[Help] Windows 10 Woes and issue (still)

So if you recall I made a thread some time ago asking about a VM installation running a raw disk partition. Well that has since been written off as "do not do" due to what it requires.

The symptoms
-System halts, freezes for about 10 seconds, displays glitch for a fraction of a second, go black, reshow whatever i was doing
-Occasionally corrupts applications visually (severity varies, no idea why)
-Keyboard, Mouse, probably USB HDD halted until visuals are restored
-Unpredictable
-Usually happens with games or GPU activity (has happened without them)
-WIndows 10 exclusive

What I have done
-Latest updates, major ones included show no change
-Extreme setting changes on GPU CPU and RAM overclocking (anything from underclocked to maximum i can go and still be stable with anything in between)
-Power supply upgrade (have been told it might be relevant, but my draw is less than this even on full load)
-Have allocated more and less power to anything that might have any bearing on this
-Have changed what ports things are displayed on (except DVI, not bothering with it)
-Have made a separate installation to test, problem still occurs on windows 10 after a fresh install (post debloat scripts running)

What I do know
-Linux doesnt have this issue
-Is likely due to running the debloat windows 10 scrips (here: https://github.com/W4RH4WK/Debloat-Windows-10)
-Using anything older than windows 10 has some issue with Ryzen CPUs because Microsoft or something. It can work but it's just annoying (lack of updates isnt an issue)

What I want to try
-Windows 8 installation
-Windows 7 installation

The problem with what I want to try
-Windows 8 isn't up to my standards, but I'd use it if I had to, is easier to install
-Windows 7 lacks USB 3.0 driver support, and due to my mobo having *all usb 3.0/1 ports* I am left with only PS/2 access (not a problem for key and mouse)
-If i want to install them on my SSD raid I need to install the drivers, which cant happen on win7, but can on 8, or I put it on my spare HDD used for console related affairs.
-I dont want to run untouched windows 10. Fuck that.

What I want to know
-Where the fuck this problem is coming from software wise
-If I can fix it without reinstalling or having vanilla 10, what do I do for that
-Is there a method to migrate all my current installation over to a different windows 7/8 installation without having to claim ownership of every single thing in the drive (like leech off the other drive i guess)
-Would having a windows 7 and windows 8 installation, each on 1 ssd be a better choice (currently only league of legends is installed on it in raid 0 til i decide what to do with it)

The goal is to not set myself back to reinstalling literally everything

Specs if interested

 

You can 'refresh' your PC. You can choose to keep your files. Only way to restore Windows 10 to vanilla without a format.

 

Sounds like an issue with the graphics card or at least the driver. What you describe sounds like when the graphics driver has crashed and recovered.

 

Which makes sense but the consistency is what I'm looking at. The card is fine as told by RMA over the phone, the same happens with my 750ti which I should have mentioned but it recovers near instantly due to power consumption difference.
It wouldn't be a problem had it been maybe once or twice but constantly over 8 months is getting to me.

Driver wise I'm sitting on the latest for all of AMD's stuff (chipset, optional revisions, not ReLive because it sucks), to which older revisions had this issue as well. Do you think I should give a Win7 installation with the same software a shot?

This is one option I considered but have tried avoiding due to potential loss I might have. My system is layers and layers of bodges and patching of anything messy to "just work" so that's rubbing me the wrong way whenever I consider it. I can rebuild it but my data means a lot to me.

For example, a recent issue no less
Was trying to install windows to my SSDs. It failed every time. No matter what version I used. Hours later I tried removing all other drives and gave it a shot. It suddenly started accepting the SSDs in RAID and in AHCI. Why? I dont know and I dont care much to find out. However, in doing so it reset my grub bootloader so I had to repair that for the third time.
Needless to say it's a mess.

 

I had similar issues with SSD as well. I 'HAD' to do a full install, I suppose when I cloned from HDD there might of been something fucky when it was SSD trimmed.

I know you're wanting to have a scrubbed W10 but honestly all those things end up screwing you in the long run. Considering you've got lots of things going on it's probably best to start again.

 

Installing Win10 without bootloader is easy, boot from media open a command prompt and use dism to restore the install wim to the partition you want.

Unfortunately, you cant moan when doing "non standard" stuff, you are expected to fix your own stuff or know how to do it.

Best way to rule out if your install is an issue? put in another hdd and install windows on it, see if fault goes away or not. If it does, reinstall your "proper" system. If not, you saved yourself some grief.

 

That's not true; Windows 7 supports USB 3.0, you just have to install drivers for it. I know because I've used USB 3.0 drives with Windows 7 and they worked at full speed like they're supposed to.

 

If the mobo is for a CPU that only supports Windows 10, you may find theres no Windows 7 drivers.

But it could just be something with a chipset you can find a driver for elsewhere YMMV.

 

Does this always happen, or does it only happen after the card is installed for a while? About a year ago or so, my PC was starting to fail to detect my GTX 650Ti. Reseating the card made the problem go away for a while... and then it became quite clear to me that the card was creeping out of the PCIe slot as things expanded due to heat. Even after cooling down, the "hardware issue" wouldn't go away until the card was re-seated.

I solved it by applying new thermal paste to the GPU (but of course, my hardware is much, much older!).

On Windows, the driver would sometimes crash. On Linux, it would boot fine... except that X sometimes fail to start or becomes unstable.

 

Installation wise no, post install yes. There's drivers for USB 3.0 from my mobo's manufacturer but the issue then is i still need to use a PS/2 kb/mouse to load it in the first place. Post install it's not an issue what so ever. The issue only existed when it was related to setting up RAID SSDs for it since loading the drivers wouldnt work without 3.0 drivers. I'm aware there's other methods but eh its minor at best now.
Intel offers a tool that can modify a live USB installer to add some drivers but after trying it (and probably wrong) I didnt bother too much.

First time ever I've had this issue and persisting for so long at that with windows 10, so in my defense this is not the usual for me. I've been using win10 since the initial leaks, even when it had this nasty bug with explorer which opened and crashed explorer forever, so I ended up operating it with task manager til the bug was fixed. My old system (i5-4440, 750ti) used a stock windows 10 and never had the issue, but this was also when I stopped updates which was some old ass build of win10. I bet if I updated it now it would probably have some issues.

I'll be doing none of the sort in the end. At the very most I'll have win7 for gaming, win10 for whatever the hell else I do if it works out in the end. I'll still try the refresh option but I know for a fact I'm not going to like it. The telemetry shit is not sitting well with me on the paranoid side. I already cant stop the Xbone I have and it's turned on accidentally from my headphones a few times so it's already something I hate having at all. I like my privacy enough to fight it.

It happens on any card and seemingly has no real indication of why. Heat hasn't ever reached higher than 50C, and I have my fans on that at 75% constantly because it's already hot enough in my room. Lately even when I decide to mess around with eyefinity and do 7680x2160 with the super resolution (it's fun and I used it for documenting an engine limitation) it's never exceeding even 45C even then. I've had multiple things going at the same time where it struggled to go even past 42C which to me is surprising at all given what I was doing.
Heat wouldnt be an issue I feel. Reseating the card multiple times hasn't really done much for me. Even my 750ti I used while my rx480 was on RMA was doing this, even lower temps, but was resolved far faster than the rx480.
Speaking of fans, I have 4 different fans for airflow, not including gpu/cpu fans. 2 Intake in the front, 2 outake put to the top and the back. I wouldn't consider the heat as an issue but I wont rule it out either.

I WILL mention this however. The card is somewhat heavy. Initially it had some sag on one side as it was installed. It bugged me a lot and now it's not sagging or its enough to not notice it. Is it possible this is contributing in some way.

My mobo DOES have win7 drivers, but how long that will last for is up in the air. Some users mentioned issues with updating on win7-8, some mentioned chipset driver wouldnt install, stuff like this. I dont see it as an issue in the long run and can be worked around at the moment as well so there's that too.


This problem has gone on for a long time like I mentioned, and I'm willing to put more time into it if I must so any suggestions are fine, but when I'll get to them is up in the air.

 

If there is an option to disable XHCI support, you could try that first. If there is an option for legacy USB support, you can try that too.

Is there another PCIe x16 slot that you can test with, in case this one is somehow faulty?

If you mean Windows update, I don't think you have to worry so much because Windows 7 is approaching EOL and Microsoft doesn't seem to provide new drivers. Anyway, they've always seemed to be very slow at updating their available drivers.

The USB standards (UHCI, OHCI, EHCI and now XHCI) allowed for standard USB drivers across all hardware manufacturers, so having Windows replace the USB driver may not cause problems (since it's supposed to be standard). In comparison to it selecting the generic NDIS driver for the Ethernet adaptor.

Windows 10 is using the generic XHCI driver for my Intel-based laptop.

 

hmm sounds like display driver crashs. Current drivers have issues with random crashs on RX 480.

Try use version 17.9.3, its running great for mine RX 480 with fall update win10

 

@PixelButts I wouldn't worry about W10 communication things. Nothing more than your phone would trace anyway.

Shouldn't stress.

Probably updates break whatever you've got modified, which is what I believe Bad was talking about as it's not officially supported.

 

In my opinion a core of the troubles in the hardware malfunction or driver software failing. Linux may uses different drivers software for your hardware, that runs in "safe mode" to prevent unstability. And Windows OS uses full acsess to your hardware via official drivers which somehow fails because of...? Just guessing.

If you familiar with PC repairs you have to test everything from the scratch to be sure what excacly cause the issue, regular routine diagnostics. So you will need basic stuff - PC test card, pyrometer, multimeter, service docs for your motherboard if available and patience (sometimes oscilloscope). First thing you have to do is disconnect motherboard and power supply from everything else. The only parts you will need for safe start is: motherboard, CPU (and CPU cooler installed), PSU, RAM (only one of them), PC test card and 1 monitor. Everything else should be disconnected for a first test. Also reset bios settings by pulling off battery for some seconds.

The whole point of the first test is to make sure that your PSU has correct voltage on pins and on all important motherboard components. If tests were passed turn it on and wait for about 30 mins without booting to any OS to monitor temperature on all the motherboard components to check if they are ok all the time during test, also run memory test for RAM from usb drive or something. If ok, now connect only 1 ssd/hdd that has windows installation (or install Win 10 again) and fill out all the RAM slots as usual. Test everything again for 20-30 mins, but now you can also run benchmarks, stress tests and heavy games/software. Then connect external GPU and test it all again with heavy benchmarks and monitoring its temperature. In this point you will probably find out where system starts to fail (if it not CPU fail). You can try to replace those components and test again. IDK maybe you know all this stuff already. i'm not familiar with detailed tracking and logging driver software fails under linux or windows, but somewhere must be tutorial.

Maybe these will help:
ifixit.com/Wiki/Troubleshooting_a_Computer_Motherboard

Spoiler

 

I'll give it a shot, i doubt it will impact me in any positive way.

And you're right but that doesnt stop it from being more up my ass than my phone.
I will worry for a number of reasons, and not just the services that destroy the OS performance which is my main focus.

 

Understood. However any MS service I honestly won't stress about.

 

Small discovery, not sure how impactful it is but I'll mention it
Decided to kill Windows Shell Experience from my processes (which restarted on its own as I later saw), and the moment my GPU started doing it's GPU stuff on all 3 displays some form of corruption happened.
Once loading into my game it did it's usual freezing/screen resetting. I want to stress that this is the second time I've ever experienced this corruption, the last time being about 5 months ago (which i incidentally had recording at the time it happened, I can link that footage if anyone wants to see it).

So I guess give it to me bluntly. Card dying or not.
Secondly, should I yell at Asus once more since the first RMA, the one they told me "everything is fine".

 

Do what Bad mentioned and install W7 on another Hard Disk. You'll have to do an isolation test to figure this out. None of us could tell you if the card is dying or not.

 

Then I guess I'll do that over the next few days.

Here's that video while I have the moment

Spoiler

 

You wrote that your previous GTX750Ti was also doing that, didn't you? So I have this feeling it is not a problem with the card (what are the odds that two cards give similar faults), but may be something related to the mainboard. Which was why I was also suggesting that you could try a different PCIe x16 slot, if there is another slot on your mainboard.