A rather unique (?) problem with my GPU

[forceinline]

Hey guys,

 

so I am having what I think is a unique problem, at least I couldn't find anyone with the same issue via Google and search.

 

I'm running Win 10 64-bit.

 

My problem is the following:

My GPU, an RX580 Nitro+, works fine... ish. The thing is, it sometimes doesn't boot as if it were broken. I get past the Windows loading screen, and then, one of three things happens:

a) My GPU boots fine and shows no signs of damage

b) My GPU does not boot, I get a black screen, and the PC eventually restarts itself

c) Basically b), except I get a screen full of artifacts which either stay or quickly disappear.

 

Whenever b) or c) happens, my LEDs on my mouse turn off, and my keyboard follows suit about after 3 seconds. If a) happens, my mouse just turns off for a split second.

 

What's weird is that it's random. Sometimes, my GPU will boot fine 20 times in a row for a week straight. Sometimes, I'll have to restart my PC and try over and over again until it finally works.

 

All of that started happening around summer 2019, bought the GPU late 2018, so it ran fine for about 9 months.

 

Here's what I already know: It's most likely not the BIOS/Motherboard as I've changed it during my ownership of the GPU and my problem persists. I also have an R9 270 lying around which I just tested and works perfectly fine.

 

From what I've read I suspect that it's either a "minor" hardware fault (very likely), or that my PSU might be the culprit (Thermaltake 600W, about 8 years old. Yes, I know it's due to replace).

I'd try reflowing it in the oven, however I'm a bit scared that might break it completely, and with the current state of the GPU market I would not like to have to buy a new GPU right now.

 

Any help is appreciated!

 

Edit: If it helps, I just cleaned the GPU.

Edited September 11, 2021 by forceinline

[HanZie82]

Re-seat the GPU and check power-cable connection.

And i would suggest trying a different PSU before doing the reflow thing.

[Guest]

16 minutes ago, forceinline said:

I'd try reflowing it in the oven, however I'm a bit scared that might break it completely, and with the current state of the GPU market I would not like to have to buy a new GPU right now.

Do not do this. Your liable to damage the card in the process. If the GPU needs reballed, it should be done by someone who has the proper equipment to do so.

 

20 minutes ago, forceinline said:

From what I've read I suspect that it's either a "minor" hardware fault (very likely), or that my PSU might be the culprit (Thermaltake 600W, about 8 years old. Yes, I know it's due to replace).

This is where I would start. Everything here lines up with my past experiences of weak/dying Power Supplies.

 

We actually just had to replace one in one of our servers where a GTX970 would cause the system to be stuck in a reboot loop, only successfully booting 1 out of maybe 10 tries only to have it artifact and eventually shutdown under load.

[char3327]

@forceinline Probably a PSU issue? Sounds like it's pretty old and needs a replacement...

[forceinline]

56 minutes ago, Nayr438 said:

Do not do this. Your liable to damage the card in the process. If the GPU needs reballed, it should be done by someone who has the proper equipment to do so.

 

This is where I would start. Everything here lines up with my past experiences of weak/dying Power Supplies.

 

We actually just had to replace one in one of our servers where a GTX970 would cause the system to be stuck in a reboot loop, only successfully booting 1 out of maybe 10 tries only to have it artifact and eventually shutdown under load.

Thank you.

 

39 minutes ago, char3327 said:

@forceinline Probably a PSU issue? Sounds like it's pretty old and needs a replacement...

Yes, it is old. Don't know why I haven't replaced it yet, guess because it worked fine.

 

What wattage do you guys suggest? Is 600W still fine or should I am for a bit more wattage? For the record, I have 4 drives plugged in, sometimes a CD drive and an R5 3600X.

[forceinline]

1 hour ago, HanZie82 said:

Re-seat the GPU and check power-cable connection.

And i would suggest trying a different PSU before doing the reflow thing.

I've re-seated it multiple times and I'd assume the power-cable connections are fine too as I've just cleaned the GPU. Unfortunately I don't have another PSU, so I'll have to get a new one, which makes sense considering how old my current one is anyway.

[Kilrah]

Would rather go for 850W or so in case the GPU is dead and you have to replace it with a recent power-hungry one since I'd imagine it's most likely the case, that GPU had a defect and should have been RMAd as soon as the problems started appearing while it was still under warranty.

[char3327]

@forceinline @Kilrah

3 hours ago, Kilrah said:

Would rather go for 850W or so in case the GPU is dead and you have to replace it with a recent power-hungry one since I'd imagine it's most likely the case, that GPU had a defect and should have been RMAd as soon as the problems started appearing while it was still under warranty.

Yah, I mean if you can verify that the GPU is not dead then you should be okay with a 650W or so... but if you do need to get a new one check that warranty to see if you can get it RMAd but if not you might have to get a new GPU and that might require a little more wattage depending on what you get.