The GPU usage is a now-known nvidia driver bug, you need the new hotfix driver to fix that
Same with GPU clocks getting stuck after quitting games and causing issues - you just need to update your drivers
Well, I switched to the other pcie cord from the psu, put the gpu back into the 1st pci slot on the motherboard, and installed the hotfix driver. Ran a few games, got great fps, gpu was throttling well. I ran a benchmark test in Warhammer: Vermintide and things were going well, averaged 123 fps. I joined a match and watched my gpu chip power draw drop to 20w and the gpu clock fell to 400 MHz (was 1780 MHz prior)
Attached is the picture of the wiring that you requested.
Thanks for your continued help with this issue!
What would you say about such high power consumption of the whole card while the GPU is just sipping power? It's not like the board has a synchro-mesh chicken plucker on it.
Anyhow - try installing the latest driver (526.xx) and choose "Custom install" and under that tick the "Clean Install" box, so driver itself cleans old installation. You can also deselect "GeForce Experience" if you really don't need it.
I followed all those steps and same issue arose. I also went into nvidia settings and made sure the 3060ti was the selected card and that power settings preferred performance.
I would also like to a picture (or four) of how this computer looks internally. That being said, here are some general thoughts.
You should never move a gpu from the primary pcie slot before you know how the secondary pcie slot(s) are wired in the motherboard. If it is not a custom job (like the junk Dell makes) you can find the motherboard manual online. This should show you which slot is the primary. Then make sure your gpu is in the primary slot. For a quick troubleshooting test you can try another slot, but for any other use always stick to the primary slot. If the slot does not work, then either your motherboard or cpu is broken.
Make sure the 8-pin power connector is properly connected. Asus says that psu has 4x pcie 6+2-pin connectors, I would assume these are two physical cables with two connectors each, for a psu of that wattage. So make sure you try the other cable coming from the psu, not just the other connector on the same cable. If they are pigtails, use the end connector. Do NOT mess around with over-/underclocking, flashing bios, and other stupid ideas like that, unless you are certain you know what you are doing and the computer is no longer under warranty. Because, depending on the age of the computer, I would look into the warranty situation here.
I have to say that second screenshot almost looks bugged honestly. 150W through the pcie slot would probably melt some traces on the motherboard if it were correct.
Do you get similar readings with hwinfo?
And finally, the drivers can mess some of this up. Not the 150W from the pcie slot, if that is a correct reading, but the other issues. Follow pavles advice above, to clean install the latest hotfix driver. A clean install of Windows is not out of the question either.
This just means the gpu is hitting max voltage before the power limit. It is normal behaviour for modern gpus to boost until it hit one or more limits. This does not have to be the power limit.
I posted a picture of the inside of the case below. I'm more than happy to take additional pictures if you can't see what you're looking for!
The GPU is back in the primary pcie slot and I recently tried the other pcie cord from the psu. I currently do not have the end connector attached to the GPU but I'll try that now. (edit: tried it, no change)
I haven't messed around with any overclocking, bios flashing or any other foolish ideas (I think)
That second screenshot is what my GPU looks like most of the time, I watched the gpu operating normally and then fall into it's "bugged" state about an hour ago. It usually occurs about 15-20 minutes after a fresh restart. I also confirmed the values in hwinfo.
I've also got the newest hotfix driver installed and I wiped the computer with a fresh install of windows when I first starting noticing the issue.
Thanks for all the notes! I'm going to send the PC back to ABS under warranty if I can't get it figured out by the end of the week. I was kind of hoping to figure out that the power supply was the main culprit, as that would be faster and cheaper to replace than having to pay the shipping for the whole rig.