Last month I had to upgrade from a GTX 1080 to a GTX 2080 super hydro copper as my GTX 1080 died. Everything is working fantastic and as expected from reading reviews. Recently I just got a bunch of new radiators for my build so after finishing my loop . My card was now running 15C cooler from its original 60C to a nice 45C I reapplied my OC of +70 on the core and +500 on the memory. I ran a few stress tests and pasted then (3dmark, real bench, superposition) and it wasn't until I decided to look into running something for a memory test that I run into problems as when I use programs such as occt, MemtestG80, and MemTestCL that my system completely locks as the screen "artifacts up". Besides I have had this card for close to 2 months now and have had no problems unless I use a VRAM stress test. The only way I can run said test is by downclocking the memory by -500.

The most frustrating part is I am able to load all 8GB of VRAM up and run the card under 100% load while not running into and problems. So is it software related or is it actually hardware related? Or does anyone have any suggestions for me?

Looks like bad memory, so you should place original waterblock on card and rma it.

And today I did every test and benchmark again and they all past. This was over 8-9 hours

CjGhost
And today I did every test and benchmark again and they all past. This was over 8-9 hours

Conclusion, software related gremlin. It not impossible memory timings were corrupted in the driver. Most folks do not know this but the driver contains hundreds of MB of bodges to get around game/application issues. Memory sub-timing is one thing the drivers can effect.

Did you put thermal paste on both the gpu die and block as I had this problem until I striped it down and redone the paste.

clone
Did you put thermal paste on both the gpu die and block as I had this problem until I striped it down and redone the paste.

It came preassembled. So I did not put thermal paste on the GPU.

Before I took the stock cooler off the thermal paste wasn't the best it may be worth a look .

The thing that bothers me though is it only has problems specifically with VRAM stress test something that will go through and check if it's bad.

CjGhost

clone
Did you put thermal paste on both the gpu die and block as I had this problem until I striped it down and redone the paste.

It came preassembled. So I did not put thermal paste on the GPU.

On second thought after reading your original post again, perhaps you had air in the block and it had a hot spot. Your model doesn't have memory temp sensors, right?

It does, the hottest the modules get are at most 45-50.