Hello everyone, I've had an issue with my 3090 for the past week and a half.

The clock speed, 90% of the time, maintains 210mhz. This also happens under load, whether I stress it with a benchmark or a game. As also stated in the title, temps are well within range. Precision shows no temp scales above 63 C whereas gpuz does show hot spots at close to 75 under load with memory (I'm assuming this is video memory) hovering at 78 regardless of load. Total tdp doesn't go far beyond 50% as well. This happens even in my roommates PC with a completely different setup.

A minor fix I've found is to max out the fan speed which will let the clock scale as it should but it tends to hit a wall somewhere every second and the clock will temporarily drop for a moment.

I've removed oc software, I've ddu'd with windows updates off to no avail. I've reached out to support hoping for an rma but it's well beyond rma, worth a shot though. If anyone has any other ideas I'm willing to hear it. If you would like to see the gpuz output, just ask.

I have a spare 1080ti that works perfectly but it's frustrating to see this better card start to brick itself.

Yes. I've added it to the description but I assume it didn't update soon enough for you. My roommates PC is an Intel build compared to my and. He also has a 1600 psu compared to my 1200.

GPUz tells me that the perfcap reason is power. I've post 3 images below. They are in reverse order of when they were taken. The first image shows Baldur's Gate 3 running with the fan speeds forced to max at 3k rpm. Second image is BG3 but without the locked fan speed. Final screenshot is the PC shortly after boot.
 
Cannot figure out for the life of me how to post an image properly.
 

Twinkiepro
Small update

I changed the power target to 119 and the gpu target to 91. It now runs stable.

Hot spot reports 73 max, memory temp reports 80 max. Gpu temps are stable and below 63.

Perfcap is still power.

How much power are you drawing? Also your pictures are not showing.

Your card has dual bios. Try switching to the other bios to see if the problem continues on the other bios.

This problem is consistent between normal and oc. I've also updated the bios to the AD version to no avail.

precision allows me to lock it at +1000, so 1200 becomes new lower limit.

I tuned it while in elden ring and the game crashed. Restarted the game and it crashed again once i loaded in so thats not the way to go.
 
What does seem to work is setting the power and temp target to max, 119% and 91C respectively. at least it works for a time and then sputters a little.

Combining it with an at least 50% fan speed seems to really help though. I'm not sure how healthy this is for the gpu in the long run but also would there be anything else i can do? And if not, Could i also add a +200 to the speed to get a near boost output?