I had a similar issue in all games, turned out it was my RAM and not the card. Suggest trying with a single stick at low speeds to ensure it's not the RAM and from there move on if this is a new PC and you didn't have another GPU in there that was working flawlessly.

that

It shows artifacts in furmark. I guess the GPU is just defective. Running a benchmark seems utopic, it does not even run AOE II without crashing - It opens but lags and has missing textures until it crashes a few seconds into the game.

Something is defective but it might not be the gpu itself. Did you just build this system? If not was it running another gpu? I currently suspecting power cables though that may be a wild goose.

Do ypu have a PSU tester? If so, Is it the old kind with just idiot lights? Or the new kind that outputs the actual voltages?

 

• Seasonic Prime Gold GX-1300 (W)

 

A 550 watt power supply? That's obviously the issue lol...

 

It shows artifacts in furmark. I guess the GPU is just defective. Running a benchmark seems utopic, it does not even run AOE II without crashing - It opens but lags and has missing textures until it crashes a few seconds into the game.

 

It seems to be defective. If you can't give it back, krisfix germany could try to fix it. But it will cost you some extra money. The other try could be a RMA if there is still a guarantee on the card.

I mean, if the Card wasn't handled properly (no support bracked used by horizontal mounting or something like that), there is a chance that there is a hardware problem like cracked soldering balls on the VRAM or something like that. That is the problem when you buy used stuff from people you don't know. I feel sorry for you.

 

NWR on a video stated 3dmark seems to show issues but furmark would work fine. He said that particular gpu had dead cores. Also heaven benchmark would artifact constantly while running.
lots of these types of issues a reball of gpu seems to fix. If you can send back gpu to seller i would do it now.
For something quick you could reflash the bios again from techpowerup bios library.

 

I had a similar issue in all games, turned out it was my RAM and not the card. Suggest trying with a single stick at low speeds to ensure it's not the RAM and from there move on if this is a new PC and you didn't have another GPU in there that was working flawlessly.

 

win 11 ? updat 24H2 ??? **** a lot pc

 

View attachment 2692161

That black sleeved cable's connector (+12v pin on the right is bent open and won't be making full contact) is damaged. I wouldn't use it, but it's probably not the source of the issue here.

it is definitely some power delivery issue.

Can't rule out a power issue, but he lack of power draw is quite possibly a symptom of any meaningful load crashing before it can demand more power, rather than it crashing because it's not getting power it demands.

I would set a fixed underclock (with the curve editor in MSI AB, where you can lock the card to a given voltage and force it to run at full speed irrespective of load) and note the point at which 3D loads start to crash.

Chances are you're going to want to send the card back to the seller while it's still practical to do so.

 

I've seen guys that could take the warranty sticker of the retention bracket so well and either mess with the card or just simply do water-cooling and in the process damaged something they either were or weren't aware and afterwards putting the cooler back on, you couldn't even tell. I would take it look under the cooler. Get a good magnifying glass and inspect the PCB and all components, you might be surprise.

 

I did that already as stated in my first post. Benchmarks don't run at all, except of furmark which seems to run and draw power but shows weird artifacts.

 

Try this: Zero/reset the BIOS and don't OC the RAM or anything else. Now run that game and see what happens. Also, try a few 3DMark benchmarks then report back.

 

A X670E is one of the most modern and high end chipsets you can buy right now. It of course downclocks but the PCI version should not really affect that it can run. How else would they have tested the 4090s PCIe scaling with versions down to 1? As I wrote, forcing 4 in BIOS does not change the issue.

 

why does it say you are runninig 16x1.1, should be 4.0 maybe even 3.0 but 1.1 means a limited mb. try it in a modern board and see what happens then rma if needed.

 

They have included a mini-test that kicks it into highest "pcie" gear for testing:

yeah - that´s its "start" button