Question PC Crashes During Graphics Testing

[EdwardJenkinsS]

My sister's PC is having issues that for years it did not have. I have done numerous things to try to nail it down.

The build: https://pcpartpicker.com/list/DLZjvj

The issue is that when using Furmark at 1920x1080 with 2X MSAA, the screen will go grey and the connection to the display will end. I have to hold the power button and restart the PC to get it to get back to the desktop.

Things that I have done.

- Do a clean install of the graphics driver with DDU, in multiple different scenarios.
- I put the graphics card which is a GTX 980, in my PC and it works fine. So its not the hardware in the GPU.
- Ive done a complete reinstall of Windows 10, I backed up data and then wiped it, and reinstalled Windows entirely. Replacing any corrupted files.
- I swapped the PSU in her computer for the one in my PC, the same issue persists with my PSU. Which works fine in my PC with a GTX 1080, doing the same Furmark tests. So the PSU can't be it.
-Ive swapped the PCIE slot that the GPU is in, same crash persists.
-Ive changed NVIDIA Control Panel settings to prefer max performance.
-I have taken out the RAM and reseated it. I am also in the process of using Memtest86 on it right now, so far no errors.
-I have done a CPU specific stress test, using Intels Processor Diagnostic Tool. The CPU passed all tests just fine.

In Event Viewer what usually preceeds the crash is an event that says "The description for Event ID 0 from source nvlddmkm cannot be found".

At this point Im kind of stumped. Ive done drivers, gpu, psu, cpu, reinstall windows and such. Any ideas on what it could be?

 

[Darkbreeze]

So the PSU can't be it.

Don't EVER, EVER say THAT. Because you know what, about 50% of the time I hear somebody say that, no matter what the circumstances are, guess what it turns out to be? Exactly. Now, I'm NOT saying the PSU is the problem, but I AM saying, don't ever say that because anything that can be wrong with any system can be a direct result of some problem with the power supply no matter what the problem LOOKS like. I can probably show you at least ten or fifteen threads where somebody even swapped a variety of PSUs around in various systems, did much the same things as you've done and in the end it was still the PSU anyhow. A PSU can absolutely have an issue when connected to one motherboard that it doesn't have when connected to a different one, and just because it seems to work ok in another system still doesn't always mean it IS ok.

Besides which, there are other criteria here. Depending on the models involved, your GTX 980 probably takes two 6 pin PCIe connectors while the GTX 1080 probably takes a single 8 pin (6+2). If there were something getting weak or failing on the circuit of one of those 6 pins, assuming you are using two different cables and not two 6 pins on the same cable, it could easily be an issue, and it's one you wouldn't see with the single 8 pin connected but you would with the two 6 pins.

Also, your GTX 980 is older. So, while both cards originally likely had something like a 520w overall PSU recommendation, the 980 has had two additional years to degrade and as graphics cards degrade they require more power to function properly. That can overtax a circuit or it could trigger some kind of protection. But since they card works ok in another system, protection triggering from the card is not likely the issue.

What are the EXACT models of the PSU in both these systems? Just so we know.

In truth, the problem is very likely the motherboard. Have you tried your GTX 1080 in her system?

 

[EdwardJenkinsS]

Don't EVER, EVER say THAT. Because you know what, about 50% of the time I hear somebody say that, no matter what the circumstances are, guess what it turns out to be? Exactly. Now, I'm NOT saying the PSU is the problem, but I AM saying, don't ever say that because anything that can be wrong with any system can be a direct result of some problem with the power supply no matter what the problem LOOKS like. I can probably show you at least ten or fifteen threads where somebody even swapped a variety of PSUs around in various systems, did much the same things as you've done and in the end it was still the PSU anyhow. A PSU can absolutely have an issue when connected to one motherboard that it doesn't have when connected to a different one, and just because it seems to work ok in another system still doesn't always mean it IS ok.

Besides which, there are other criteria here. Depending on the models involved, your GTX 980 probably takes two 6 pin PCIe connectors while the GTX 1080 probably takes a single 8 pin (6+2). If there were something getting weak or failing on the circuit of one of those 6 pins, assuming you are using two different cables and not two 6 pins on the same cable, it could easily be an issue, and it's one you wouldn't see with the single 8 pin connected but you would with the two 6 pins.

Also, your GTX 980 is older. So, while both cards originally likely had something like a 520w overall PSU recommendation, the 980 has had two additional years to degrade and as graphics cards degrade they require more power to function properly. That can overtax a circuit or it could trigger some kind of protection. But since they card works ok in another system, protection triggering from the card is not likely the issue.

What are the EXACT models of the PSU in both these systems? Just so we know.

In truth, the problem is very likely the motherboard. Have you tried your GTX 1080 in her system?

So I tried my PSU in her PC, and still the same issues. I also didnt use her PSU modular cables, I brought mine.

I also just discovered something, the system does not crash with the same Furmark test at 1080p, with 4x MSAA with Debug Mode enabled. The same test thatd crash when everything is the exact same. I went to the NVIDIA Control Panel, enabled debug mode. And boom its fine now.

PSU Models are
Corsair RM750 - Model #: RPS0119

Corsair AX760I - Model #: 75-001302

TY for replying btw, nice to have some new ideas. I have not tried my GPU in hers, I tried her GPU in mine. Ive avoided doing that so far because I figured if it was a power supppy issue, it may damage my card. But I can definitely try it

 

[Darkbreeze]

So I tried my PSU in her PC, and still the same issues.

Totally not what I asked. I don't mean to be short, or a jerk, or anything, but totally NOT what I asked. And, when it comes to this kind of stuff, it MUST be exacting, or we can't learn anything.

What I ASKED was, did you try her CARD in YOUR machine? Because, if the card works fine in YOUR motherboard, and given all the other things you've eliminated, then it almost HAS to be her board. But we don't convict until we've done all the testing we are able to do , and we haven't yet because I don't think you did that.

 

[EdwardJenkinsS]

In truth, the problem is very likely the motherboard. Have you tried your GTX 1080 in her system?

Oh my apologies, I thought you asked if I tried my GTX 1080 in her machine. To answer your question, I tried her GTX 980 in MY machine. When doing the same tests it passed with flying colors.

 

[Darkbreeze]

So her graphics card works FINE, in YOUR system?

And her card works FINE, in YOUR system?

And, in BOTH cases, that is factual no matter which PSU is used in WHICH system? Or, you didn't try BOTH PSUs, in BOTH systems, with BOTH video cards? Because if you did not, for any reason, not do any ONE of those things, you ought to. And then let me know. But if you did, then tell me, we will move forward.

 

[EdwardJenkinsS]

So her graphics card works FINE, in YOUR system?

And her card works FINE, in YOUR system?

And, in BOTH cases, that is factual no matter which PSU is used in WHICH system? Or, you didn't try BOTH PSUs, in BOTH systems, with BOTH video cards? Because if you did not, for any reason, not do any ONE of those things, you ought to. And then let me know. But if you did, then tell me, we will move forward.

Alright, I will report back after I have done this. Thank you for the suggestion.