Question R9 290 Unstable at stock (Memory) frequencies. HELP!

[novalyfe]

Hey guys,

So I got a R9 290 reference card thats unstable at the stock frequencies of 947Mhz core and 1250Mhz memory. I monitored the card behaviour and found that the memory frequency jumps from 150 to 1250Mhz every second or 5. Via MSI Afterburner I downclocked the memory to 1001Mhz making it more stable, but every so often the frequency spikes to 1250Mhz again risking a blackscreen and hard reset of my entire PC. The AMD Driver performance tuner doesnt let me lower the Max. frequency of the card. What I think is that the AMD Performance tuner AND MSI Afterburner and each taking a turn with the frequencies of the card.
As you can see in the image below, the Max. memory frequency spike is 1250Mhz eventhough my MSI Afterburner profile is set to a max of 1001Mhz. What causes these spikes? And how do I turn down the Max. frequency in the AMD performace tuner?

Specs:

i5-3570K OC to 4.1Ghz
Gigabyte z77xd3h
16GB Corsair Vengeance 1333Mhz DDR3
R9 290
OCZ modxstream pro 500w
The most recent AMD Driver 20.2.2

https://pasteboard.co/IY9OaAq.png

Could't upload the image differently so I posted the URL, not sure if allowed (Sorry in advace if not allowed)
I get a "something went wrong" error when I try to upload an image.

 

[ChumP]

Have you bought the R9 290 used?
Did the card work normally beforehand?

 

[novalyfe]

Have you bought the R9 290 used?
Did the card work normally beforehand?

I got the card from my girlfriends cousin who said it used to work properly but stopped working properly after one year of him using it. The same black screen issues bugged him. He bought the card "refurbished" from a respected hardware store in the area we live in.

 

[anotherdrew]

@novalyfe
I had an R9 290 for about 3-4 years. It runs hot. Fans were always screaming and temps were through the roof. What you describe could be a heat issue (it could also be damaged VRAM). IF you want to use this card and are willing to spend $30 and are willing to say "Oh well" if you kill the card ... I would consider re-pasting it. The thermal paste manufacutres use is cheap and they smear it everywhere. On top of that, the thermal pads they use on the memory chips is low quality. There are a number of good guides online .. I would read a couple to get a good idea of what you are in for. Oh, a fan pointed at the back of this card also helps.

I re-pasted and brought down my temps about 10 C.

As an alternative, set a max frame rate. It keeps your card from working as hard and limits heat build up.

 

[novalyfe]

@novalyfe
I had an R9 290 for about 3-4 years. It runs hot. Fans were always screaming and temps were through the roof. What you describe could be a heat issue (it could also be damaged VRAM). IF you want to use this card and are willing to spend $30 and are willing to say "Oh well" if you kill the card ... I would consider re-pasting it. The thermal paste manufacutres use is cheap and they smear it everywhere. On top of that, the thermal pads they use on the memory chips is low quality. There are a number of good guides online .. I would read a couple to get a good idea of what you are in for. Oh, a fan pointed at the back of this card also helps.

I re-pasted and brought down my temps about 10 C.

As an alternative, set a max frame rate. It keeps your card from working as hard and limits heat build up.

Thanks for the massive reply the issue with heat I saw that today! FurMark stress test showed a 94c temp so I'm definitely considering doing what you said.
However even if the card is at idle, and around 50c, it will sometimes spike to 1250Mhz and crash.
I'm starting to think either the Vram is damaged or the card on its whole is showing its age.
Is damaged vram a thing you can test? Either through software or just opening the card up?

 

[anotherdrew]

Thanks for the massive reply the issue with heat I saw that today! FurMark stress test showed a 94c temp so I'm definitely considering doing what you said.
However even if the card is at idle, and around 50c, it will sometimes spike to 1250Mhz and crash.
I'm starting to think either the Vram is damaged or the card on its whole is showing its age.
Is damaged vram a thing you can test? Either through software or just opening the card up?

You can't see damaged RAM (unless it's just totally fried). There are some apps that claim to test VRAM, but I don't know of one that I would recommend.

The 290 was a capable card, but it ran hot. The spiking upto to 1250 MHz is a symptom of the card working hard ... nothing more. Limit the work it does and you will limit the memclock speed.

Time and temperature may have already wreck havoc on your card ... or you might get some more time out of it if you can get temps down (and/or limit how hard it works). Is it worth a couple dollars to give it shot or is it time to drop $300-$400 on something newer and faster ... you're call.

 

[novalyfe]

You can't see damaged RAM (unless it's just totally fried). There are some apps that claim to test VRAM, but I don't know of one that I would recommend.

The 290 was a capable card, but it ran hot. The spiking upto to 1250 MHz is a symptom of the card working hard ... nothing more. Limit the work it does and you will limit the memclock speed.

Time and temperature may have already wreck havoc on your card ... or you might get some more time out of it if you can get temps down (and/or limit how hard it works). Is it worth a couple dollars to give it shot or is it time to drop $300-$400 on something newer and faster ... you're call.

Its this behaviour thats confusing me:

https://imgur.com/khOQfW3

View: https://imgur.com/khOQfW3

https://imgur.com/0cY4mpa

View: https://imgur.com/0cY4mpa


In the second image I drew around the readings where my screen became black, the PC temporary froze and I had to wait to my PC becoming stable again. As you can see theres massive spikes in GPU Usage eventhough it was idling. Both the Core AND Memory frequency spiked to 947Mhz and 1250Mhz respectively eventhough my MSI Afterburner profile that was loaded is set to 856Mhz and 1001Mhz respectively. So the million dollar question for me is: how does this happen? Why does the card spike to full load causing it to be unstable EVEN WITH an Afterburner underclock profile loaded? The temperature reading is on the higher side for idling, but not lethal for the card. Please does anybody have answers?

 

[anotherdrew]

I would uninstall Afterburner and try to use just the AMD software to control your GPU (I haven't used it in a long time, but I know it use to allow that).

I would also pull up your task manager, click performance and then GPU ... find out what is using the GPU when everything is closed. It really should be nothing.

 

[novalyfe]

I would uninstall Afterburner and try to use just the AMD software to control your GPU (I haven't used it in a long time, but I know it use to allow that).

I would also pull up your task manager, click performance and then GPU ... find out what is using the GPU when everything is closed. It really should be nothing.

The AMD software doesn't allow me to lower the memory frequency below 1250Mhz, the result being an unstable card.