Support NeoGAF

[Rudimental]

I will do this tomorrow. Only reason I hadn't earlier was because it was in the middle of the CPU test. Appreciate it - if some dumps show up I'll PM them to you if you don't mind.

Yeah, for sure. I'm roughly in the European timezone so might be a delay due to that, otherwise shouldn't take long to look at. Believe me, I know how frustrating trying to troubleshoot without having spare hardware to swap out is, which is why I downloaded the debugger and learnt to use it to read crash dumps.

 

[Resilient]

Ok so Unigine ran for about 4 hours, then it froze, but I could still use the PC, I could alt tab but it kept bringing me back to the Unigine frozen screen of the dragon. So I created a second desktop, opened task manager and ended the task from there. I'm taking that as a pass for the GPU.

Sounds like the issue's the PSU. Switch it out.
After that, I'd start testing DRAM modules/slots. I doubt it's your GPU (you basically tested that) or CPU (rare).

Stop fucking around with it until you get the new PSU to test it out. You're playing with fire.


Yep. I'd say we're at a good 70% chance of it being a PSU issue.

I'm gonna swap the PSU out now and just use the PC as normal. Then I'll see if it pops up again and do some more testing from there. Not sure how to test DRAM modules but I'll google it.

 

[adamantypants]

So it seems like you're turning off the CPU OC while you figure it out. OC is a hairy variable to throw in even if you know it's a solid overclock.

I had a similar experience with my PC. I had a solid overclock but would freeze when playing games. Turns out, it wasn't the powering up of the cpu that caused the issues, it was the powering down. The PSU will pump appropriate voltage into the CPU under load but powers down way too much when the CPU begins to idle. The CPU get starved for power and shuts down. Double check that your voltage doesn't drop too much, I can't really remember the settings in my bios but there was some gnarly sub menu I got real familiar with.

 

[Resilient]

So it seems like you're turning off the CPU OC while you figure it out. OC is a hairy variable to throw in even if you know it's a solid overclock.

I had a similar experience with my PC. I had a solid overclock but would freeze when playing games. Turns out, it wasn't the powering up of the cpu that caused the issues, it was the powering down. The PSU will pump appropriate voltage into the CPU under load but powers down way too much when the CPU begins to idle. The CPU get starved for power and shuts down. Double check that your voltage doesn't drop too much, I can't really remember the settings in my bios but there was some gnarly sub menu I got real familiar with.

Unfortunately, the PC still freezes when playing games with the OC turned off. It passes the stress tests fine but even in game, still persists. I haven't swapped the PSU out yet. Honestly I thought it was the OC too but the bastard crashed 4 times in one night with no OC. I've never ran an OC on the GPU though, that's always just been stock.

 

[JaseC]

Have you tried increasing the CPU voltage even when running at stock? Your experience mirrors mine when I tried in vain across several weeks to maintain 4.3GHz on my 6800K. As the instability was completely and utterly random (sometimes occurring after minutes, other times several days) and never reared its ugly head during stress tests, only in games, there came a point where I believed I'd successfully identified the PSU as the culprit. Once it became abundantly clear that a stable 4.3GHz wasn't going to happen without breaking the 1.4v barrier, I dropped down to 4.2GHz, slightly reduced the voltage accordingly, and haven't had any issues since.

 

[Resilient]

OK, so, new PSU is in. Had another damn crash when gaming/testing another issue (if Geforce Experience was the problem), so I've slotted it in.

Tried:
Turning off CPU OC - crashed when gaming.
Swapping SATA ports around with SSD/HDD - crashed when gaming.
8 hour prime95 28.10 stress test; no errors found
1 hour Unigine Heaven stress test; crashed, took the PC with it
4 hour Unigine Heaven stress test; crashed, didn't crash the PC
Turning off Geforce Experience - crashed when gaming

Haven't tried:
Testing DRAM
LatencyMon (should have done this sooner, sorry LilJokka)

Let's see what happens with the new PSU..

 

[oxrock]

OS on ssd and games/media on hdd?

It sounds to me like you mostly or only crash when gaming, have you checked the drive your games are on to see if there's errors? Running CHKDSK can't hurt at least and could possibly lead to a solution.

 

[BMX Bandit]

OK, so, new PSU is in. Had another damn crash when gaming/testing another issue (if Geforce Experience was the problem), so I've slotted it in.

Tried:
Turning off CPU OC - crashed when gaming.
Swapping SATA ports around with SSD/HDD - crashed when gaming.
8 hour prime95 28.10 stress test; no errors found
1 hour Unigine Heaven stress test; crashed, took the PC with it
4 hour Unigine Heaven stress test; crashed, didn't crash the PC
Turning off Geforce Experience - crashed when gaming

Haven't tried:
Testing DRAM
LatencyMon (should have done this sooner, sorry LilJokka)

Let's see what happens with the new PSU..

Post back if it fixes it. Always nice to have resolutions in these kinds of threads for future users experiencing similar issues.

 

[Liljagare]

Any USB spheripals? Had a issue like this that was driving me mad, noticed when I disconnected the wireless keyboard + mouse that it went away. USB3.0 conflicting driver, had to update BIOS.

Longshot, but I too hate when this sort of stuff happens..

 

[Sevenfold]

Post back if it fixes it. Always nice to have resolutions in these kinds of threads for future users experiencing similar issues.

Nothing worse than Google leading you to the only person in the world with the exact same problem, only to have the the thread abandoned or even better, a "Fixed it now thx"

 

[Manatox]

I had amout the same problem two times over the years i have been using a PC.

First time this happened, i still had an AthlonXP CPU back then, the problem was the motherboard that got defective for some reason, the problems disapeared after that got changed.

The second time happened with my current system. For some reason, my system started restarting itself and to freeze even doing nothing during staying idle on desktop, similar problems that happen when the voltage is not high enough on CPU (and sometimes RAM). I had my CPU overclocked that time too, but for some reason, the voltage that was rock stable for 6+ Months was not enough anymore to run it stable for whatever reason (yes, it was still the higher then stock voltage setting i had after overclocking the CPU to its limit).
I solved this with increasing both the CPU and RAM voltage through BIOS, though i still don't know which of those two needed more voltage, as i didn't bother to check afterwards when it started to run stable again (and still is even now).

I bit of a long post i know, but you might want to try bumping the voltage on both the CPU/RAM up and try it like that, because i know in fact that the normally stable voltage may suddenly not be enough for whatever reason there is.

 

[Resilient]

Fuck me.

New PSU in, the crashing still happens.

I'm at an absolute loss here. If anybody has any other ideas...please let me know

 

[knitoe]

Fuck me.

New PSU in, the crashing still happens.

I'm at an absolute loss here. If anybody has any other ideas...please let me know

Seems to be crashing when gaming, is your video card OC, by user or from the factory? If yes to either, downclock it and run at the normal speeds for that GPU. You can use software, like MSI Afterburner, to do that. Also, they show you to ingame / stress testing current temps.

 

[Resilient]

Seems to be crashing when gaming, is your video card OC, by user or from the factory? If yes to either, downclock it and run at the normal speeds for that GPU.

It's an EVGA GeForce GTX 1070 SC GAMING ACX 3.0, with no user OC.

Should I just use MSI Afterburner to downclock?

 

[knitoe]

It's an EVGA GeForce GTX 1070 SC GAMING ACX 3.0, with no user OC.

Should I just use MSI Afterburner to downclock?

SC = factory overclock.

Yes, use Afterburner and downclock to normal 1070 speed, Graphics Clock (MHz) =1506 and Memory Clock (MHz) = 2000.

 

[Evolution of Metal]

Fuck me.

New PSU in, the crashing still happens.

I'm at an absolute loss here. If anybody has any other ideas...please let me know

I am very surprised that you still test your system with all components in.

Start from the bottom: Mobo, one RAM stick, CPU and main drive. If your system isn't stable, then one of those components is faulty.

If it works well without crashing, start adding RAM sticks and eventually GPU. If it all works, then your secondary drives might be at fault.

Eventually, you will be able to diagnose the problem. It is a long process but nobody in here can tell you what is wrong in this case.

 

[Resilient]

SC = factory overclock.

Yes, use Afterburner and downclock to normal 1070 speed, Graphics Clock (MHz) =1506 and Memory Clock (MHz) = 2000.

This is what I have atm. Do I need to go lower?

I am very surprised that you still test your system with all components in.

Start from the bottom: Mobo, one RAM stick, CPU and main drive. If your system isn't stable, then one of those components is faulty.

If it works well without crashing, start adding RAM sticks and eventually GPU. If it all works, then your secondary drives might be at fault.

Eventually, you will be able to diagnose the problem. It is a long process but nobody in here can tell you what is wrong in this case.

I get what you are saying, but I have essentially done that.

I'm not trying to be rude, but I'm not understanding what else I should do. Apologies if it comes off that way - but this is what I've done so far.

I tested one stick of RAM
I tested the other stick of RAM in the other DIMM slot. No issues. Tested using memtest86.

I can use the PC fine and do anything on it with all components in. The crash so far has only ever happened when gaming.

8 hour CPU stress test, no errors. The Unigine test requires the GPU to test, and the results of it I posted above (Unigine crashed after 4 hours but not the system).

I swapped the PSUs and the crash still only happens when in a game.

I swapped the SATA ports for my SSD and HDD, it still happens.

I can't play a game without the GPU so I can't isolate the other components and test them in a game to see if they are causing the crash.

I've lowered the clocks on the GPU. If it happens again, I'm going to disconnect the secondary drive and uninstall all programs from it. Note - the game I'm playing most atm (FF14) is on the SSD - the SSD I replaced a week ago (replacing SSD didn't fix the problem).

 

[knitoe]

This is what I have atm. Do I need to go lower?

Sure. Try, lower. Use -200 Core and -200 Memory. If it stops your PC from crashing, it's a bad video card and you will need to RMA it.

 

[JaseC]

If GPU downclocking doesn't get you anywhere:

Have you tried increasing the CPU voltage even when running at stock? Your experience mirrors mine when I tried in vain across several weeks to maintain 4.3GHz on my 6800K. As the instability was completely and utterly random (sometimes occurring after minutes, other times several days) and never reared its ugly head during stress tests, only in games, there came a point where I believed I'd successfully identified the PSU as the culprit. Once it became abundantly clear that a stable 4.3GHz wasn't going to happen without breaking the 1.4v barrier, I dropped down to 4.2GHz, slightly reduced the voltage accordingly, and haven't had any issues since.

Start by bumping it up to something that'd you'd typically associate with a generous overclock, such as 1.36v.

 

[Evolution of Metal]

I can't play a game without the GPU so I can't isolate the other components and test them in a game to see if they are causing the crash.

I've lowered the clocks on the GPU. If it happens again, I'm going to disconnect the secondary drive and uninstall all programs from it. Note - the game I'm playing most atm (FF14) is on the SSD - the SSD I replaced a week ago (replacing SSD didn't fix the problem).

Sure you can. You have integrated GPU in your CPU. Just fire some tests, use low settings to play games. It is by far the easiest way to see if it is a dedicated GPU problem.

FF14 will work on low.

 

[Resilient]

If GPU downclocking doesn't get you anywhere:



Start by bumping it up to something that'd you'd typically associate with a generous overclock, such as 1.36v.

Ah, sorry I missed your post.

I originally had an OC on the CPU but have since taken it off since the problems appeared.
The crashes happened both when I had the OC and off. But, I've never once crashed in an OC stress test, or a stock clock speed stress test. The problem persisted regardless

Sure you can. You have integrated GPU in your CPU. Just fire some tests, use low settings to play games. It is by far the easiest way to see if it is a dedicated GPU problem.

FF14 will work on low.

You're right. My bad mate. If the crash occurs again with the reduced clock speeds as knitoe suggested, I'll take it out and play on Low. Sound like a plan?

 

[JaseC]

Ah, sorry I missed your post.

I originally had an OC on the CPU but have since taken it off since the problems appeared.
The crashes happened both when I had the OC and off. But, I've never once crashed in an OC stress test, or a stock clock speed stress test. The problem persisted regardless

Is that to say you left the voltage as-is (at the overclock level) when you went back to stock clockspeeds?

 

[Resilient]

Is it possible it's not the actual GPU itself, but the motherboard in some way?

I ask because my old card (7850 OC Radeon card (2gb, Gigabyte))did the exact same thing in this PC.

And that old card never had an issue in my old PC (i5 3550) ..

Is that to say you left the voltage as-is when you removed the overclock?

Correct, on Auto/default in BIOS settings

 

[Baleoce]

I had similar issues to you. Posted on gaf multiple times. In the end I just took the plunge and decided to buy a new PSU as that's what I suspected was the issue. Not had a problem since then. Not to say that this is guaranteed to work for you. But try to borrow a psu and test it out if you can't afford a new one outright. Or check to see if your warranty is still active on your current psu.

Edit - sorry, I Should've caught up first.

 

[Resilient]

I had similar issues to you. Posted on gaf multiple times. In the end I just took the plunge and decided to buy a new PSU as that's what I suspected was the issue. Not had a problem since then. Not to say that this is guaranteed to work for you. But try to borrow a psu and test it out if you can't afford a new one outright. Or check to see if your warranty is still active on your current psu.

This is my 2nd PSU and the error just happened again

 

[JaseC]

Correct, on Auto/default in BIOS settings

I actually meant at the overclock level but realised that wasn't overly clear and edited, haha. As I mentioned earlier, try bumping up the voltage to, say, 1.36v but not overclocking. There's a slim chance that the default voltage is too low.

 

[Baron Aloha]

No, I didn't (sorry). How does this affect my driver updates in the future, though? That's what I was a bit confused about.

Someone can correct me if I'm wrong but you could always re-install it, update drivers, and then uninstall again. I don't think the drivers are tied to the application. Of course, the nice thing about having the application is that it knows when updates have been put out - but unless a specific game(s) you are using have problems you shouldn't need to update drivers as often as they seem to put them out.

It sounds like you have tried some other things in this thread. You may as well try uninstalling, rebooting, and seeing if you run into the issue. At this point you have nothing to lose. Like I said earlier, I went through the same thing as you - replaced literally everything at one point - and this is what worked for me.

 

[knitoe]

Is it possible it's not the actual GPU itself, but the motherboard in some way?

I ask because my old card (7850 OC Radeon card (2gb, Gigabyte))did the exact same thing in this PC.

And that old card never had an issue in my old PC (i5 3550) ..

Yes, it's possible that a bad MB can cause many weird issues. Have tried your video card in the other lower PCIe slot?

 

[owasog]

Is it possible it's not the actual GPU itself, but the motherboard in some way?

I ask because my old card (7850 OC Radeon card (2gb, Gigabyte))did the exact same thing in this PC.

Do you have a second PCIe slot on your motherboard? Stick the GPU in there to troubleshoot.

Also, do you know how to clone your OS drive? If you have a spare drive somewhere you could eliminate a faulty drive that way.

 

[MarkMclovin]

Issues like this I've diagnosed by forcing a crash dump file at the point of the freeze. Assuming the OS is running to some capacity.

https://msdn.microsoft.com/en-us/library/windows/hardware/ff545499(v=vs.85).aspx

You can then analyse it and see whats triggering it.

 

[Resilient]

I actually meant at the overclock level but realised that wasn't overly clear and edited, haha. As I mentioned earlier, try bumping up the voltage to, say, 1.36v but not overclocking. There's a slim chance that the default voltage is too low.

But if the system passed a stress test at the default settings, wouldn't that suggest the CPU is not the issue (no crash when testing)?

I'll give this a shot if it crashes after lowering the GPU clocks. Thank you.

Someone can correct me if I'm wrong but you could always re-install it, update drivers, and then uninstall again. I don't think the drivers are tied to the application. Of course, the nice thing about having the application is that it knows when updates have been put out - but unless a specific game(s) you are using have problems you shouldn't need to update drivers as often as they seem to put them out.

It sounds like you have tried some other things in this thread. You may as well try uninstalling, rebooting, and seeing if you run into the issue. At this point you have nothing to lose. Like I said earlier, I went through the same thing as you - replaced literally everything at one point - and this is what worked for me.

I'll try this next (after sussing out the GPU underclock).

Yes, it's possible that a bad MB can cause many weird issues. Have tried your video card in the other lower PCIe slot?

No - I thought about this today though. I'll give that a shot next.

Do you have a second PCIe slot on your motherboard? Stick the GPU in there to troubleshoot.

Also, do you know how to clone your OS drive? If you have a spare drive somewhere you could eliminate a faulty drive that way.

I'll try the second PCIe next.

I've replaced the SSD (drive that the OS is installed on) as of 2 weeks ago; crashes still persisted.

Issues like this I've diagnosed by forcing a crash dump file at the point of the freeze. Assuming the OS is running to some capacity.

https://msdn.microsoft.com/en-us/library/windows/hardware/ff545499(v=vs.85).aspx

You can then analyse it and see whats triggering it.

So I'm forcing a system crash and then reviewing the dumped data? interesting...what would it generally return (out of curiosity)?

 

[Resilient]

So for those who think it's a GPU issue..what could be the problem?
Overheating? The temps have always stuck at about 72c, I've never seen them exceed that.

 

[JaseC]

But if the system passed a stress test at the default settings, wouldn't that suggest the CPU is not the issue (no crash when testing)?

Not necessarily. The 4.3GHz overclock on my 6800K I tried in vain to stabilise never became unstable during stress tests, just during games (primarily The Witcher 3 and Just Cause 3). That and the sporadic nature of the instability lead me to falsely conclude that the PSU was the culprit. I haven't had any issues whatsoever since dropping down to 4.2GHz and slightly lowering the voltage accordingly.

 

[PaNaMa]

Fuck me.

New PSU in, the crashing still happens.

I'm at an absolute loss here. If anybody has any other ideas...please let me know

Plug into a wall socket on a different circuit in your house.
Use a UPS if possible to prevent browning out (heh).
Really tho. I had an issue similar to yours a couple of years back, and I thought it was PSU but it was just plain old power being wonky from my outlet.

 

[Resilient]

OK - i'll give that a shot next.

Just crashed again with the lower clocks.

what should I try next? remove GPU all together or put it in the 2nd pci-e slot?

 

[knitoe]

OK - i'll give that a shot next.

Just crashed again with the lower clocks.

what should I try next? remove GPU all together or put it in the 2nd pci-e slot?

First, since you can use onboard video, try without video card. Then, try video card in a lower PCIe slot.

 

[MarkMclovin]

So I'm forcing a system crash and then reviewing the dumped data? interesting...what would it generally return (out of curiosity)?

No not forcing a system crash, but forcing the creation of a dump file at the time of the freeze. A lot of freezes doesn't actually give you a crash and a dump file, which is what you'd prefer when figuring out whats going wrong. But seeing as you're getting a freeze over a crash then you can force it.

EDIT - Have you tried uninstalling any recent updates? Assuming it started recently.

 

[Easy_D]

Man I hope you can get this solved OP. On the bright side, at the end of it, you'll be a fucking master of troubleshooting PC hardware. Unfortunately I don't really have anything to contribute other than a good luck

 

[Engell]

OK - i'll give that a shot next.

Just crashed again with the lower clocks.

what should I try next? remove GPU all together or put it in the 2nd pci-e slot?

yes try another pciexpress slot
also make sure you have the newest EVGA bios for your 1070 card, there was problems with the memory modules in the start.(even if they state better overclocking, some users had problems at stock speeds)
http://forums.evga.com/EVGA-GeForce-GTX-1070-BIOS-Update-v8604500070-m2565056.aspx

 

[Inkwell]

Maybe I read that wrong, but it still sounds like you have only tested your ram with memtest. I highly suggest testing each stick on their own while doing whatever it is that crashes/freezes your system the most. Outside of this, it sounds like the most likely culprit could be the motherboard. I'll tell you what happened with me.

I had constant freezes and crashes when I built my current PC in 2010 (it's the same outside of upgrading the ram and getting a new video card). It was happening multiple times a day and very frustrating. It went on for months.

To be quite honest I'm not sure what the actual problem was, but I did a few things at once. First, I had sent my ram into the manufacturer and had it replaced. That didn't seem to fix the problem, but supposedly faulty ram could have corrupted some files in Windows previously. I simultaneously did 2 other things:

1. I went into the bios and turned off everything I wasn't using. This includes extra SATA ports. I know someone else mentioned something issues with specific SATA ports earlier in the thread.

2. I reinstalled windows.

Here's the thing. Since then I haven't had a single issue that I can remember. I'm not sure if it was faulty ram that corrupted windows, or whatever I turned off in the bios.

I'll be repeating myself a bit, but I would do all of these things in this order:

1. Turn off all unused hardware/features in the bios.
2. Test ram isolated in conditions that most likely cause freezing.
3. Test separate slots on the motherboard for both ram and video card, plus whatever else you may have in there.
5. Try a new motherboard.

Maybe not #5 unless you completely exhaust all other potential fixes. I really hope if any of this doesn't help, someone else figures it out before you start buying more hardware.

 

[MrFlooD]

It's 100% Mobo or RAM. The chances you have two faulty GPU's is slim to none. Especially since you know the 7850 worked in an old build.

As recommended above, test the RAM one stick at a time. If still freezing, RMA the motherboard.

 

[Brerlappin]

OP, are you on Windows 10? Would it be feasible to install Windows 7? I say this because i had no end of issues with memory leaks in Windows 10 that would lock my PC up for ages. I fixed it eventually by a registry edit.

Also, are you using onboard sound, or a sound card? I had alot of freezing/hard lock ups in certain games caused by a realtex onboard sound card. Try disabling sound, play a game for a while, and see if it still locks up.

 

[LordCiego]

yes try another pciexpress slot
also make sure you have the newest EVGA bios for your 1070 card, there was problems with the memory modules in the start.(even if they state better overclocking, some users had problems at stock speeds)
http://forums.evga.com/EVGA-GeForce-GTX-1070-BIOS-Update-v8604500070-m2565056.aspx

Check what he said, knowing the other things you tried this could be problem.

http://www.theverge.com/circuitbrea...vga-geforce-1080-1070-1060-graphics-card-fire

 

[Baron Aloha]

It's 100% Mobo or RAM.

Not necessarily. See my posts. GeForce Experience has been freezing up people's PCs. With a little googling you can find a bunch of people who had their freezing issues resolved after removing it (myself included).

 

[Leatherface]

Also, are you using onboard sound, or a sound card? I had alot of freezing/hard lock ups in certain games caused by a realtex onboard sound card. Try disabling sound, play a game for a while, and see if it still locks up.

I agree with this.


I think this may be a motherboard issue. I've had some wacky problems related to onboard sound myself where I had to modify the latency timing or something to that effect to fix it. I would try disabling onboard sound completely first.

 

[Resilient]

Ok, everyone - here is my next plan of attack.

'To do' before further testing:

1. Enable the Crash report. I tried this using the method Rudimental outlined, but none of the crashes have created a Crash report/Minidump yet. I'll set this up using the method MarkMclovin posted.
2. Update BIOS of the EVGA 1070 SC card.
3. Turn off all unused hardware in the Motherboard BIOS (if someone could point me in the right direction as to what I need to turn off, it'd be appreciated).

Then I will test under the following conditions (in order):

1. Put GPU in the 2nd PCIe slot, game, test for a crash.
2. Remove the GPU and game using the Integrated Graphics, test for a crash.
3. Remove 1 RAM stick, game using 1 stick in DIMM A1, test for a crash.
4. Remove 1 RAM stick, game using 1 stick in DIMM B1, test for a crash.
5. Remove the 2nd storage drive (HDD), game, test for a crash.

JaseC - I'm still pretty unclear on what I'm doing with the clock speeds of the CPU using your method. It's basically me setting the clock speeds myself rather than letting the mobo run it at the default settings? So I should set the CPU frequency to maybe 37, how would I work out the voltage to use then? I'm tempted to make this the first step, but I feel like it requires heaps of attempts before I can rule it out as the issue (i.e fiddling with voltages each time until I decide, fuck it, it's not the CPU voltage).

If NONE of this resolves the issue...what do people suggest next?

Thank you for all your input so far peeps.

 

[Timu]

What is going on with your PC?

 

[Resilient]

What is going on with your PC?

No matter what game it is, the game will just crash. It might be 4 hours into a session or it could be 1. But the game just crashes, screen freezes on the game, and sometimes the audio goes weird (screeching, distortion, audio stuttering), the PC loses functionality (can't open Task Manager, can't alt+tab), and I need to hold the power down for 5seconds and turn it back on again. all fans are still on, all lights are still on etc.

 

[jahepi]

Hi Resilient,

I´ve been following the thread and you´ve already swap the PSU, tested each module of RAM, changed SATA ports, tested the CPU with Prime, you did experience same symptoms with your previous video card so i don't think the problem is your current card so the only component missing is the motherboard, it is going to a pain in the ass to get another motherboard but i don't see any other way around.

 

[Engell]

No matter what game it is, the game will just crash. It might be 4 hours into a session or it could be 1. But the game just crashes, screen freezes on the game, and sometimes the audio goes weird (screeching, distortion, audio stuttering), the PC loses functionality (can't open Task Manager, can't alt+tab), and I need to hold the power down for 5seconds and turn it back on again. all fans are still on, all lights are still on etc.

try uninstalling the audio driver, literately go to the device manager select your sound card, click on uninstall, if you get a little check box that says delete driver then check that as well... click ok reboot and let windows install the standard vanilla windows HD audio driver(hopefully it works). if possible try doing this while unplugged from the internet since shitty windows update potentially could start downloading and installing the same driver again...
Do we even have games that can run offline anymore .. see if you can test like this, if that works you can try to see if windows update will leave the standard driver alone, otherwise maybe you will have to kill windows update(if that is even possible anymore)
Did it on the release version of win10 home on my laptop, also because of audio driver issues(wouldn't detect when i plugged in headphones)