Question GPU being finnicky - possibly messed up PCIE slot but not sure

[Jorvalt]

Okay so this all started with the fans on my current GPU crapping out. Not entirely dead, but failing and making a lot of noise. I eventually settled on replacing it with my previous card for the time being while I wait on XFX support, but now I'm facing the issue of it just crashing. I'm not really sure what it is, but the computer will still be on, though both the GPU and CPU fans stop moving which leads me to believe they just stopped, for some reason. It's possible that it's overheating since it only happens while playing games but I have good airflow in my chassis (or at least, as good as I can possibly get it). I haven't caught it at high temperatures because of course Radeon doesn't warn you when your GPU is about to overheat and doesn't even tell you that after it overheats. I also have used this card before and never had any problems with overheating, even using Radeon's terrible default fan settings. I'm worried that somehow the slot was deformed (the new card is a big card, though I do have a brace on it) since this only started happening AFTER I swapped out the GPUs. It took considerable force to actually remove it which seemed unusual. And no, I'm not a moron, I pushed down on the tab thing to unlock it. The slot itself doesn't look deformed, but I'm not really sure what else it might be.

System specs:
GPU (the newer one, in case it's relevant): Radeon RX 5700 XT
GPU (the old one, which I'm using now): Radeon RX 590
CPU: AMD Ryzen 5 3600 6-Core Processor
RAM: 16GB
Motherboard: B450 Tomahawk Max
PSU: Thermaltake M750W

 

[vov4ik_il]

I assume you used DDU, right? One would have to remove clocking/voltage/power profiles when swapping GPUs of the same brand... otherwise, they will be applied to the other card and might not fit. Make sure to remove Afterburner and similars before using DDU.

 

[Jorvalt]

I assume you used DDU, right? One would have to remove clocking/voltage/power profiles when swapping GPUs of the same brand... otherwise, they will be applied to the other card and might not fit. Make sure to remove Afterburner and similars before using DDU.

I actually didn't, I'll try that and get back to you.

 

[Jorvalt]

Nope. I thought it was working last night, but this morning I experienced the same problem.
I suppose it's also worth noting that sometimes even when rebooting the computer doesn't exactly boot up fully. It gets past the BIOS and then the monitors disconnect and it just kind of hangs there.

 

[vov4ik_il]

I would try running stress tests with OCCT for one thing at a time to test individual components and then all together to stress PSU and see where it would fail.

 

[Jorvalt]

I would try running stress tests with OCCT for one thing at a time to test individual components and then all together to stress PSU and see where it would fail.

I tried every test (30 mins for each) and nothing. You can't run more than one test simultaneously but it does have a "power" test. It seems the only way I can get it to crash consistently is when I'm playing a game, and I have no clue why because even doing the monitoring test while playing a game doesn't help any. And OCCT doesn't seem to leave any log files after the crash.

 

[vov4ik_il]

it does have a "power" test

So you did run that test too, right?
It might be the PSU or the VRM.

 

[Jorvalt]

So you did run that test too, right?
It might be the PSU or the VRM.

Yeah. I ran all of the tests as I said, including the power one. No issues.
Since the only way I could get it to happen was in a game I thought that maybe it was another issue. I disabled all of the settings that Radeon likes to apply to stuff (don't like those anyway) and THOUGHT I was onto something because it went a pretty long time before crashing again. So back to square one I guess.

 

[vov4ik_il]

Yeah. I ran all of the tests as I said, including the power one. No issues.
Since the only way I could get it to happen was in a game I thought that maybe it was another issue. I disabled all of the settings that Radeon likes to apply to stuff (don't like those anyway) and THOUGHT I was onto something because it went a pretty long time before crashing again. So back to square one I guess.

For mobo VRM, disable all the CPU boosting features in BIOS, that will lower VRM load spikes. The card should be ok with stock settings.

 

[Jorvalt]

For mobo VRM, disable all the CPU boosting features in BIOS, that will lower VRM load spikes. The card should be ok with stock settings.

I never changed any of those settings, though. And as I said before, I never had this issue until the time of posting this thread.

 

[Jorvalt]

It crashed again and now it won't boot. I didn't change the BIOS settings yet.
The case fans are on (correction, were, they don't turn on now) and the built in lights on the motherboard are on, but both the CPU and GPU seem to be off, the power light isn't on and neither is that blinking hard drive indicator thing.

 

[Jorvalt]

It boots again! I think one of my friends suggested that I try booting with each of the RAM sticks one by one. I have two 8GB sticks, so first I tried removing stick 1 and leaving stick two, nothing. Then removing stick 2 and leaving stick 1, and it booted.
So I don't know if one of my RAM sticks just decided to die on me or what, but I'll let you know if the issue is still happening in this state.

 

[Jorvalt]

Unfortunately, computer still crashing. It took a while this time.
At least it'll still boot.

 

[Jorvalt]

Unfortunately, computer still crashing. It took a while this time.
At least it'll still boot.

And just as I say that, now it won't boot again, regardless of which stick is in what slot.
I want to stab someone.
Edit: Now it boots again. I still don't know what the hell is going on.