Jump to content
HWBOT Community Forums

hansglans

Members
  • Posts

    17
  • Joined

  • Last visited

Posts posted by hansglans

  1. 19 hours ago, sokdak said:

    Here's some test result about 1T and 2T on z490-i unify.

    either 12Z and 12T bios cannot boot above 4300 freq.

    and i've found some strange that 1N(Fake 1T) option mentioned as equals to N:1N(N=3), but latency is not the same.

    KakaoTalk_Photo_2020-09-27-12-09-03.png

    Amazing, thanks for your effort! Do you mind sharing which ram and voltages you use?

    I believe N=3 and 1T are actually the same and only differ because of benchmark variation. 1T here is probably the "old" 1T which has N=3 by default.

  2. 9 hours ago, Sparky's__Adventure said:

    @hansglans Technically, it's not necessary, but if you want your computer to function properly it is. I ran into significant instability when not using it. I have a theory that it trying  to boost and throttle as it normally would while trying to use my manually set voltage, which would crash at the lighter threaded turbo speeds.

    When using the patch, my settings would work fine. Unfortunately, that patch isn't persistent between boots.

    I'm not sure about the transferrability of OC profiles, I wasn't able to save the profiles from my non-modded Dark bios anyways.

    Thanks, can you show what exactly the task does?

  3. 8 hours ago, Sparky's__Adventure said:

    @yaqy123 I can make a bios for the atx, but if you'd like to do it yourself:

    Download and unzip all the things I provided. Open the bios image in MMTool. Go to CPU Patches and delete all of them. Save image. You can now flash that bios and it should work just fine.

    Once you're in OS, go to thing I have as "tool" and look for a task called "patch for no-ucode." Then go to task scheduler and schedule that patch to run on every login.

    For best results, you want to also be using InSpectre to disable in-os mitigations.

    Please report results.

    What does the "patch for no-ucode" do? I have task scheduler disabled, is the tool really necessary?

  4. 19 hours ago, technikswd said:

    ….."Now we need true 1T by having them set "Command Rate Support" to 0".....

    what do you mean - "set "Command Rate Support"?

    Where is that Option? 

     

     

     

        Command Rate Support | VarOffset: 0x212, VarStore: SaSetup (0x16)
              Disabled: 0x0
              1 CMD: 0x1
              2 CMDs: 0x2
              3 CMDs: 0x3 (default)
              4 CMDs: 0x4
              5 CMDs: 0x5
              6 CMDs: 0x6
              7 CMDs: 0x7

    It's a hidden setting changeable with Grub. I set it from 0x3 to 0x0 and see some improvements:unknown.png?width=1440&height=588

    The weird thing is that MemTweakIt still shows 0x0 as ?T (or  N:1T) - it should show 1T.
    The test on the right is after setting it to 1T in MemTweakIt.

  5. https://drive.google.com/drive/folders/1eD1QxbQfun1M5tMYvPyxMUc5yueZLN3S

    New v12Z shared by Toppc Lin. It has N:1T support (fake 1T) and Power Down Control for ITX. (ATX still not fixed afaik)
    Now we need true 1T by having them set "Command Rate Support" to 0.

    Note: MTI = MemTweakIt.
    Bios CR1 equals to "?T" (or N:1T) in MTI.
    Changing "?T" to "1T" should change it to actual 1T until reboot, but not sure if it's actually applied as the difference is small.

    1Tbios_TMTI.png

    1Tbios1TMTI.png

    2Tbios2TMTI.png

  6. https://drive.google.com/drive/folders/1eD1QxbQfun1M5tMYvPyxMUc5yueZLN3S

    New v12Z shared by Toppc Lin. It has N:1T support (fake 1T) and Power Down Control for ITX. (ATX still not fixed afaik)
    Now we need true 1T by having them set "Command Rate Support" to 0.

    Note: MTI = MemTweakIt.
    Bios CR1 equals to "?T" (or N:1T) in MTI.
    Changing "?T" to "1T" should change it to actual 1T until reboot, but not sure if it's actually applied as the difference is small.

    1Tbios_TMTI.png

    1Tbios1TMTI.png

    2Tbios2TMTI.png

×
×
  • Create New...