capabilities.
     
Topic Status:
Not open for further replies.

Similar Topics

Add New Comment

You need to be a member to leave a comment. Join thousands of tech enthusiasts and participate.
TechSpot Account You may also...