3DMark05 scores courtesy the Inq

Discussion in '3D Hardware, Software & Output Devices' started by trinibwoy, Sep 28, 2004.

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

    Banned

    Joined:
    Jan 31, 2002
    Messages:
    3,266
    Likes Received:
    24
    Well, in a parallel to DST, 3Dc could be an option (have 3Dc and DXT5). Sure, it means more work for FM but there you go.

    DST should have been an option to enable, not an option to disable, by default. Same with 3Dc if it was used.
     
  2. kyleb

    Veteran

    Joined:
    Nov 21, 2002
    Messages:
    4,165
    Likes Received:
    52
    Yet DST doesn't change image quality, does it?
     
  3. Dave Baumann

    Dave Baumann Gamerscore Wh...
    Moderator Legend

    Joined:
    Jan 29, 2002
    Messages:
    14,079
    Likes Received:
    648
    Location:
    O Canada!
    Inclusion or lack of 3Dc is not an issue, its somewhat of a deflection and just an easy parallel. The point being is that the default path is using non-defined (within DX), IHV specific features, something that FM had said they would not do before (the deflection comment is if you are going to do it for one, why not the other).
     
  4. Hanners

    Regular

    Joined:
    Jul 12, 2002
    Messages:
    816
    Likes Received:
    57
    Location:
    England
    Actually, it seems that it does.
     
  5. kyleb

    Veteran

    Joined:
    Nov 21, 2002
    Messages:
    4,165
    Likes Received:
    52
    A very valid argument, but I am sure how you can see where the improvement in image quality offered by 3dc over dxt5 would be troublesome within the context of 3dmark as image quality comparison has become a large part of their benchmark.

    But with crap like that the aforementioned argument goes right out the window. I for one am very interested in seeing what becomes of "Futuremark approved drivers" in regards to this.
     
  6. Pete

    Pete Moderate Nuisance
    Moderator Veteran

    Joined:
    Feb 7, 2002
    Messages:
    4,936
    Likes Received:
    333
    So, are there vertex instancing gains to be had from GT2? If so, is ATi missing out, or does 3DM accept the fourcc detection hack?

    So it'll take "a dozen of the biggest game devs" before FM considers including 3Dc support? How far out is that?

    Also, does enabling colored mips change the scores for either ATi or nV, both with and without optimizations?

    (BTW, HW.fr also illustrated the diff b/w nV's hardware DST and the "non-DST" rendering paths.)
     
  7. Reverend

    Banned

    Joined:
    Jan 31, 2002
    Messages:
    3,266
    Likes Received:
    24
    The largest part of their benchmark are the WOW factor of the game tests and the ORB. Image quality comparison is a tool they provide as an extra, as is all their other purely synthetic tests.

    We're going OT here wrt to what Dave is really talking about. And that has to do with the value of impartiality in an app that is promoted as a benchmark (which, of course, must be absolutely impartial to have any credibility).

    I think FM has focussed more on anticipating how "games of the future" will be made (with their game tests) instead of what 3DMark really is (or more importantly, what FM thinks it is), a benchmark meant to show how various hw perform on an absolutely level playing field.

    The more I think about this, the more upset I get.
     
  8. PowerK

    Regular

    Joined:
    Jun 12, 2004
    Messages:
    312
    Likes Received:
    0
    Location:
    Seoul, Korea
    Well said. I agree 100%.
     
  9. Fodder

    Fodder Stealth Nerd
    Veteran

    Joined:
    Jul 12, 2003
    Messages:
    1,112
    Likes Received:
    9
    Location:
    Sunny Melbourne
    Where do you suggest the line is drawn? Should FutureMark only exploit features that;

    a) All vendors support
    b) Are part of the DirectX spec
    c) Don't impact IQ
    d) ....
     
  10. Reverend

    Banned

    Joined:
    Jan 31, 2002
    Messages:
    3,266
    Likes Received:
    24
    If FM wanted to show how certain tech/features can be a boost to games (in private emails, it is said by FM that most games that use depth shadow map will use DST, which is a main contributant to why it is enabled by default, which is to say they're looking at 3DMark as agme basically), these features/tech should NOT be enabled by default unless all hardware supports them. DST is not supported by all hardware. 3Dc is not supported by all hardware. Using either as default is Not Right. Not in a benchmark. We're not talking shader model support here.
     
  11. Dave Baumann

    Dave Baumann Gamerscore Wh...
    Moderator Legend

    Joined:
    Jan 29, 2002
    Messages:
    14,079
    Likes Received:
    648
    Location:
    O Canada!
    Nor is it defined or controlled in DirectX and you would not be able to get the same functionality in the DX reference rasteriser.
     
  12. Bjorn

    Veteran

    Joined:
    Feb 6, 2002
    Messages:
    1,775
    Likes Received:
    1
    Location:
    Luleå, Sweden
    If 3D Mark shouldn't focus on anticipating how "games of the future" will be made then they need to skip the game tests altogether imo. Since they have the game tests and they're there to give us a hint of future game performance, then i'd say that talking to game developers and trying to do the same thing that they're doing is a good thing.
     
  13. PowerK

    Regular

    Joined:
    Jun 12, 2004
    Messages:
    312
    Likes Received:
    0
    Location:
    Seoul, Korea
    Hasn't 3DMark always been about measuring DirectX (D3D) performance ? The line drawn is pretty clear to me.
     
  14. kyleb

    Veteran

    Joined:
    Nov 21, 2002
    Messages:
    4,165
    Likes Received:
    52
    Then it is quite possable it will be another thing like the buffer shadows in Splinter Cell, eh? Where it was an exclusive feature of nvidia cards and now it doesn't even work with newer drivers.
     
  15. Reverend

    Banned

    Joined:
    Jan 31, 2002
    Messages:
    3,266
    Likes Received:
    24
    I didn't say they shouldn't focus on it -- the game tests serve a purpose IMO. I said they focussed too much on it and as a consequence probably "forgot" that 3DMark is "just a benchmark".
     
  16. Bjorn

    Veteran

    Joined:
    Feb 6, 2002
    Messages:
    1,775
    Likes Received:
    1
    Location:
    Luleå, Sweden
    Perhaps. I would agree with you if enabling 3Dc lead to a performance increase vs DX5. If not then i don't see the problem.
     
  17. Fodder

    Fodder Stealth Nerd
    Veteran

    Joined:
    Jul 12, 2003
    Messages:
    1,112
    Likes Received:
    9
    Location:
    Sunny Melbourne
    What about vendor specific features like PS2_b? Sure it's a DX standard and anyone can support it, but anyone could implement 3Dc too. I think the DST/PCF default (and perhaps inclusion full stop) is pretty poor as the IQ impact is quite noticeable, but what about features that provide a performance boost with no side effects?
     
  18. Reverend

    Banned

    Joined:
    Jan 31, 2002
    Messages:
    3,266
    Likes Received:
    24
    Bjorn, like Dave said, 3Dc is mentioned to provide an example. Look at the bigger picture -- what should the guidelines be behind a benchmark app?
     
  19. GraphixViolence

    Newcomer

    Joined:
    Jul 3, 2002
    Messages:
    194
    Likes Received:
    2
    I think even more important than the question of whether or not Futuremark should use vendor-specific features in the benchmark, is how can they possibly justify using one vendor's features but not another's?

    Right now, 3DMark05 is using SM3.0, partial precision, and DST by default. All three of these features benefit only NVidia hardware... not ATI's, S3's, XGI's, 3DLabs', Matrox's, etc. In the case of DST, we're talking about a feature that is not officially supported in D3D, and visibly affects image quality. In contrast, 3Dc, which falls into the exact same category is not even available as an option, let alone enabled by default.
     
  20. Ichneumon

    Regular

    Joined:
    Feb 3, 2002
    Messages:
    414
    Likes Received:
    1
    Without quoting the several posts this refers to (the Nvidia shadowing on by default bits)...

    What it sounds like is that Nvidia marketing got Futuremark to believe that Nvidia is going to have DST in everyones games so Futuremark should have it on by default... and that is rediculous!

    If people thought Futuremark had sold out before, just wait until this really gets out there. I feel sorry for them. I think they've really shot themselves in the foot wrt that decision. All of their fighting for a level playing field in 3dm03 was all a sham, and now that they got green coming in from Nvidia they just pretend 3dm03 never happend. What a waste of a lot of talent.

    If Futuremark wants to build in IHV specific optimizations, that should be a benchmark option. So users can see the difference with optimizations on and off.... but that should absolutely Not be the default behavior.
     
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...