1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

GPU-Z and PowerColor Graphics Hardware Giveaway

Discussion in 'Contests' started by W1zzard, Aug 31, 2011.

Thread Status:
Not open for further replies.
  1. JohanDoe66 New Member

    Joined:
    Sep 23, 2010
    Messages:
    4 (0.00/day)
    Thanks Received:
    0
    that could be true if you are "lucky" enough to get to 1 fps per test and if the ORB will show it in the end, I got 2 frames at the first test and the ORB didn't show any result as if like it didn't run at all
     
  2. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314

    Actually, I was referring to a total of 1 frame, not 1FPS.

    And yes, the benchmark is flawed, as several times times I have personally observed a test render two to four frames, and still give a N/A. It may or may not be possible to get a valid score with only one frame rendered.

    I believe it bears repeating; there is as luck much involved in obtaining a successful result as there is skill.
     
  3. Chri New Member

    Joined:
    Sep 18, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    Location:
    Alps
    I´m able to nearly control the Benchmarks for each frame and can tell you that 1 Frame (not FPS) will always result to N/A.
    with two to three frames it`s a bit tricky, sometimes i get n/a , sometimes i get a valid value.
    the problem is to get them at all four GT`s at the same run :cry:

    how exact after the decimal point does the benchmark calculate?, or is it possible to view the 3dr with an other editor so see more accurate values ?

    :) mfg Chri
     
  4. m00nr4k3r New Member

    Joined:
    Sep 3, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    1st of all , open the results in excel , ull see detialed numbers up to 10 decimals.
     
  5. m00nr4k3r New Member

    Joined:
    Sep 3, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    There is NO luck involed , simply conditions for each test to give a result ,
    and there is no rounding after the summ of the 4 parts , the rounding is before the summ , its a Very different result that way.
     
  6. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    Umm... OK, I understand what you are saying, but I would ask how you manage to control Windows background processes and services to the extent that each and every run is exactly the same?


    Interesting, where did you find this information?

    And how is the rounding done?

    I will give an example:

    http://3dmark.com/3dm03/6440101

    If you take the FPS and apply the multipliers, you get the following:

    GT1-0.12 X 7.3 = 0.876 (from what I was taught in elementary school), I would round that up to 1.

    GT2-0.07 X 37 = 2.59 rounded up to 3.

    GT3-0.08 X 47.1 = 3.768 rounded up to 4.

    GT4-0.09 X 38.7 = 3.483 rounded down to 3.

    1 + 3 + 4 + 3 = 11 ≠ 8.

    If you add up the scores without rounding; 0.876 + 2.59 + 3.768 + 3.483 = 10.717, which would again round up to 11.

    So, it is obvious that the method above is not how the 3D Mark score is calculated.

    Could you perhaps explain exactly how the final score is derived?

    Thanks
     
  7. Chri New Member

    Joined:
    Sep 18, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    Location:
    Alps
    The excel export is only a PRO feature ?!?!? (grey button) , i don`t have a Pro registration :cry:

    mfg Chri
     
    Last edited: Sep 22, 2011
  8. catnipkiller

    catnipkiller

    Joined:
    Apr 14, 2010
    Messages:
    1,493 (0.82/day)
    Thanks Received:
    500
    Location:
    canada
    Benched @ 5.5k i down clocked it a bit and now the sound test wont work but the sound works fine on youtube and stuff. The old card this ibm came with was only dx7 ready lol.
    It clocked @ 1.7ghz atm will down clock it and test it again when i have the time.
    [​IMG]
    [​IMG]
     
  9. m00nr4k3r New Member

    Joined:
    Sep 3, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    go back to page 9 of this topic and read .
     
  10. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    I see, that is why only one frame rendered will not give a valid result.

    Therefore, the lowest score is three, with two frames rendered in each test.

    Thanks
     
  11. m00nr4k3r New Member

    Joined:
    Sep 3, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    U are wrong about it once again.
    just to mention. ...u will not recive result if u only got 2 frames rendered in gt 1 or even gt2.....

    just read the whole topic(s)maybe?

    ofc im happy to help after u done.
     
  12. Chri New Member

    Joined:
    Sep 18, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    Location:
    Alps
    I can`t follow you with your calculation ?

    So with one frame it would end with zero - n/a (0,017/0,5/0,654/0,569) right ?

    calculate with 2 frames: a three (0,173/1/1,308/1,138)

    But but what if the Benchmark for example would be rendered with 1/1/1/2 frames, this should end up to 1 Point ?

    Hmm, where do i have a mistake in this calculation ?

    :confused: Chri
     
    Last edited: Sep 23, 2011
  13. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    After looking at it again, I believe you are correct.

    Since all digits to the right of the decimal point are not used when adding up the total score, it appears that a zero may be possible.

    If all tests render only one frame:

    GT1: 1/42 = 0.02380952380952380952380952380952

    GT2: 1/74 = 0.01351351351351351351351351351351

    GT3: 1/72 = 0.01388888888888888888888888888889

    GT4: 1/68 = 0.01470588235294117647058823529412

    Applying the multipliers:

    GT1: 7.3 X 0.02380952380952380952380952380952 = 0.17380952380952380952380952380952

    GT2: 37 X 0.01351351351351351351351351351351 = 0.5

    GT3: 47.1 X 0.01388888888888888888888888888889 = 0.65416666666666666666666666666667

    GT4: 0.01470588235294117647058823529412 = 0.56911764705882352941176470588235


    As all the results are less than one, and no digits to the right of the decimal point are used, the result would be 0 3D Marks.

    And, GT1 could render five frames and still give a zero.

    I wonder if rendering of partial frames is possible?
     
    Last edited: Sep 23, 2011
  14. m00nr4k3r New Member

    Joined:
    Sep 3, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    exactly why did u ignore the fact that none of the tests will give u result with 1 frame rendered? !
     
  15. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    While I believe what you say is true, and I mean no offense, but I wonder how it is you know this?

    If I understand what you are trying to tell me;

    GT1 will give no result with less than three frames rendered?

    GT2 will give no result with less than three frames rendered?

    GT3 will give no result with less than two frames rendered?

    GT4 will give no result with less than two frames rendered?

    Please be so kind as to correct my mistakes.

    Thank you
     
  16. Chri New Member

    Joined:
    Sep 18, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    Location:
    Alps
    I`m aiming to 3/3/2/3 since a week now and didn`t managed to get them at the same run. :ohwell:
    With only two frames at GT4 i always get n/a.

    so now i think you know my 3DMark Points :D

    Chri

    p.s.: would be great if Techpowerup would list how many people managed to get to the sub 10 Point range after the event. :rolleyes:
     
  17. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    I agree, a list of the scores (with ORB links, if possible) would be nice.
     
  18. m00nr4k3r New Member

    Joined:
    Sep 3, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    Well if u still not understand how u get your scores lets start at the Very begining.

    Run 3dmark03 with commandline parameters.
    3dmark03.exe -verbose
    Ull get a report about the run , open it , this should Tell u how long the actual tests WAS , not like it matters at ALL , and just forget the timmer on screen during tests , will also tell u some more.

    With the numbers in the report thats actually the same as the numbers in excell , just use the math explained before and ull end up haveing the proper numbers.

    PS: RTFM (ReadTheF*ckingManual)
     
    Last edited: Sep 24, 2011
  19. koorosh New Member

    Joined:
    Dec 1, 2009
    Messages:
    39 (0.02/day)
    Thanks Received:
    9
    Location:
    Iran
  20. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    Thank you for trying to help me, I am sure this is all very frustrating.

    Unfortunately, I only have the free version, for which cmdline options are unavailable.
     
  21. m00nr4k3r New Member

    Joined:
    Sep 3, 2011
    Messages:
    14 (0.01/day)
    Thanks Received:
    0
    well u could save yourself a lot of trouble if u not just run the tests and relax but in the meantime u take the effort to read the topic, the manual and the whitepaper, its a great disadvantage to slack on those while also useing the free edition only.

    ofc if u still have questions after u done with the "homework" say so, someone will be able to help
     
  22. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    Thank you.

    I believe I understand how the scores are calculated.
     
  23. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    Taking m00nr4k3r's advice, I re-read the manual.

    Under the Non-comparable Results section, I see this:

    Requirements for a 3DMark03 default score:

    The graphics driver settings should be set for maximum quality, since mipmap bias can othwerwise be downgraded and other similar tweaks producing less than the desired image quality.
    If a separate slider is available for texture quality or mipmap settings, these should be set to maximum quality or 'bias=0', or the value that produces the DirectX default mipmapping.
    No graphics card tweak software may be used when running a default score. The system should be freshly booted, and as many background programs as possible should be switched off or disabled.
    Forced AA or higher quality texture filtering should be turned off, since this might produce a lower score than the system is capable of. All this kind of settings should be set to 'application specific'.
    The graphics driver may not display the content differently than how it was originally meant to be displayed / rendered. For example textures that are not originally compressed may not be compressed by the driver.
    Vertex shaders may not be forced to be run on the CPU in a default run. There is an option in 3DMark03 to run all vertex shaders on the CPU, but the graphics driver may not alter this setting in any way.
    In general, the graphics driver may not have any 3DMark specific settings, it should run in a mode as default as possible. 3DMark is meant to measure general 3D gaming speed, and this is not gained in 3DMark specific driver modes.


    I would like a clarification from W1zzard on this point specifically:

    No graphics card tweak software may be used when running a default score.

    Does this include RivaTuner, NVidia Inspector,et al?

    Also, does this:

    In general, the graphics driver may not have any 3DMark specific settings, it should run in a mode as default as possible.

    mean using nVidia Control Panel to adjust AA, AF, etc. is not allowed?

    Thanks
     
  24. Helltech New Member

    Joined:
    Aug 27, 2011
    Messages:
    21 (0.02/day)
    Thanks Received:
    1
    Arctucas, he already clarified that in post page number 1.

     
  25. Arctucas

    Arctucas

    Joined:
    Jul 14, 2006
    Messages:
    1,814 (0.57/day)
    Thanks Received:
    314
    I know, I just want to be absolutely certain that no 'technicalities' arise.
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Thread Status:
Not open for further replies.

Share This Page