Page 1 of 5 123 ... LastLast
Results 1 to 10 of 43

Thread: Nvidia cut 3/4 of the gtx480 speed:

  1. #1
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default Nvidia cut 3/4 of the gtx480 speed:

    http://forums.nvidia.com/index.php?s...&#entry1028870

    "Yes, full-speed double precision performance is a feature we reserve for our professional customers. Consumer applications have little use for double precision, so this does not really affect GeForce users. Having differentiated features and pricing is actually fairer for all. Given the option of enabling all professional features on GeForce and having gamers pay for them, or disabling them on GeForce and offering a more compelling price, we feel the latter is the better choice."

    LOL Nvidia GTX480 customers are very very very stupid and dump !

    the hd5870 is 4 time faster on that and is already cheaper!

  2. #2
    Join Date
    Jan 2008
    Posts
    772

    Default

    Nvidia's been doing this sort of thing for years, as have some other vendors. It's essentially the DRM mindset applied to hardware instead of software or media.

  3. #3
    Join Date
    Aug 2007
    Posts
    6,645

    Default

    And of course ati has got no firegl cards, right? Every vendor tries to sell professional cards with extra features, therefore the driver restricts the consumer cards - with basically the same chip. But it does not mean that there are no tricks to "fake" quadro/firegl cards.

  4. #4
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    Quote Originally Posted by Kano View Post
    And of course ati has got no firegl cards, right? Every vendor tries to sell professional cards with extra features, therefore the driver restricts the consumer cards - with basically the same chip. But it does not mean that there are no tricks to "fake" quadro/firegl cards.
    i can not see the trick@fireGL i try a hacket driver for my 4670 to get the fireGL version of the driver... and there is no diverend for me!

    trick? can't see it!

    in germany all PC letters like PCGH post benchmarks and they test on a fakes nvidia hardware gtx480 and they tell us the card do have the full speed like a tesla ! but its just fake! nvidia just send tesla cards to the reporters !

    in the end DP nvidia's PR talk to us 480 is faster than a 5870 but in real! the AMD is 4 time faster! yes 4Time!

    and AMD do not slowdown the 5870....

    amd just optimise the driver for special software to bring more speed as normal.

    no fake! i can get DP 0,550Tflops on 5870! but i can't get DP 0,650tflops on the nvidia gtx480!

    nvidia is just a fake!

  5. #5
    Join Date
    Aug 2007
    Posts
    6,645

    Default

    Well the main difference is in the opengl code for professional apps. Also for benchmarks like specviewperf, you can easyly verify that...

    Btw. the major difference between a normal gtx and a tesla is that ecc ram is used as bad ram is often a critical point. fermi allows to detect (and fix if simple enough) ecc ram errors, ati chips do not allow ecc ram usage, or do you see it mentioned anywhere? it does not only matter how fast you can calc something it has to proven to be correct too.

  6. #6
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    Quote Originally Posted by Kano View Post
    Well the main difference is in the opengl code for professional apps. Also for benchmarks like specviewperf, you can easyly verify that...
    i do not use specviewperf or maya or c4d,catia and so one.....

    I use Blender! and the Firegl fglrx do not accelerate Blender!

    for me as a Blender user FireGL cards are just a Joke!




    Quote Originally Posted by Kano View Post
    Btw. the major difference between a normal gtx and a tesla is that ecc ram is used as bad ram is often a critical point.
    no ! no! no! you can test the ram in a 5870 to! if the ram do have a problem you can use your Guaranty to chance it into a well working one without ram problems!

    ECC ram only cut cost for the company to save costs on Guaranty because you do not need to repleace the card!

    Guaranty means for you ECC check is not needet!

    i never use the ECC in my opteron system!

    because normal systems i check the ram manualy and if the check was fine i use the system if not i use the Guaranty to become a new ram!



    Quote Originally Posted by Kano View Post
    fermi allows to detect (and fix if simple enough) ecc ram errors, ati chips do not allow ecc ram usage, or do you see it mentioned anywhere? it does not only matter how fast you can calc something it has to proven to be correct too.

    ati chips can detect ram errors to!

  7. #7
    Join Date
    Jul 2008
    Location
    Berlin, Germany
    Posts
    847

    Default

    You seem very confused regarding the purpose of ECC and double-precision FP capabilities of the GTX480.

    GDDR5 has built-in error correction during transfer. But this is not the same as ECC. An important purpose of ECC is to detect spurious errors in memory, such as the ones caused by cosmic radiation (I don't think warranty covers cosmic radiation, does it? ). Also it detects when a memory chip on the card goes bad. This will allow you to decide which part of the calculation to throw away in case of error.

    Double-precision is not widely used outside scientific calculations as it gives you no significant advantage in 3D or video rendering.

    Providing the full extent of those features only to Tesla customers is called market segmentation and is a very common business practice as pointed out by Ex-Cyber already.

  8. #8
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    Quote Originally Posted by chithanh View Post
    GDDR5 has built-in error correction during transfer. But this is not the same as ECC. An important purpose of ECC is to detect spurious errors in memory, such as the ones caused by cosmic radiation (I don't think warranty covers cosmic radiation, does it? ). Also it detects when a memory chip on the card goes bad. This will allow you to decide which part of the calculation to throw away in case of error..
    ecc and error detection is not the same thats true!

    but if you only wana detect the error the 5870 will do the same work!

    for exampel why a gtx480 is pointless:

    if you buy 2 of 5870 ans you calculate the exakt same! you can detect errors by crosswatch the result!

    in SP the 5870 is much faster 2.7Tflos! vs gtx480 1,5tflops
    and in DP the 5870 is much faster to 0,55tflops vs 0,1xtdlops!
    the 5870 consumes 100watt less power!

    you also can buy a 5970 and crosscheck the result from GPU1 to GPU2!

    thats ALL cheaper and better! as a GTX480 with only 1,5/0,1x Tflops!!!

  9. #9
    Join Date
    Aug 2007
    Posts
    6,645

    Default

    Just that you can play more games with linux and have got vdpau as extra

  10. #10
    Join Date
    Nov 2008
    Location
    Germany
    Posts
    5,411

    Default

    Quote Originally Posted by Kano View Post
    Just that you can play more games with linux and have got vdpau as extra
    o yes.. nvidia is for games only but all the linux games run on amd to..

    only 'wine' have a main fokus on nvidia'

    vdpau... i do not watch movies very often (means never)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •