Question R9 280x grey screen crash

[luks104]

Hello everyone,
I encountered a weird problem with my system the other day. I was playing PUBG with my friends, when suddenly the pc freezes and a grey screen with vertical lines appears. System was unresponsive and I had to reset it. Played 2 more rounds and it was okay, no crashes. I then tested the system with different benchmarks (Furmark, Firestrike, Unigine Valley..) and it ran all of them completely fine. Since I didn't play so many games in the last year I thought it was just a PUBG related issue. I then proceeded to play GTA V to see, if the problem appeared there and sure enough, after 1,5 hours of game play with texture and shadow popping (which didn't occur about half a year ago) I got the dreadful grey screen with vertical lines again.

Link to identical picture of the grey screen: https://www.overclock.net/photopost/data/1193077/9/9c/9c3ef323_HD5000VS.jpeg

After searching for answers online I was sure that the problem was indeed in the VRAM modules of the graphics card (known problem for the R9 200 series GPUs) and asked a friend, if I could borrow his R9 380x, which he replaced with a 2070 and didn't need anymore. I uninstalled the drivers, replaced the 280x with the 380x and installed the correct drivers. Now comes the weird part. I tested the system with the Unigine Valley benchmark and 3 seconds into the benchmark the system froze with a black screen and a hard reset was required. After the reset, there was no video output and I heard the windows welcome sound. I pressed the power button to power it off, and restarted the system. I got into windows normally, like nothing happened. Re-ran the benchmark and exactly the same happened again. Uninstalled the drivers, powered off the system, removed the graphics card and ran the benchmark with the integrated graphics - no problem appeared. I am now puzzled as to what is causing this problem. I am thinking the PSU is on it's last legs, but I am unsure and don't want to spend money on a new unit to replace a functioning one.

Full specs of my system should be in my signature, but I am going to post them anyway:
Intel Core i5 4670k |
Sapphire R9 280x Vapor-x |
Gigabyte GA-H97-HD3 |
Kingston HyperX 16GB |
Gigabyte GreenMax 650w |
Antec P100 case |
2x 1TB WD Green |
128GB Samsung 840 EVO

So I am once again asking for your help since I do not know what to do next. I am not in a hurry, especially since I don't plan on playing any games anyway, but want to solve this problem. Thanks in advance.
Cheers

 

[Metal Messiah.]

Gigabyte GreenMax 650w |

That PSU might be the culprit, IMO. That's a mid-end unit. But I'm not 100% sure. IF possible, replace this PSU asap, and grab yourself a HIGH quality, upper TIER PSU. Can you test your GPU by using some other PSU model ?

EDIT: How OLD is this PSU ?

 

[luks104]

Thanks for the answer. I bought the unit in the begining of 2014, so it's 5 years old.
Honestly, it never caused me any sort of problems through the years I had it, and it even wasn't cheap. After all it's 80+ bronze certified. It's not the best unit, no, but it isn't the worst. But it could be on it's last legs by now.
But if it was the psu, I don't know why the much less power demanding 380x (2x6 pin connectors) crashes just by running at full load after 2 seconds. The 280x (2x8 pin connectors) crashed just two times with several hours of max load. My logic tells me that the crashes would be more frequent with the more power hungry card, if it was the psu. But I don't know if the 380x is more sensitive to voltage spikes.
That's why I am puzzled.
Edit: I currently don't have access to a powerful enough psu