Futuremark Announces Patch for 3DMark03

Discussion in 'Graphics and Semiconductor Industry' started by Nick[FM], Nov 11, 2003.

  1. Sharkfood

    Regular

    Joined:
    Feb 7, 2002
    Messages:
    702
    Likes Received:
    11
    Location:
    Bay Area, California
    You would actually need to use some form of digital capture device in order to objectively make this determination.

    NVIDIA has already been accused of performing a single frame re-render when a frame buffer lock is detected, this invalidating screenshot grabs from comparison. This makes it so even 3DMark's own image comparisons may be comparing non-valid captures.
     
  2. demalion

    Veteran

    Joined:
    Feb 7, 2002
    Messages:
    2,024
    Likes Received:
    1
    Location:
    CT
    It would be nice if we could, as one example, take the camera off the rail for ourselves and note any performance and/or visual rendering descrepencies that might result with varioud IHV driver versions.

    This remains one of the tools Futuremark could conceivably provide us with to make our information about defeated cheats (and the questions about the possibility of cheats that aren't defeated) more complete, but it is one they have apparently decided not to provide to a wide selection of consumers.
     
  3. digitalwanderer

    digitalwanderer Dangerously Mirthful
    Legend

    Joined:
    Feb 19, 2002
    Messages:
    17,120
    Likes Received:
    1,653
    Location:
    Winfield, IN USA
    Forceware (Detonators) 53.03 for win2k/XP that I got the heads up from off of nForceMan over at nVnews.
    I'm just wondering if these "fix" that 13-15% loss FX owners have been getting since the 3dm2k3 3.40 patch came out...
     
  4. 3dilettante

    Legend Alpha

    Joined:
    Sep 15, 2003
    Messages:
    8,117
    Likes Received:
    2,858
    Location:
    Well within 3d
    I think there are a few posts around that state that users have gotten back a good amount of the lost points. Considering the seeming simplicity of Futuremark's changes (even I understood a good amount of what they did), I guess it wouldn't be too difficult to add a second set of shader fingerprints for 3dmark.

    Now, my following statement is contingent on the following things being true--if these performance gains in 3dmark are present, and if they have been gained through replacement shaders:

    Nvidia driver department--you are hacks. I'm sorry to say, that regardless of whether some of you do not work on pointless driver bloating replacement shaders, you are by association members of a first rate collection of software hacks. Not hackers, who do things exemplified by a mastery of code and elegant examples of cleverness, but hacks in the sense of lousy after school special writers, or a thousand monkeys at an approximately equal number of typewriters.
    I'm pretty sure it's not entirely your fault, and that you normally probably wouldn't think about doing this. I'm sure a number of you wish you didn't have to do this. However, that only makes me empathize with people who were unwillingly forced into being hacks.
    In addition, most of you are probably stand-up individuals, who would be fun to get to know and invite to some kind of social function, but--and I give my condolences--you have been reduced to digital hacks, with whatever skills and capability you have debased by actions seemingly more appropriate to a middle-schooler peering over someone's shoulder to get an answer to a math quiz.
    Regretfully, even should the latest drivers be the best Nvidia ever made (which is somewhat questionable, since they were written by hacks), there will always be the asterisk over your achievement, since you are part of a team of hacks. Hopefuly, there will come a time where you can simply be one of the premier driver development teams anywhere, as opposed to ridiculously overqualified hacks. Here's to hoping your employer gets its act together soon.

    edited to change some wording
     
  5. XForce

    Newcomer

    Joined:
    Jun 12, 2003
    Messages:
    58
    Likes Received:
    0
    Off course they do!

    As confirmed by bkswaney here: http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=20892&pagenumber=3

    Like I said before: wash - rinse - repeat :roll:

    You gotta love those "leaked" drivers..
     
  6. Tim Murray

    Tim Murray the Windom Earle of mobile SOCs
    Veteran

    Joined:
    May 25, 2003
    Messages:
    3,278
    Likes Received:
    66
    Location:
    Mountain View, CA
    http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=21177&pagenumber=3

    Yup. 3DMark scores are back up, PS2.0 scores do not increase according to Shadermark, but there are also some oddities:

    1. Ragtroll increase 1.7 FPS (over 10%) between 52.16 and 53.03 while CPU scores remain the same?

    2. PS2.0 test *drops* 7 FPS?

    3. Vertex shader up 4 FPS (from 13.1 to 17.1--almost 30%)?

    Haven't had time to try anything else yet.
     
  7. Dave Baumann

    Dave Baumann Gamerscore Wh...
    Moderator Legend

    Joined:
    Jan 29, 2002
    Messages:
    14,079
    Likes Received:
    648
    Location:
    O Canada!
    Just remember:
    <center>An optimization must produce the correct image
    An optimization must accelerate more than just a benchmark
    An optimization must not contain pre-computed state
    </center>
     
  8. davefb

    Newcomer

    Joined:
    Feb 7, 2002
    Messages:
    148
    Likes Received:
    1
    Location:
    manchester- UK
    ahh but the general principle of "sticking our own shader in" will optimize other games ;)

    i was thinking about how other companies could do this.. i reckon VIA should release new drivers which recognise that sisoft sandra is being run.. Now since the only memory write which really matters is the last one , the drivers might as well ignore all others.. Just THINK how fast the memory will be reported !!!

    -dave-
     
  9. XForce

    Newcomer

    Joined:
    Jun 12, 2003
    Messages:
    58
    Likes Received:
    0
    Dave B.:
    is that from FM's or NV's guidelines? I honestly can't remember.

    other Dave:
    don't give them any ideas.. :D
     
  10. digitalwanderer

    digitalwanderer Dangerously Mirthful
    Legend

    Joined:
    Feb 19, 2002
    Messages:
    17,120
    Likes Received:
    1,653
    Location:
    Winfield, IN USA
    Yeah, what's your point? Aren't you just glad that nVidia "fixed" their problem? ;)

    EDITED BITS: Unchecked "Disable HTML in this post" so Dave's quote worked better. ;)
     
  11. nelg

    Veteran

    Joined:
    Jan 26, 2003
    Messages:
    1,557
    Likes Received:
    42
    Location:
    Toronto
    Does the new driver score the same on the 330 ver. Of 3DMark03 ?
     
  12. Tim Murray

    Tim Murray the Windom Earle of mobile SOCs
    Veteran

    Joined:
    May 25, 2003
    Messages:
    3,278
    Likes Received:
    66
    Location:
    Mountain View, CA
    Word on the grapevine is yes.
     
  13. Hanners

    Regular

    Joined:
    Jul 12, 2002
    Messages:
    816
    Likes Received:
    57
    Location:
    England
    That was from nVidia's own guidelines.
     
  14. jb

    jb
    Veteran

    Joined:
    Feb 6, 2002
    Messages:
    1,636
    Likes Received:
    7
    Maybe we should change guidelines to "guidelines" :)
     
  15. John Reynolds

    John Reynolds Ecce homo
    Veteran

    Joined:
    Feb 7, 2002
    Messages:
    4,490
    Likes Received:
    262
    Location:
    Westeros
    I prefer guidelies.

    :twisted:
     
  16. digitalwanderer

    digitalwanderer Dangerously Mirthful
    Legend

    Joined:
    Feb 19, 2002
    Messages:
    17,120
    Likes Received:
    1,653
    Location:
    Winfield, IN USA
    Mebbe we should change "guidelines" to "PR lip-noises to appease the gullible"? :|
     
  17. digitalwanderer

    digitalwanderer Dangerously Mirthful
    Legend

    Joined:
    Feb 19, 2002
    Messages:
    17,120
    Likes Received:
    1,653
    Location:
    Winfield, IN USA
    Oooh, that one was much better than mine! :lol:
     
  18. tay

    tay
    Newcomer

    Joined:
    Nov 19, 2003
    Messages:
    14
    Likes Received:
    0
    RTCW: Enemy Territory

    Anyone have a chance to try ET with these drivers? 52.xx seem to cause a hard lockup.

    thx
     
  19. Deathlike2

    Regular

    Joined:
    Aug 17, 2003
    Messages:
    542
    Likes Received:
    5
    It sounds more like...

    An optimization must produce the correct image, whether or not our hardware can produce it normally with the optimization.
    An optimization must accelerate more than just a benchmark, especially if its a popular one.
    An optimization must not contain pre-computed state, except if it doesn't use our proprietary extensions or special coding to produce a similar effect.
     
  20. bloodbob

    bloodbob Trollipop
    Veteran

    Joined:
    May 23, 2003
    Messages:
    1,630
    Likes Received:
    27
    Location:
    Australia
    Best drivers are 2.08 because it had AA on my TNT 2.

    Yeah people are getting like 800 points back.

    On the futuremark board someone was saying supposedly the only app specific code that is in the drivers is for vertex shaders as they currently have no vertex shader dynamic recompilation module in the driver.
     
Loading...

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...