3dmark2001 and GLexcess 6094 vs 6102

Discussion in 'Overclocking, Benching & Modding' started by Renegade[CSR], Jun 27, 2002.

  1. Renegade[CSR]

    Renegade[CSR] New Member

    Joined:
    May 7, 2002
    Messages:
    396
    Likes Received:
    0
    Trophy Points:
    0
    Ok well here are my specs.

    AMD Athlon 1700+ @ 1900+
    Radeon 8500 @ 292/288
    384 mb DDR 266 pc2100 ram

    Everything is being run at balanced settings, driver defaults.

    3d mark 2001: Here are some surprising results. I had expected an increase in this area. Well, I got a decrease in performance.

    6094: 8792 3d marks
    6102: 8712 3d marks

    It isn't noticeable at all, and I wasn't concerned about D3D anyways, as its performance for the card is already more than adequate.

    Onto Glexcess. In this benchmark, I had already noticed a, not to shabby 1000 point XS mark increase from 6093 to 6094. My results for this set were satisfying to say the most.

    6094: 7019 XS marks
    6102: 7057 XS marks

    Opengl is still not where it should be in my opinion. Once it reaches about the mid-8000 range, I think that that will be adequate.

    Ah but, all hope is not lost yet. I ran MOHAA with the new drivers and noticed a considerable performance increase in high stress areas (such as the first snow level) It was noticeably faster.

    Nothing really new to report with these drivers, other than they provide a slight increase in Opengl, which still needs work, and D3D is pretty much where it should be as always, any increase in performance is always a bonus. Hopefully for my next benchmarks Ill have Commache 4 and Vulpine GL to try out. :)
     
  2. HardwareHeaven

    HardwareHeaven Administrator Staff Member

    Joined:
    May 6, 2002
    Messages:
    32,274
    Likes Received:
    163
    Trophy Points:
    88
    interesting your Glexcess mark increases with this new set and yet peroni had a decrease, I suppose the marks are so close its within system variables.
     
  3. Renegade[CSR]

    Renegade[CSR] New Member

    Joined:
    May 7, 2002
    Messages:
    396
    Likes Received:
    0
    Trophy Points:
    0
    Yes. I find that each benchmark can go plus or minus 20.

    Although, I find that glexcess is much more consistent in its results.
    The only thing that can hinder it is those happy MSN messenger windows that ALWAYS seem to decide to come up right at the time im benching. :uhoh: They seem to deal a hit on any opengl program.

    Ah well, no one benchmark is perfect ;)
     
  4. HardwareHeaven

    HardwareHeaven Administrator Staff Member

    Joined:
    May 6, 2002
    Messages:
    32,274
    Likes Received:
    163
    Trophy Points:
    88
    I would certainly kill MSN and as many running processes as possible when benching for sure.
     
  5. T2k

    T2k Guest

    Re:

    Agreed. :)


    Just like on my side.
     
  6. merry

    merry gargouille

    Joined:
    Jun 21, 2002
    Messages:
    962
    Likes Received:
    1
    Trophy Points:
    28
    Re:

    I'm not benching a lot, but I noticed - under 98SE - that killing everything else - CPUidle and stuff - would improve 3DM score by a few %
     
  7. raptora12

    raptora12 Mystic seeker

    Joined:
    May 10, 2002
    Messages:
    193
    Likes Received:
    0
    Trophy Points:
    0
    why don't you msconfig msn messenger never had 1 problem

    just a idea :hmm:
     

Share This Page

visited