Rivatuner 12.1 + SoftR9700 released!

Discussion in 'RivaTuner Advanced Discussion forum' started by Hilbert Hagedoorn, Jan 13, 2003.

  1. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    No problems, English is not my natural language so I realize that I can make mistakes and don’t feel shame for that. If you wish, we can switch to my natural language (Russian), mr. Polyglot. ;)
  2. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    I'd like to see results of SoftR9x00, not Wizzard's drivers. But I don't think that it's really matter coz none of sw modded drivers can cause such thing even in theory on the GPU with non-faulty HSR unit.
    Is there any difference in performance with regular driver, wizzard's driver and wizzard's driver with disabled HZ?
  3. Robbeh

    Robbeh Guest

    Unwinder, could the strange results pussycat is experiencing be be caused by some sort of protection in the ati drivers? Anyone with a 9700 want to try soft modding to see if they also get checkerboards?
  4. Molestro

    Molestro Guest


    What Brand do you guys suggest I purchase for the 9500 non-pro. Which one for sure has the 256-bit memory? Assistance in this matter would be greatly appreciated.

  5. ectx

    ectx Guest

    Re: Must be blind

    First, to answer your questions:

    Yes, so far (before ATI forced the manf. to not use 9700pcb) all 9500NP's can be modded (of course, some may have artifacts, so before you buy one consider ymmv).

    Now, are you saying that you applied the script to a 9500pro and get artifacts? Did you overclock when you get them? Did you experience artifacts w/o the script????

    PLEASE CONFIRM THIS. Because, according to this article:

    Excellent read from Digit-Life - Search for HSR to see it

    The HyperZ for 9500Pro's should be functional - there should not be artifacts due to defective hyper z unit (or hierarchical Z unit??)!! If you mod a 9500Pro, there should be no performance increase (since it already has 8 pipelines) and there should be no artifacts due to the eabling of the hyer Z functions!!!!!!

    This would support Pussycat's claim that there is driver incompatibilities...

    Folks, before you jump up and down and accuse me of being too stubborn to believe in the "truth" (that the artifacts are due to ATI's using defective chips), please attack my logic - no personal attacks. Let's have some real discussions. As far as I am concerned the jury is still out and the fat lady has not sung yet.
  6. FinalFantasy

    FinalFantasy Guest

    Ectx , are you "naturally stupid"???. Hehehe, I was just joking. Please do not take it seriously.

    Ectx, the fact that SOME people with the hardware mod get those artifacts, and the fact that NOT all people with the software mod get those artifacts means one thing and ONLY ONE THING:

    If it was a driver problem then everybody must experience the artifacts.

    You should defenitely go and take some classes in philosophy and logic.

    No body knows what kind of miss pussycat or molestero may have done with their system configuration and the drivers before they got the artifacts they claim they got.

    And let us be logical again please: molestero did not even know that soft9700 does not work on 9500pro, then how could he ever know how to install soft9700 properly??? ( No offence meant).
  7. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    Good news

    Good news, I've disassembled the OpenGL driver and found the way to disable HierarchicalZ via the registry in OpenGL too (ATI encrypt registry entiry names in NV-styled manner now, that naive guys think that it will help 'em to hide it :)).
    Registry entry name is An6aYj2vwKuKd9brxl7
    Registry entry type is REG_DWORD
    Registry entry handling algorithm: 0 disables HZ, 1 enables HZ, hardwired to 0 on R9500.

    It's located in the driver's base key in \OpenGL\Private subkey. If it's not a problem for you - just add this entry to specified subkey, if you're not sure that you can locate the driver's key correctly - I've created small database for RivaTuner ( http://www.nvworld.ru/temp/HZDatabase.rar ), which allows to control HZ in both D3D/OGL on Catalyst 62xx under Windows 2000/XP.
    Just download the file (~300 bytes) then open it via RivaTuner's PowerUser tab (Open database button). This will allow you to edit HZ related registry entries directly via PowerUser tab.

    I'm waiting for replies from the users who have checkerboards in OpenGL.
    Last edited: Feb 4, 2003
  8. ]SKyRuNNeR[

    ]SKyRuNNeR[ Guest

    Re: Brand

    every 9500 non pro with 128 ram has 256 bit memory bus
  9. Molestro

    Molestro Guest


    Yea I should not be taken seriously at all seeing as how I didn't even know what god damn card the script even applied too :)
    However I am in the process of getting a 9500 nonpro and I did apply the patch correctly to the 9500 pro, gained no performance understandably, and did get artifacts from using the patch. No I did not overclock it, but the script set the frequencies to that of a 9700. That also seemed to have no positive effect. Then, from reading on the forums I pieced together something about rTool and hyperZ and screwed around. When I disabled hyperZ the artifacts disapeared, and once again same performance. All of this should be fairly redundant at this stage seeing as how I don't even have the damn card you are all talking about but my question is this: Aside from flashing the bios (or does that even work) is there a way to overclock the 9500 pro at this stage?
    I tried a series of programs all unsucessful, the only one that "came close" was powerstrip, which gave me the like 1 more fps. And if I went to far with powerstrip the card faught back with a vengance (distortion-o-rama). So if any1 has been sucessful in overclocking the 9500 PRO could you please tell me how it was achieved? Thanks in advance.
  10. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    Re: Re: Brand

    AFAIK Sapphire stopped manufacturing 256-bit R9500's, new 128MB revisions must be equipped with 128-bit bus.

  11. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    Re: Overclocking

    I'm starting getting tired of the users fantasies. WHAT are you talking about?! Script set you new clock frequencies?! Sorry, but it's full bull****.
    And if you really get rid of 'artifacts caused by patch' via disabling HyperZ (it disables not only HierarchicalZ but all bandwidth saving techniques including Z-buffer compression), you _must_ loose performance.
  12. vtsantzalos

    vtsantzalos Guest

    Re: Good news

    Could someone tell me which drivers are the Catalyst 62xx? I still havent figured out which one is which. Thanks for your time.
  13. Creig

    Creig Member

    Likes Received:
    Great job on the OpenGL find, Unwinder. I see how you got your screen name now. Even though I was one of the lucky ones to get a fully functional 9700 out of my 9500 it's nice to see there's still hope for those who received cards that show artifacts.

    Way to go!
  14. As I sayd, WFP was undoing the patching.

    Your FAQ states that the drivers are identical, however. Are they not?

    Q: Is there any difference in performance between 
    distributed patched driver and RivaTuner's patch script?
    A: No. Both of them give absolutely equal performance .
    Well, it did. I can't explain it either.

    Well, as I sayd I have uninstalled the driver. I will see if there's still a recovery point to get back to the old situation.

    Currently, it's:

    Dragothic high detail 1280*1024*32*6xAA + 16 X Aniso @ 60 Hz, vsync off:

    44,4 fps w/ HZ,
    42,3 fps w/o HZ.

    Both valeus tested twice, benchmarks right after one another.

    Same setting, but with 2xAA / 8 x Aniso / HZ on gives 67,5 FPS.
    Last edited by a moderator: Feb 4, 2003
  15. Final, I did get those artifacts. I do not make them up, neither am I unexperienced with PC's. I am not even a miss.
    I am also very interested in politics and I am very good at logical conclusions / things that are not. I am studying at a university right now, and that is NOT because I am stupid / |\|008 / whatever.

    Now we've got this settled, I don't understand the problem either. I am thinking of some driver incompatability with certain board types. Since it's also hardware-caused, I do not think it's unwinder's fault.

  16. merou13

    merou13 Guest

    YES! YES! YES!

    Unwinder > thank you, my lord

    Your last proposal with the rivatuner database is perfect. It removes the last artefacts, I have in opengl games and demos. Thouch my artefacts are not checkboards but a line of 12-16 dots (5*5 pixels), they have vanished now. I have already get rid of my artefacs with Rtool in direct3D games.

    Great job!!!

  17. lopoetve

    lopoetve Guest

    Re: Good news

    I DL it, and winRar says the file has a CRC error....

    Any other place I can get it? I'll paypal someone $2 for the file...
  18. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    Re: Re: Good news

    It looks like you just have old version of WinRar ;)
    Anyway, database is a plain text file so you can create it yourself. Just create text file named "Catalyst 62xx w2k.rtd" and add the following lines to it:


    Name = Catalyst 62xx Direct3D settings
    Key = %DisplayDriverPath%

    Value0 = HierarchicalZEnable
    Param0 = 1,system dependent,,NotAvailable.rth


    Name = Catalyst 62xx OpenGL private settings
    Key = %DisplayDriverPath%\OpenGL\Private

    Value0 = An6aYj2vwKuKd9brxl7
    Inter0 = 0,false,0,true,
    Param0 = ,system dependent,HierarchicalZEnable,NotAvailable.rth
  19. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    Re: YES! YES! YES!

    Cool, I'm glad that it really helps to remove artefacts in OGL. Now, after getting confirmation that it really works I can either include the database for Catalyst in RivaTuner to allow contolling HZ in both D3D/OGL or create the script wich will block HZ directly in the driver.
  20. Unwinder

    Unwinder Moderator Staff Member

    Likes Received:
    They perform the same task (override PCI DeviceID at the driver level) but in different places of the miniport's code. However, I'd like to exclude any side effects and ensure that it will also happend with RT's script.

Share This Page