AMD Cayman Architecture article online

Article: AMD's Cayman GPU Architecture
By: Vincent Diepeveen (diep.delete@this.xs4all.nl), December 17, 2010 7:02 am
Room: Moderated Discussions
DaveC (davecuny@gmail.com) on 12/16/10 wrote:
---------------------------
>EduardoS (no@spam.com) on 12/16/10 wrote:
>---------------------------
>>an (an@yahoo.com) on 12/16/10 wrote:
>>---------------------------
>>>In other words, Nvidia products are equally powerful, but more flexible and easier to get the performance out of.
>>
>>In real life, nVidia products have half (at best) raw performance for int and float and so for anything bound on this.
>>
>
>It looks like "an" and I are asking the same question:
>
>Fermi looks to be a better GPGPU architecture than Cayman, but Fermi is getting
>about 50% efficiency with a theoretical max of 500 GFlops (i.e. real-world 250 GFlops).
>So, since Cayman has a theoretical max of 675 GFlops, meaning 250 GFlops is 37%
>efficiency for it . . . so who would win in a pure FLOPs contest?

You're looking wrong. You buy 2 cards with 2 gpu's for a 2000 euro. You buy a single machine for a 1500 euro or so.

For 4000 euro you've got a node with 2 cards and 4 cayman gpu's delivering single precision double digit teraflops.

Now please calculate the price of a similar nvidia solution.

Of course you're not going to do this if you want to write a program within 2 days, being a lazy university dude, just in order to 'test' gpu performance.

You do the above if you really need performance, then you have a budget also to pay for a good gpu coder.

Assume 50% IPC for such project out of AMD and 25-30% out of nvidia.

Most those projects you don't write double precision. It's the lazy university dudes that want double precision. You can do transforms also in single precision and use CRT to effectively use more bits. CRT = chinese remainder theorem.

Google for it.

In my case it's about artificial intelligence. Everything is single precision anyway there.

Raw performance is what matters.

All you want is raw performance. Then you hire someone to get it out of the card.

It's wishful thinking that with 2 days of coding you can get any good performance out of any gpgpu solution anyway.

Vincent
< Previous Post in ThreadNext Post in Thread >
TopicPosted ByDate
AMD Cayman Architecture article onlineDavid Kanter2010/12/15 07:39 AM
  AMD Cayman Architecture article onlineBryan Catanzaro2010/12/15 12:25 PM
  AMD Cayman Architecture article onlineCarsten Spille2010/12/15 02:51 PM
  AMD Cayman Architecture article onlineDaveC2010/12/15 05:17 PM
    AMD Cayman Architecture article onlineAntti-Ville Tuunainen2010/12/16 01:13 AM
      AMD Cayman Architecture article onlinePing-Che Chen2010/12/16 02:39 AM
        AMD Cayman Architecture article onlineEduardoS2010/12/16 01:54 PM
          AMD Cayman Architecture article onlineDavid Kanter2010/12/16 03:46 PM
            AMD Cayman Architecture article onlineEduardoS2010/12/16 06:03 PM
    AMD Cayman Architecture article onlineVincent Diepeveen2010/12/17 07:16 AM
  AMD Cayman Architecture article onlinean2010/12/16 12:39 PM
    AMD Cayman Architecture article onlineRichard Cownie2010/12/16 01:51 PM
      AMD Cayman Architecture article onlineVincent Diepeveen2010/12/17 07:31 AM
        AMD Cayman Architecture article onlineRichard Cownie2010/12/17 09:22 AM
    AMD Cayman Architecture article onlineEduardoS2010/12/16 02:01 PM
      AMD Cayman Architecture article onlinean2010/12/16 02:43 PM
        AMD Cayman Architecture article onlineEduardoS2010/12/16 02:51 PM
          AMD Cayman Architecture article onlineDaveC2010/12/16 03:41 PM
            AMD Cayman Architecture article onlinehobold2010/12/16 03:56 PM
              AMD Cayman Architecture article onlineDaveC2010/12/16 05:31 PM
            AMD Cayman Architecture article onlineVincent Diepeveen2010/12/17 07:02 AM
        AMD Cayman Architecture article onlineAaron Spink2010/12/16 03:39 PM
        AMD Cayman Architecture article onlineDavid Kanter2010/12/16 03:48 PM
          AMD Cayman Architecture article onlineVincent Diepeveen2010/12/17 07:07 AM
        AMD Cayman Architecture article onlineVincent Diepeveen2010/12/17 06:56 AM
  Logic error on the articleHeikki Kultala2010/12/17 03:59 AM
    Good pointDavid Kanter2010/12/17 11:21 AM
      Good pointTriskaine2010/12/17 01:02 PM
        Good pointDavid Kanter2010/12/17 04:45 PM
      Good pointJohn2010/12/20 08:05 PM
  WavefrontsMoritz2010/12/20 04:11 AM
    WavefrontsSeni2010/12/20 01:07 PM
      TexturesMoritz2010/12/21 01:41 AM
        TexturesGabriele Svelto2010/12/21 02:21 AM
        TexturesAntti-Ville Tuunainen2010/12/21 02:33 PM
  Integer computerRobert David Graham2010/12/21 04:45 PM
    Integer computeranon2010/12/21 08:22 PM
      Integer computeretzel2010/12/21 09:12 PM
  TransposeMoritz2010/12/23 03:44 PM
    TransposeMatt Sayler2010/12/23 07:58 PM
      Transpose/RotateMoritz2010/12/24 01:38 AM
        Transpose/RotateMatt Sayler2010/12/24 06:45 AM
          Transpose/RotateMoritz2010/12/25 05:33 AM
  control-data associated with scalarsMoritz2010/12/23 03:59 PM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell avocado?