cancel
Showing results for 
Search instead for 
Did you mean: 

PC Graphics

Gpu core stuttering and blackscreens ingame

So a few days ago it started stuttering after I died and respawned in Arma 3. So I thought it might be the overclock. I disabled the oc and it worked fine. Now it stutters in 3dmark and ingame. I've tried all the steps like ULPS disable and power save disable and it is not working. Also now when I try to enter arma 3 my screen just turns black and the game crashes, like a bad oc crash. My card doesn't overclock nearly as close as it did before. Before I could get a 1400mhz oc without problems. Now I can't even get a 1350. Have I broken my card?

Temps are fine. Yesterday arma 3 was blackscreening but I fixed it by disabling power state enhancement and I got around 52c max on normal core clocks. Today the game won't even start and 3dmark is stuttering whatever I do. I've tried installing newer driver, older drivers, removing every third party oc program.. Nothing is working please help I am thinking of just throwing the card into the bin.. One thing that did work was when I tried 3dmark when I had no drivers installed, It had no stutters and was running fine but you can't really run the pc without gpu drivers in a daily basis.

My setup:

i7 9700k at 5ghz

Samsung 850 250gb ssd

2tb seagate barracuda

Asus Rx480 strix 8gb. I know it is big bottleneck but I don't feel like upgrading my gpu right now.

Noctua D15 cpu cooler

Corsair cx650m psu

0 Likes
1 Solution

Yes! I fixed it. I opened the gpu and saw that there was almost not thermal paste on the actual dye. It was all around it and I have a theory that the thermal paste conducts small amount of electricity and it shorted something and made it unstable. I scraped off the thermal paste and placed it where it should be. Now it is running great but a little hot so I'll replace the thermal paste and buy noctua coolers. Someone had the same issue and apparently changing thermal paste fixed it, never thought I'd have the same problem. 

View solution in original post

8 Replies

I now tried to up the power limit to +50 and It persisted so I added a little bit of voltage and it seems to be fine. This should not be the case tho since I am running stock clocks.. I hope I haven't damaged my card but it looks like I have..

0 Likes

It certainly isn't impossible to damage. And if stock settings on the same drivers once worked with zero issues that's a big indicator. But since the Wattman drivers came out a little over a year ago, the default setting honestly don't work well for a lot of RX 480 and 580 users. I can't play any game on my RX 580 with default settings. Never could from day one getting the card. I now have to raise the power limit on all three of my older AMD cards HD 7870, 7950 and R9 380x. None will now run in current drivers without doing this. I absolutely agree stock settings should work, but they don't and AMD hasn't acknowledged this yet officially in over a year. Yet this is one of the most common fixes we giver here every day.

Here is to hoping you card is not damaged!

Also before deciding it is a hardware issue. Try that card in another computer if available. 

Thank you for the response.

I think I might've damaged my card tho sadly.. I am running 150% power limit and still getting artifacts then blackscreen and crash. I guess i'll install the latest driver again since i tried the oldest one I can get and up the voltage a bit and try again. Man I shouldn't have even tried to oc it..

0 Likes

Go into Radeon Settings / Gaming / Global Settings / Global Wattman: Raise the "Power Limit" slider to it's maximum. The black screens are a sign your cards power ceiling isn't high enough. This is extremely common or RX 4 and 5xx series cards and raising the power limit fixes this. If that doesn't do it alone you may still be throttling and a custom fan curve will help. I included my RX 580 config for reference. Only follow the Power Limit and Fan Settings as reference as your card wont handle my speed settings. 

Or if you are on pre 2019 drivers

Thank you for the response.

I have done this, my power limit is att +50 and I do have a good fan curve. Like I said it hit about 57c max if the temp meter is correct and not damaged. I guess i'll try upping the voltage a bit and try again.. 

0 Likes

I did replace the thermal paste with noctua one I got with the cpu cooler, I feel like I put too little on there though as there wasn't much left in the tube, I am thinking of buying thermal grizzly kryonaut but it might be a waste of money if my card already is damaged and I should get a new one.

0 Likes

Yes! I fixed it. I opened the gpu and saw that there was almost not thermal paste on the actual dye. It was all around it and I have a theory that the thermal paste conducts small amount of electricity and it shorted something and made it unstable. I scraped off the thermal paste and placed it where it should be. Now it is running great but a little hot so I'll replace the thermal paste and buy noctua coolers. Someone had the same issue and apparently changing thermal paste fixed it, never thought I'd have the same problem. 

I refurbished a GTX 750 with MX-4 and the card works well now. been abusing it with every demanding game I own to see if I can warm it up, it's no use, the card is stable now

0 Likes