Black screen with sound followed by manual restart with no dump

Hopefully things work out which one you decide to try with.

And yeah even with stress testing you get conditions like full load and a constant demand under a stable test condition which can still be a good stress test but the rippling and fluctuations as the GPU powers up and down and adjusts is what tends to be a issue as it can for a few milliseconds hit above what the rated and measured specs show.

That was AMD's issue with the Vega though it persists with Navi too but it draws less so the spikes are overall lower.
Still could hit close to double over the rated maximum power draw and amperage values for but a moment and that was pretty hard on current power supplies in fact as I remember some manufacturers even made new models that could handle the extra load.

The opposite used to be true too but from 2014 and newer most PSU's have the ability to draw a lower power amount without shutting down, Intel CPU sleep state problems and also AMD now though there's a bios toggle for it on their systems for low load or typical load which draws a bit more but keeps things stable on older hardware.
(5+ years though and I'd be looking into a replacement unit they can last but it's not a component you want to chance with.)


EDIT: There's a article on that too actually.
https://www.igorslab.de/en/graphics...s-cause-and-correct-power-supply-measurement/


Bit of a peak there alright.

It can be a bit of a mess honestly keeping track of some of these things, I was kinda unaware of most of this too until more recently as I started looking into the Navi GPU stability issues.

Just being aware the GPU's can pull a bit above the rated levels usually works without having to specialize in these things, can be helpful to look into when there are problems but it's a bit more specialized.

 

I just realized, does max power draw affect anything unless I'm at close to 300W? Thing is, playing at 144fs being limited by the game engine my "max" power draw is at around 200w

 

You can't have RMA rates for products you don't sell.

 

That's a good question, I'm on AMD and Wattman has a direct impact on the power draw when adjusting the slider.
Makes it a good idea to see how much is actually needed so you don't needlessly increase the slider and cause the GPU to draw more than it would need.

Was thinking it'd be similar for NVIDIA it draws what it does and then this increases or decreases it by the percentage value set but perhaps it only affects the max allowed and then this can be increased or decreased further if desired.


EDIT: Hmm wonder how the GPU clock speeds are reacting.
If it's trying to boost because there's room for it as it's not hitting any thresholds perhaps that's what is hitting the OCP when that kicks in.

Can't say for certain though I am not very good at how NVIDIA's newer boost algorithms behave.
Other than that it works really well from what I can tell from reviews and such as it's hitting up to 2 Ghz or above even with voltage reduction via GFE or Afterburner.

And if monitoring is only updated every few milliseconds those bursts might not be registering as it's a very quick shift.
Both AMD and NVIDIA have a pretty dynamic clock speed now and how the GPU tries to hit this target value if the various sensors and such indicate that there's no current limits for it.

 

Unlikely.
TBH I dont even know what you're trying to do. What kind of OCP should get triggered at around 200W of semi-light gaming.

Right.
And since it's a well established fact that French have banned HX, HXi and AX, no need for any sources. As usual...

 

@EvilDoll89

Q: Did you get any BSOD with the 2nd monitor switched OFF?

 

haven't really tried that, and it's going to be a hard one due to the nature of my work. I'll see if I can pull this off somehow. I'm currently running driver verifier and ordered a new rm1000i PSU from corsair as I got a good deal from it.
At least I'll know for sure if it's the PSU or the GPU that way.

 

Just had another black screen .
Driver verifier wasn't running.

I increased the tdr delay to 10 seconds( a few days ago) from 2 seconds.
And for the entire ten seconds I could:
- talk to someone over voip(microsoft teams)
- restart the display driver using menu+ctrl+crap+b > I could hear the beep, but after the 10 seconds the PC still completely froze.

edit: oh an ofc ,no dump

at this point I'm waiting for my new PSU to swap it in.

 

I plugged in my old GPU after using DDU in Safe mode.
Once my new PSU comes I'll swap to the new PSU + new GPU.

Updates will follow as the "situation" unfolds

 

around 3 days with my old gpu plugged in, with no black screens. My new PSU should arrive in a few days and I'll swap in the 3080 in again.

 

new PSU came Tuesday, installed it and the 3080.
It's Friday, The PC ran more or les 24/7. No black screen, no errors, it's even boosting to higher clock speeds on default settings.
Also installed a 900W UPS, albeit that probably doesn't matter.
Looks like it was indeed the PSU, as I was getting 1 or 2 black screens a day, if not more.

 

Awesome, game on!

 

Did the new PSU completely solve this? I am having the exact same problem in my brand new PC. Black screen, still can hear the audio then PC restarts itself. Was that the same you had? I also get the volmgr in eventviewer. No dump.

 

i had the same issue, PSU fixed it for about 2 months, now its back again, mind you i RMAd every single part of the PC and the same thing started happening a few months down the line. reseating the RAM also fixed it momentarily.

 

it was definately the PSU, sorry for the late reply. Been almost 4 months now with no more issues.

 

I have the same card except it's Rev. 2, and was 100% having the same issue. It's a messy driver issue. NVIDIA pushed out a bad driver which caused the issue then a few days later pushed a fix. The fix apparently didn't overwrite everything necessary, or there was some sort of Windows related issue. I had to fully uninstall the driver using device manager. After reinstalling the latest driver everything was back working as usual. I'm sure you have your problem resolved by now, but wanted to post the resolution for posterity. I spent hours trying to fix it, and hopefully can save other the trouble.