BloodyHell619 Yeah, I confirm, I have faced with the same effect. As soon as I update motherboard painful BIOS or insert "infectious" video card and then I cant repair normal state as before. The only thing that helps is to restore previous state regarding components and firmware and then reinstall windows|OS and in doing so, you have to disable update system even down to disconnecting any kinde of network.

9 days later

You know what I'm just thinking now? Maybe it wasn't the dual GPU trick that fucked up my system in the first place. I got my gtx 970 back from my friend and so the card was on his system with a very recent mobo. maybe his goddamn mobo infected the card and it got passed on to my mobo. If there is 1 thing thats for sure now its that the mobo is the god damn defect.

BloodyHell619 How about the power supply? Try plugging the computer and monitor directly into the wall if you're using some kind of multi extension cord.

Also consider trying a higher or lower monitor cable, test display port vs hdmi if available, or try changing hdmi or display port version on the monitor or bios. After digging into EDID and display driver suff on Mac I realized just how many different types of compression and changes can happen there, based on what bandwith or capabilities the graphics card and os thinks are worthwhile to use.

    8 days later

    maybe his goddamn mobo infected the card and it got passed on to my mobo

    That is not possible.

    A BIOS upgrade itself won't do anything to a video card. However an upgrade can change settings. Have you gone through every page and tried every setting in the Bios? particuarly anything to do with PCI-e timing or anything related to video card voltages

    I've been getting various messages from a couple of games about how my vulcan graphics driver was too old (Win 10 1511, 970 card and driver version 391.55). Now this has been a good strain-free setup for a long time now, apart from it starting to age out re various bits of software and eye strain with a few Unity games.

    But I've got backups… so I thought just this one time I'd try a newer driver… 411 something… uh-oh eye strain feelings detected. So I pushed on trying another even newer driver version 441.12 or similar. It still felt a bit off - those games were running though.

    Sadly after a couple of days I knew this wasn't going to fly with creeping eye strain. Next, I rolled back with a bare metal restore from backup. And… it STILL felt off. Yikes. I've tried a few more sessions and the eye-strain is definitely there and significant.

    Everything's exactly the same as a week before my driver shenanigans. But now there's eye strain. You start to wonder what those newer drivers did to my graphics card.

    So all I can try now is a full ground up reinstall of 1511… (fingers crossed)

    I really am my own worst enemy. I KNEW I shouldn't have messed with my setup. But sometimes you can't help but get so damn tired of all the eye strain everywhere and not being able to run big screens, big graphics cards, and better mobos/cpus.

    But it DOES make you wonder what kinds of settings can 'stick' on your graphics card, much as has been discussed in this thread.

      6 days later

      So I've now done a ground up OS installer rebuild of my PC - exact same hardware and using the same Win 10 1511 OS installer as was working OK for me before.

      Sadly, I am still getting bad eye strain. The eye strain is even evident on a pure 1511 install before ANY (1511 related) patches are installed.

      I am now utterly convinced that a new Nvidia driver HAS changed something fundamental in my graphics card.

      I did try reflashing both VBIOS with the same version so see if that helped but no joy there.

      I'm dejected… and my only hope is that one of my spare 970s doesn't exhibit the same problems. Trouble is they're in live PCs that are running the newest graphics drivers, so I'm not optimistic.

      • glvn replied to this.

        AgentX20

        Sadly, I am still getting bad eye strain.

        I did try reflashing both VBIOS with the same version so see if that helped but no joy there.

        I have one stupid, weakly reasoned idea 🙁

        Reflash the MB BIOS (not VGA) at least to the same version. Some people in this forum claimed that flashing the MB BIOS to an older version helped them. I assume that it could have helped them not to reflash the old version firmware, but the flashing procedure itself (this is based on my old experience, when the firmware update procedure could additionally clear the service areas of the BIOS, like DMI...)

        This is a very big assumption for modern UEFI BIOS, but I have no other assumptions: where which ones can be saved settings when completely reinstalling the OS.

        To be completely sure of the purity of the flashing, it would be a good idea to buy a clean BIOS chip and flash it on a hardware programmer (this is a lot of trouble, don't go down this path if you haven't done something like this before). This may be necessary, since I am almost sure that updating the BIOS in the standard way is does not provide a 100% guarantee of rewriting all UEFI modules/areas. I don’t know if I was able to convey the main idea with my English...

          glvn To be completely sure of the purity of the flashing, it would be a good idea to buy a clean BIOS chip and flash it on a hardware programmer (this is a lot of trouble, don't go down this path if you haven't done something like this before). This may be necessary, since I am almost sure that updating the BIOS in the standard way is does not provide a 100% guarantee of rewriting all UEFI modules/areas. I don’t know if I was able to convey the main idea with my English...

          You're not crazy. I know for a fact in some cases internal settings of the firmware can be persistent across supposed resets, so I wouldn't be surprised if software based flashing doesn't clear everything either.

            JTL

            in some cases internal settings of the firmware can be persistent across supposed resets

            is this fact from the "world of x86" hardware?

            what year is it approximately? (UEFI or BIOS)

            • JTL replied to this.

              glvn is this fact from the "world of x86" hardware?

              Yes

              what year is it approximately? (UEFI or BIOS)

              Hard to say. I think it's a generic property of certain UEFI BIOS's, so starting sometime past 2010 would be my best guess.

              glvn Reflash the MB BIOS (not VGA) at least to the same version.

              I'm not optimistic but it's worth a shot. Thanks for the suggestion.

              • glvn replied to this.

                @AgentX20 Oh boy do I feel you mate. You cannot belive how much I wish I could go back to that day when I stuck that wrecked 970 into my system. My life has been a total wreakage since then. And a huge peice of advice DO NOT STICK ANY OTHER CARD IN THAT SYSTEM. Every card I have put in that system now dithers even if I put it in aonther system!! You will be stuck in a cycle like me infecting card after card, PC after PC. I have tried every possible BIOS version that works with my CPUs on my PC and the horrific dithering is still there. I am convinced that there is now one and only one solution and that is to buy the exact same mobo and vga and replace my infected ones. I do not know what the fuck NVIDIA has done, how the fuck they have made this shit stick so hard in our systems. I messaged them again the other day about this issue and begged them to give me a solution. I even begged them to just tell what in my PC has gotten infect so I could replace it. They no longer even respond. The issue was marked as "researching" for weeks and then it just got flared as "unresolved" without even a word of reply.

                async I have tried all that mate. Nothing worked. different PC with different power supply, different screen, different outlet, ….

                first, a quick Google search showed the presence of the entity NVRAM in UEFI BIOS(not CMOS)

                NVRAM is a special data area that stores those UEFI variables that have the Non-Volatile attribute set. The most popular variables of this kind are Setup, which stores most of the current settings from the BIOS Setup

                you can try to view these variables from EFI SHELL, http://www.opslib.com/2012/01/efi-shell-commands.html

                dmpstore     -- Displays all NVRAM variables

                set         -- Displays, creates, changes or deletes

                default         -- Sets, Resets, or Clears default NVM values

                pci         -- Displays PCI devices or PCI function config space

                maybe the names of the variables will tell you something about setting dithering..

                and secondly (for example), the AFUDOS flashing utility has an additional key /N - clear NVRAM

                This is not an instruction manual, the utilities for flashing and the procedures for flashing the firmware itself are very different for different MB

                  glvn

                  Are you implying it's possible to stop dithering by changing nvram variables ?

                  • glvn replied to this.

                    Liberator005

                    I'm trying to logically understand where the dithering setting can be saved if a clean reinstallation of the operating system does not help with this (and reflashing the VBIOS too).

                    NVRAM - suitable place for this

                    AgentX20

                    "Load BIOS defaults" or ClearCMOS in modern UEFI BIOS should perform some kind of cleaning of the nvram, but I’m not sure that a complete cleaning of this area is performed. You can try this before reflashing MB bios.

                    "I am still getting bad eye strain" Is there this feeling on the UEFI BIOS setup screen?

                      glvn

                      Maybe the dithering setting is just hardwired in the gpu hardware ?

                      Though if the software "ditherig" works (at least with intel gpus), it must be changing the setting somewhere. Could the software "ditherig" have access to nvram values ?

                      • glvn replied to this.

                        BloodyHell should sell their machine, and start over from scratch with a different machine.

                        Liberator005

                        Maybe the dithering setting is just hardwired in the gpu hardware ?

                        Maybe … BUT, AgentX20 say:

                        (Win 10 1511, 970 card and driver version 391.55). Now this has been a good strain-free setup for a long time now

                        I've got backups

                        just this one time I'd try a newer driver… 411 something… uh-oh eye strain feelings detected

                        Next, I rolled back with a bare metal restore from backup. And… it STILL felt off.

                        I've now done a ground up OS installer rebuild of my PC - exact same hardware and using the same Win 10 1511 OS installer as was working OK for me before. … Sadly, I am still getting bad eye strain.

                        I did try reflashing both VBIOS with the same version so see if that helped but no joy there.

                        …seems like dithering was not originally hardwired

                        This is just my guess, but this is how I imagine the sequence of changing dithering as a variable for the programmer:

                        a) the initial value of the variable from the video bios (or video chip hardware), in old cards is not set or is set as OFF

                        b) a variable in the uefi nvram of the motherboard, the initial value is inherited from point a) or is forcibly set to ON by “new” motherboards because it is useful for “image beauty”. This can try to explain the fact that “good” video cards perform “good” on old MBs but “bad” on new ones (including even before loading the OS).

                        c) a graphics pipeline consisting of the OS(version), graphics driver, DWM, registry, color profile settings and M$ knows what else…. Or perhaps it inherits the initial dither value from b), or perhaps new operating systems are simply forced to be set to ON for “image beauty” on modern cheap (6-bit) monitors. This can try to explain why dithering is enabled more often on “new” OS/drivers. Why the state of the variable is then stored in b), (if this is at all possible), I cannot explain. This is all despite the fact that MS/Nvidia technical support denies the very fact of dithering.

                        .. too many "perhaps" (

                          Liberator005 Could the software "ditherig" have access to nvram values ?

                          as far as I understand - NO

                          this software only changes the state of the current graphics pipeline( point c) ), and only until the next reboot (or video mode change??)

                          p.s. "access to nvram values" from os/driver - just my guess, but how else can you explain the persistence of dithering after a complete reinstallation of the OS?

                            dev