Hi everyone,

I just got my 4090 FE a few days ago (after a 2 month wait for them to come back in stock where I live) and I’m running into some really puzzling issues. Hope maybe someone can help or I can help others because as far as I can see there are not many people reporting this.

I bought this card for work (I do 3D stuff) and first thing I did was load up my C4D+Octane combo which had been working flawlessly thus far. From the first render I started experiencing random crashes, sometimes midway through, sometimes renderer would not start and throw out a CUDA 700 error. After a few frustrating days trying everything including but not limited to:
- wiping the computer and doing a clean win 11 install with just C4D and Octane installed
- wiping the computer, downgrading to windows 10 and installing just C4D and Octane
- Trying different versions of the Nvidia Studio driver, cleaning it out with DDU properly every time
- Swapping PSUs (Seasonic GX-1000 and Seasonic Vertex PX-1000, both fully compatible with ATX 3.0 and 12vhpwr)

I thought I’d change my approach and try to test things in a different way since my system is running just fine with no errors with my 2080ti. So I tried to stress test the card and I ran Time Spy and it’s simply crashing at random points with the error in the picture. I could never complete the benchmark because of this. The crashing pattern is random, it crashes once, tries to continue and then crashes again with the error. Superposition ran ok twice but third time crashed as well.
After looking around I found a memory test for GPUs: Memtest Vulkan on Github that seems to find errors in the VRAM. Running it on my 2080 gives no errors.
My questions are:

- are there any other ways to check the VRAM of the card?
- has anyone here encountered this issue before?
- is it time to give up and RMA the card?

System specs:
i9-13900,128 GB ram Corsair, 1TB Samsung 990 PRO SSD, 1000W Seasonic ATX 3.0 PSU, Windows 10 & 11, everything stock, no OC.