6800NU's 3DMark2K3 score is leaked by Galaxy! Updated!

Discussion in '3D Hardware, Software & Output Devices' started by solobird, May 20, 2004.

  1. Bjorn

    Veteran

    Joined:
    Feb 6, 2002
    Messages:
    1,775
    Likes Received:
    1
    Location:
    Luleå, Sweden
    Higher fillrate doesn't necessarily mean more shader power. Not that i know if this isn't the case with the X800 Pro vs 6800 NU though.
     
  2. snakejoe

    Newcomer

    Joined:
    Apr 27, 2004
    Messages:
    51
    Likes Received:
    3
    latest update:
    the first batch 6800nu maybe is 16ppc. 8)
    because the second batch 6800nu cant get that result under same condition.

    if you know chinese, look at here
    http://bbs.gzeasy.com/index.php?showtopic=170038&st=30
     
  3. jimmyjames123

    Regular

    Joined:
    Apr 14, 2004
    Messages:
    810
    Likes Received:
    3
    I don't think it really matters too much whether or not v61.11 drivers were used, because the main performance boost from those drivers was in FarCry, and performance actually went down in some games too. Notice that a 6800UE clocked at 450Mhz core, using the 60.72 Futuremark approved drivers, still scored about 1000 points higher in 3dmark03 than the X800XT, even though it still gives up fillrate to the X800XT PE:

    http://www.firingsquad.com/hardware/geforce_6800_ultra_extreme/page5.asp
     
  4. jimmyjames123

    Regular

    Joined:
    Apr 14, 2004
    Messages:
    810
    Likes Received:
    3
    This doesn't sound very likely either, because the 3dmark03 score is well below the other 16 pipe variants.
     
  5. jimmyjames123

    Regular

    Joined:
    Apr 14, 2004
    Messages:
    810
    Likes Received:
    3
    This is not necessarily true. The X800Pro has slightly more fillrate than a 6800GT, but the 6800GT generally has higher shader performance in Shadermark 2.0.
     
  6. Pete

    Pete Moderate Nuisance
    Moderator Legend

    Joined:
    Feb 7, 2002
    Messages:
    5,779
    Likes Received:
    1,816
    This thread is about the $300 6800 (12 pipes, 325/350).

    It's possible PCPop's 6800 has 16 pipes, as its 9K 3DM03 score roughly jives with a 6800GT's 11K, considering their "6800" only gives up 200MHz to a GT in memory speed. Memory speed may be a bottleneck with 16-pipe GPUs, even in shader-heavy 3DM03. I say this looking at TR's 3DM03 numbers: the 6800U loses ~1K with a ~10% slower core than a 6800UE, and the 6800GT loses another ~1K with a ~10% slower core and ~10% less memory bandwidth. I can see a 16-pipe 6800 at 350/800 (essentially a 6800GT with ~20% slower memory) give up 2K to a full 6800GT (11K -> 9K). These are all very, very rough numbers, though.

    Bjorn, I know that, I was just going on the assumption that R420 and NV40 have--for the moment, at least--similar IPC per pipe. I imagine benchmarks after a two months of driver development and game updates should put a stop to my gross generalizations.
     
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...