Suddenly can't put any load on week old 2080 Ti without crashing

Questloved · 4 Oct 2018 at 04:29

Specs:

Mobo: ASRock Z370 Extreme4
CPU: Intel Core i7 8700k
GPU: Palit RTX 2080 Ti GamingPro OC
RAM: 2x8GB Corsair Vengeance LPX DDR4-3000
PSU: Corsair TX750

Was playing Monster Hunter World for hours earlier, quit the game and left the room for about 2 minutes until I heard a fan spinning really fast on and off in my PC. Came back to a black screen, but managed to reboot. Since then if I put any load on the GPU, even something like 40% sitting in menus on Monster Hunter, I'll get the same problem. Temps are fine as I'm not even able to run the card long enough to heat it past 50ish degrees, CPU and memory tests run fine. Absolutely everything seems normal apart from this crashing. When I replicate the error in Far Cry 5 I get the normal 160-200 FPS until the system dies after the 10 seconds or so. Reinstalling Nvidia drivers had no effect. I'm going to try and swap it out for a different GPU tomorrow to see if it happens with that too, in which case I'm assuming it would be the power supply, right?

Does anyone here have any ideas of what else I could do to try and narrow down a cause? I'm feeling really stressed by the whole PC building experience right now.

Apologies if this is in the wrong place, I'm slightly panicked right now

Vega · 4 Oct 2018 at 05:20

Ya I'd say power supply most likely. Is your PSU modular so you can switch power outputs on the PSU? Check that all power cables are seated well? It could be the GPU power section, but usually those are pretty robust and don't fail.

GordyR · 4 Oct 2018 at 06:13

Is your PSU multi rail by any chance? If so, switch it over to single rail mode.

I’ve heard of others having the exact same issue with some of the high end Corsair power supplies that offer switchable rail modes.

Switching to single rail mode (either by a switch on the PSU itself, or on software like Corsair Link) solved it for them.

Bailey · 4 Oct 2018 at 09:50

GordyR said:
Is your PSU multi rail by any chance? If so, switch it over to single rail mode.

I’ve heard of others having the exact same issue with some of the high end Corsair power supplies that offer switchable rail modes.

Switching to single rail mode (either by a switch on the PSU itself, or on software like Corsair Link) solved it for them.

Good Call

@Gibbo just showed me this thread and said the exact same thing!

Bailey

Nasher · 4 Oct 2018 at 10:18

Shouldn't suddenly happen after a week though. I'm betting it's the card.

spikeot · 4 Oct 2018 at 10:43

The advice above is sound though, and there’s only one way to find out. My money would still be on the PSU.

Marine-RX179 · 4 Oct 2018 at 11:03

In general I would agree with the what's said above, but does this only happen on Monster Hunter World? What about other games and benchmarks?

The reason I am asking this is because specifically for Monster Hunter World, I recall reading there's been quite a lot of stability issue with Nvidia GPU. It could well be Nvidia's done a pants jobs on their driver in the support for this game , and if that's the case changing different drivers won't help, unless Nvidia got a driver that's specifically for fixing this issue. To be frank some people also reported new drivers has caused some stability issue in general as well, but not sure for 2080Ti specifically.

Distracted · 4 Oct 2018 at 11:55

Bailey said:
Good Call

@Gibbo just showed me this thread and said the exact same thing!

Bailey

Is this a defect with the PSU, or design fault or is it something else?

Distracted · 4 Oct 2018 at 12:02

Okay I've looked it up and it seems to be an issue with more than 40 amps on a single rail tripping the psu protection when in multi rail mode.

The fix is either switching to single rail OR having each 8 pin on separate rails to ensure you don't reach 40 amps on a single rail. Instead of using 2 plugs on one cable from one 8 pin port on the psu, use 2 cables from different 8 pin ports on the psu.

This is specific to the corsair multi rail 'issue' which is actually the psu working as intended. This may or may not be the case for you.

Questloved · 4 Oct 2018 at 12:31

GordyR said:
Is your PSU multi rail by any chance? If so, switch it over to single rail mode.

I’ve heard of others having the exact same issue with some of the high end Corsair power supplies that offer switchable rail modes.

Switching to single rail mode (either by a switch on the PSU itself, or on software like Corsair Link) solved it for them.

I believe my power supply is single rail, it's a Corsair TX750. That said I've never even heard of the difference until now and I'm not sure how to check.

Nasher said:
Shouldn't suddenly happen after a week though. I'm betting it's the card.

The suddenness of it is what's spooking me, it went from fine to this in literally minutes for no obvious reason.

Marine-RX179 said:
In general I would agree with the what's said above, but does this only happen on Monster Hunter World? What about other games and benchmarks?

The reason I am asking this is because specifically for Monster Hunter World, I recall reading there's been quite a lot of stability issue with Nvidia GPU. It could well be Nvidia's done a pants jobs on their driver in the support for this game , and if that's the case changing different drivers won't help, unless Nvidia got a driver that's specifically for fixing this issue. To be frank some people also reported new drivers has caused some stability issue in general as well, but not sure for 2080Ti specifically.

No, it's any game or stress test I try. It even happened once on the Windows login screen.

Thanks for the advice guys, I'm gonna reseat the card and make sure everything's properly connected and hope for the best. I've got my old GPU I can slap inside to see if I can replicate the problem on there as well if that doesn't work.

Quartz · 4 Oct 2018 at 13:33

Are you sure the fans on the GPU are working? I recall my Titan XM suffering infant mortality in that regard.

GordyR · 4 Oct 2018 at 13:51

Kieran said:
But I would still try each 8 pin on a different cable from the PSU because I believe that over voltage protection can be used for each output on a single rail PSU.

I agree entirely. If the PSU is single rail only, which it does indeed appear to be, then ensure that each 8 pin connector has it's own cable.

Then force the fans to 100% on in Afterburner/Precision, check they're working fine, and fire it up again for testing.

The PSU/power is still the most likely culprit here. Do you have access to another that you could try?

Although not impossible, my view is that until you've ruled everything else out, it's unlikely to be a defective card. Given that this issue appears reasonably common and many people have been able to fix it by replacing/re-configuring their PSU's, I would want to be 100% certain before considering the RMA route.

On a probably completely unrelated note, my 2080 Ti hard locks the windows desktop without fail if it's not using the actual NVidia driver, and is instead using Microsoft's "safe" ones.

I had to connect my monitor to my 8086k's iGPU, boot up, install the NVidia drivers, then reboot a couple of times until windows initialised the correct drivers upon start.

azazl187 · 4 Oct 2018 at 14:24

Please ensure you do not have excessive GPU sag. Have you taken the card out and re seated it at all? What were the results from saying that you were going to do this in a previous post?

dvs8 · 4 Oct 2018 at 14:24

Aforementioned other thread OP here! *waves*

After a very lengthy thread on Reddit (over 30k views) we eventually figured out:
* switching to single-rail mode on my PSU fixed my 2080Ti issues
* multi-rail mode isn't "BAD", the combination of multi-rail mode, 2080Ti's excessive power needs, AND running a single cable from PSU to GPU is the real culprit
* for the best setup, use either single-rail or multi-rail mode on your PSU, but most of all make sure you're running 2x cables from the PSU to the GPU

VincentHanna · 4 Oct 2018 at 14:27

I don't even know what a PSU rail is!

I have a 1200W Super Flower Leadex Platinum. Never had an issue.

HazardO · 4 Oct 2018 at 15:41

VincentHanna said:
I don't even know what a PSU rail is!

I have a 1200W Super Flower Leadex Platinum. Never had an issue.

"Multi Rail" PSUs are essentially two separate power circuits in one unit. Single rail PSUs have just one. IIRC. So a 1000Watt dual rail PSU behaves more like two 500W PSUs working in tandem. If you power a highly overclocked CPU and a powerful GPU from just one of the two rails, it will shut down as it tries to draw more than 500W even though it's a "1000 W" PSU. Even on a single rail PSU you probably don't want to be using a pair of 8 pin PCIe connectors which are connected to the same port on the PSU as you can draw too much power from that port.

Vega · 4 Oct 2018 at 15:47

dvs8 said:
AND running a single cable from PSU to GPU is the real culprit
* for the best setup, use either single-rail or multi-rail mode on your PSU, but most of all make sure you're running 2x cables from the PSU to the GPU

Not to throw stones, I would think that would have been pretty obvious...

Questloved · 4 Oct 2018 at 16:11

So the troubleshooting I did today actually seems to have made things worse for the 2080 Ti:

1. Reseated the GPU without changing the PCI-E cable setup - no change
2. Tried a separate PCI-E cable - couldn't actually boot into Windows anymore, fans immediately go to 100% and I get a black screen
3. Reseated the GPU with the new PCI-E cable setup - no change, still can't get into Windows
4. Went back to the old PCI-E cable setup to see if I could at least get into Windows, no dice
5. Put my old R9 390x back in - boots fine, has been running at max load in Far Cry 5 for about 20 minutes with no problems

Is it possible that it could still be the power supply or have I basically got a £1200 brick on my hands? I'm considering ordering the modular EVGA Supernova G3 750W for tomorrow or Saturday to make 100% sure, because this one I have right now is a pain in the arse with all the useless cables in the way anyway.

GordyR · 4 Oct 2018 at 16:21

Questloved said:
So the troubleshooting I did today actually seems to have made things worse for the 2080 Ti:

1. Reseated the GPU without changing the PCI-E cable setup - no change
2. Tried a separate PCI-E cable - couldn't actually boot into Windows anymore, fans immediately go to 100% and I get a black screen
3. Reseated the GPU with the new PCI-E cable setup - no change, still can't get into Windows
4. Went back to the old PCI-E cable setup to see if I could at least get into Windows, no dice
5. Put my old R9 390x back in - boots fine, has been running at max load in Far Cry 5 for about 20 minutes with no problems

Is it possible that it could still be the power supply or have I basically got a £1200 brick on my hands? I'm considering ordering the modular EVGA Supernova G3 750W for tomorrow or Saturday to make 100% sure, because this one I have right now is a pain in the arse with all the useless cables in the way anyway.

When you say that it can't boot into windows, could you be a little bit more specific?

Does it hang on the login screen? Does it turn black on the login screen? Does the boot process appear to complete? Does it lock while the windows loader is spinning?

My 2080 Ti runs perfectly, but if I were to swap it out with another video card temporarily, then switch back as you've done, then mine would lock at the windows login screen every single time until I force it to switch over to the NVidia drivers, rather than the default windows "safe" ones.

You might have fixed the original issue, but are now experiencing what I have.