3dmark03 & patch 330 = slow

Discussion in 'Benchmarks & Performance' started by Ice-Freezer, May 24, 2003.

  1. Ice-Freezer

    Ice-Freezer Guest

    driver 3.4 catylsit
    cpu xp1800@1900
    mem 512sdram cas3 7.5ns 141mhz
    mobo jetway v266b
    hdd seagate baracuda v 80gb
    vid card powercolor 9700 non-pro modded to pro
    core 270@360
    mem 270@288 cant go higher and be stable :(

    3dmark03 313 = 4776
    after patching to 330 version i got 4605 score (twice)
    went back to 313 version got 4776 again

    mother nature test slowed down from 95fps in begging to 60 on 330 version

    i thought the patch had something to do with nvidia not ati :)
     
  2. neo

    neo Guest

    I am getting the message DrawIndexedPrimitive failec error.
     
  3. Sazar

    Sazar F@H - Is it in you? Staff Member Political User Folding Team

    Messages:
    14,905
    Location:
    Between Austin and Tampa
    ati has a 1.9% expected drop in score from tests done by futuremark.. almost all of it stemming from around an 8% odd drop in game test 4 (nature)

    the score SHOULD drop because either there is a valid bug in game test 4 or ati is running application specific coding in catalyst 3.4 version...

    the overall 1.9% score variance lies within futuremarks stated 3% variance unlike nvidia's 24% odd variance overall and over 50% variance in a couple of tests... ati's variance == around 2-2.5 fps in gt4...

    you score WILL drop... it is expected...
     
  4. Octopus

    Octopus Moderator

    Messages:
    1,200
    when you update to ver. 3.2 a message will appear saying that with DX9 you get higher score than DX9a
     
  5. Electronic Punk

    Electronic Punk Administrator Staff Member Political User Folding Team

    Messages:
    18,590
    Location:
    Copenhagen, Denmark
    Mine was faster.
     
  6. Ice-Freezer

    Ice-Freezer Guest

    that might be it cuz i got dx9a installed :)
     
  7. Gus K

    Gus K NTFS abuser

    Messages:
    380
    Location:
    USA
    It's only a silly benchmark. It's such a drag all the time, effort, and resources that companies put into getting good benchmark scores, instead of real world performance.