Question Gigabyte RTX 2080 Super dying

eternalabys

Honorable
Nov 8, 2018
155
5
10,615
Hello everyone,

As the title reads, I recently got a GPU (Card in question is a Gigabyte RTX 2080 Super Gaming OC) for free from my cousin as it seemed to have an issue seemingly out of no where after he came back after 3 weeks of not being home. He tried playing Genshin and it kept on black screening after a minute, sometimes even less.
Any kind of 3D load on stock settings makes the system black screen and forces the GPU fans to go 100%, audio continues to play in the background however.
As far as I understand, it's either the Core (highest likelihood), or the VRM's. VRAM is perfectly fine as it has no issues with artifacting.

On desktop or 2D /3D very light games, it works fine. But if I try and play Far Cry 5, EFT, basically anything harder to run than osu!, it black screens and needs to be power off.
Only way I've gotten it to run the higher load games is to neuter the Core clock to -650MHz and Power limit down to 60% via EVGA Precision X1.

Aside from that, main reason I believe it's the Core dying is due to the fact it had never been repasted or maintained in any other way aside from dusting it out once in a while (a year or more) since it was bought back in 2019. Before repasting it, I downclocked the GPU and ran Furmark, which almost immediately made the temps shoot to 88C, at that point I switched it off and went ahead with repasting and cleaning the card, temps dropped down to 60-70C under load, but whenever it reached close to 180-200W power draw, it would black screen.
His PC has a Corsair RMx 750W PSU, mine has a TX 550W. So it's not a power delivery related issue as his PC is running fine on my current GPU (EVGA GTX 1080 SC). Reset Bios and reinstalled drivers and it did nothing, so it's definitely a hardware issue.

Is this GPU worth attempting to fix? Is there anything I missed that might be causing the issue?
 

eternalabys

Honorable
Nov 8, 2018
155
5
10,615
Is the issue present when installed on the PC with the Corsair RMx 750W PSU?
Yes, it was in my cousins build before we switched GPU's. And like I mentioned, the PSU isn't the issue, HWinfo shows that the GPU isn't even using close 200W, my GTX 1080 drew 230W on full load. My PSU is more than capable of handling the 2080 Super in it's current state, even if it drew 300W, my CPU doesn't get close to 100W, so I still have overhead. It is a decent quality PSU
 
2080's have bad power spikes which is probably causing your issue and the PSU cant handle it. If everything is fine till you get into demanding games its a power delivery issue. You need to take into account what the GPU, CPU, and the rest of the board is doing. 2080's are rated at 250w which is half of your PSU's wattage add in the CPU and then the power spikes of the 2080 and you have a recipe for disaster.

I had the same issue with my 2080 Ti with a EVGA 850w G2, i just so happen to have a 1600W sitting in an old case and havent had an issue since. I have yet to repast my 2080 Ti and run it at 124% power target averaging around 2000Mhz on the clock which is well above the boost speed.
 

Phaaze88

Titan
Ambassador
Any kind of 3D load on stock settings makes the system black screen and forces the GPU fans to go 100%...
Was about to suspect power, UNTIL...

audio continues to play in the background however.
That's memory. Artifacts aren't the only sign of bad memory.
Memory failures seemed to be more common with the 20 series too.


Is this GPU worth attempting to fix?
Might not be. Especially if the memory chips require a reball - that's not without risk.
Taking it to repair shop is your call though.
 
Yes, it was in my cousins build before we switched GPU's. And like I mentioned, the PSU isn't the issue, HWinfo shows that the GPU isn't even using close 200W, my GTX 1080 drew 230W on full load. My PSU is more than capable of handling the 2080 Super in it's current state, even if it drew 300W, my CPU doesn't get close to 100W, so I still have overhead. It is a decent quality PSU
When the GPU was repasted, did the thermal pads were also replaced?
If the pads are no longer touching the RAM, VRMs, capacitors, etc. that could cause the GPU to crash.
 
Last edited:

eternalabys

Honorable
Nov 8, 2018
155
5
10,615
When the GPU was repasted, did the thermal pads were also replaced?
If the pads are no longer touching the RAM, VRMs, capacitors, etc. that could cause the GPU to crash.
I didn't replace the pads, but they were clearly squished between the VRM's, VRAM and heatsink. And I doubt heat would be the issue as even if it's a cold boot in the morning, if I even try loading up furmark, Heaven or anything of the sort, it blacks out immediately. Temps are also reported to be well under control in HWInfo.
 

eternalabys

Honorable
Nov 8, 2018
155
5
10,615
2080's have bad power spikes which is probably causing your issue and the PSU cant handle it. If everything is fine till you get into demanding games its a power delivery issue. You need to take into account what the GPU, CPU, and the rest of the board is doing. 2080's are rated at 250w which is half of your PSU's wattage add in the CPU and then the power spikes of the 2080 and you have a recipe for disaster.

I had the same issue with my 2080 Ti with a EVGA 850w G2, i just so happen to have a 1600W sitting in an old case and havent had an issue since. I have yet to repast my 2080 Ti and run it at 124% power target averaging around 2000Mhz on the clock which is well above the boost speed.
Thing is, my cousin had this card in his build for 4 years, his PSU is fine as he's running my current GPU without any issues, and as I said, it draws close to 250W with an OC. Both of our PSU's are fine and the issue persists across both systems. It's a component issue with the GPU, I'm beginning to suspect it's the VRM and not the core as with 50% reduced power it can play demanding games like Tarkov and Far Cry 5 for quite long periods of time with no issue.
 

eternalabys

Honorable
Nov 8, 2018
155
5
10,615
Was about to suspect power, UNTIL...


That's memory. Artifacts aren't the only sign of bad memory.
Memory failures seemed to be more common with the 20 series too.



Might not be. Especially if the memory chips require a reball - that's not without risk.
Taking it to repair shop is your call though.
I believe Micron had most of the issues with VRAM failues, while I was repasting I checked to see if it was one of the bad batch chips from them, but the memory is all from Samsung. Not to mention I clocked down the memory clock, yet it still seems the card dies if I give it anywhere between 65-75%+ of it's original power. VRM and VRAM repair is a tricky thing so it really might not be worth it afterall, at least it's not the Core lol