Announcement

Collapse
No announcement yet.

AMD OverDrive Overclocking To Finally Work For Radeon Navi GPUs With Linux 5.5 Kernel

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Originally posted by Danny3 View Post
    Still disappointed that there's no official control panel so we can control this as human beings instead of this command line crap.
    That is the community job to step in making such control panel using human interface guideline like the author of WattmanGTK did.

    Comment


    • #12
      Originally posted by faldzip View Post
      That's cool. I am using those features with my RX570 as it normally clocks even 200mhz less than max stock clock in Witcher 3. The thing is, that increasing power1_cap doesn't work as expected. With the 0xffffffff od features I am able to set it to 180W, and while it exceeds then the default 120W, it is not rising over 130W while it drops a clock and temperature is totally ok (not exceeding 70C). Has anyone faced such issue?
      PS. On Win 10 it easily reaches 174Wat full clock.
      I have strangely experienced something like this on Windows 10 with a Vega card. It was reaching 100% clock (cores and memory) when I first got it, but then after changing some settings in the official Radeon control panel software it was always hanging around 80% or less clock in all games. I restored the settings to defaults and made 100% sure the powersaving software features were all disabled and it didn't fix it. The only thing that fixed it was uninstalling the software and using the official Radeon tool to remove all traces of the driver and its settings from my system. After a re-install it was fixed.

      I only mentioned this, because it seems like Radeon software is sometimes a bit buggy in very strange and unintuitive ways. If I experienced that terrible issue on Windows 10 (their primary desktop platform of interest), then I wouldn't be surprised if their were equally buggy & unintuitive behaviours in their GNU/Linux drivers. In your situation, I'd find out where the Radeon driver stores data (if anywhere) and delete it (moving it is better). I'd also remove any custom tweaks you may have made (e.g. ones stored as kernel params or as scripts in something like rc.local that get run on boot).

      Comment


      • #13
        Anyone know if this will support fan-curve adjustment? I really love having the ability to have the fans turn off completely when the card is idle. My Windows 10 system is configured to do that with my Vega card.

        Comment


        • #14
          At least on Linux, you are not forced to suffer screw ups by some bugged GUI. Windows forums are full of people having issues with Wattman, or be it just its terrible custom fan control feature which doesn't interpolate between steps and has no hysteresis.
          I much prefer to control things via sysfs. Once you set everything up, it's basically immune to breakage and other bugs. And it's not that hard after all.

          Btw. also MUCH better than the Nvidia crap, who offer no custom fan curve without performance issues and allow no voltage modification at all.

          Comment


          • #15
            Originally posted by aufkrawall View Post
            skeevy420 Setting power_dpm_force_performance_level to manual is not necessary, as long as you don't want to have additional control like locking of single pstates etc (which the driver probably still forgets after suspend etc.).
            You also waste some power at low load by giving all pstates from 3-7 the same voltage. You also only need to specifiy the pstates you want to alter.
            I know. Except for the 0 states, everything is tweaked from stock with a lowered voltage (my GPU does not like the 0 states tweaked at all).

            The same voltage numbers on 3-7 is because I was doing the lower voltage by 5mv, stress test, play a game, repeat until crash and got state 7 to where state 6 was, got those down to 5, got those down to 4, and you see the results when they all got to 3. By then, the reason I was undervolting my card didn't matter anymore so that's just where I left it...it ran very, very hot and thermal throttled itself into crap performance whereas now it runs like a champ and doesn't get hot enough to throttle.

            One of these days I'll eventually finish that process and dial in the minimal necessary voltage for 1080p.

            My memory speeds don't look overclocked, but they are. Mine was one of those 580s that came with 1750 instead of 2000 so my overclock is what everyone else calls stock.

            Comment


            • #16
              Originally posted by cybertraveler View Post
              Anyone know if this will support fan-curve adjustment? I really love having the ability to have the fans turn off completely when the card is idle. My Windows 10 system is configured to do that with my Vega card.
              This in particular is not effected by these patches, but I (the author of these patches), put in a separate patch to enable sysfs fan control a month or two ago, so your standard fan control methods should work fine!

              Comment


              • #17
                Good day, are pp_sclk_od or pp_mclk_od still valid? You should now that custom pstates may destroy your idle power consumption.

                Comment


                • #18
                  I've managed to undervolt my MSI 5700 XT Gaming X using upp so it is working as I had hoped (less power usage, a lot cooler and quieter but only a few fps slower). However, the voltages reported via upp are 4 times what the actual values are supposed to be (at least what seems to be reported under Windows according to online screenshots ). From pp_table dump:

                  smcPPTable: [...]
                  MinVoltageGfx: 3000
                  MinVoltageSoc: 3000
                  MaxVoltageGfx: 4800
                  MaxVoltageSoc: 4200

                  I've set the MaxVoltageGfx to 4400 (-100mV or 1100 mV from 1200 mV supposedly). I'm worried that future updates elsewhere to the way pp_table handles voltages (starting to use the actual values) may cause the upp settings to damage the card if the values are set several volts higher than they should? Tested under amd-staging-drm-next (drivers via oibaf) as of today (2019-12-27), didn't test to set any values but the same values were reported under kernel 5.5-rc2.

                  Comment


                  • #19
                    Good day again, is OC by percentage still valid?

                    Comment


                    • #20
                      Originally posted by artivision View Post
                      Good day again, is OC by percentage still valid?
                      Afaik this is not possible anymore since pstate overdrive was implemented.

                      Comment

                      Working...
                      X