Hi everyone, 
 
Hope you are all well and looking for some help from you please :). I recently purchased a Evga 3090 Ultra and after installing the precision software everything has been running great. Recently I noticed I could up the power target to 107% and this is where I started noticing issues with my computer randomly restarting during gameplay with no warning. 
 
If I leave leave the power target at 100% I have no issues what so ever. During gameplay this happens at very random times and there is no clear pattern or thing that I do to make it happen. So far I have done the following - 
1. Swapped the PCI cables between the graphics card and PSU. 1 PCI cable per plug. 
 
My specs are 
10900k @ 5.1 all cores
asus rog maximus xii extreme
EVGA 3090 Ultra
Seasaonic Prime Titanium 1000W PSU
 
I know this is first world problems and 100% power target is probably fine, however I figure if I've paid for the priveledge I might as well give it a go. 
 
I was thinking about testing another PSU next and if it still does the same random reboots, looking to RMA the card? 
 
Your advice is greatly appreciated. 
 
Cheers
 

I had been experiencing a similar issue with my 3080 FTW3 Ultra, and had also tried swapping PSUs to see if that was the issue. I also (currently) have a Seasonic Titanium 1000W PSU, Intel processor, and Asus motherboard (full specs are in my signature below).
 
I filed an RMA and am awaiting a replacement GPU at the moment. I'm hoping that fixes the problem, otherwise, it could be my motherboard. 
 

I've tried disabling XMP and also going back to stock CPU clock and voltage settings. I've also got my most recent motherboard BIOS. Also ran memtest86.

At first this only seemed to happen in Cyberpunk 2077, and then only when I did something to pause the game, never during gameplay. But then it got worse and started rebooting my PC in Heroes of the Storm and Europa Universalis IV.
 
***Update: I got my replacement in today. Black accents (not the red lips this time), serial number first 4 digits 2114, packaged in USA. Just for reference, my previous card was: serial number 2014, made in Taiwan. Everything seems to be okay; no reboots yet with OCCT 3D stress test--tested on shader complexity levels 3, 5, and then 8. I'll be fairly satisfied that the rebooting issue is fixed if I continue to pass gaming/benchmarks/stress tests over the next week or so. 
 
Initially, I had a loud clicking noise coming from one of the fans, but I updated the firmware to the latest and it went away, although the standard fans ramping to 100% during the firmware update scared me...the clicking was very, very loud. But then I rebooted my PC myself, and the fans always spin up during reboots, just as they do during power on, and they didn't make any noises then. I ran the stress test and a few benchmarks and sometimes I'll hear a faint clicking/rattling when the fans ramp up, but it stops. I've checked that I don't have anything up in the fan shroud/heatsink, but I could check again I suppose. I've checked to make sure all fans are running at roughly the same RPM, and they are, so I don't think any of them were damaged. 
***

You have no idea what his temps are as he didn’t state them or case airflow. This has been a known problem since launch. I RMA’d my first card back in October because of this issue. That was in a well ventilated case, 3 different power supplies, and two motherboards. If I’m buying an overclocking card I’d expect it to at least handle its own voltage range. This issue is all over the forums already. RMA that sucker and get what you paid for OP.

Voltage range is not the same as the amount of current that the card is requesting and using. (Power) Wattage is a product of both current and voltage.

There’s also a lot of whiners on this forum thinking they are entitled to something. You bought an OC card because it’s overclocked from the factory already. Anything else is dependent on silicon lottery and is a bonus if you can get more out of it. If the goal was simply to say I can get +500mhz out of the card, manufacturers would just market their cards with a lower rating just so you can claim gains of 500 if it makes you feel better. Being able to do factory overclocks at all is a privilege.

With your logic, if the slider goes up to +1000mhz in X1 do you feel entitled to use 1000? I’m sure EVGA and anyone else will tell you to get bent when you don’t get that speed.

What I do know is that more power creates more heat. Heat definitely causes the card to crash, especially pushed to the limit. And to get the most performance out of the card, you don’t have to rely on the power slider to be at the max. It’s overkill until you’re hitting higher and higher clock speeds. You might even get better performance if you undervolt and overclock because of the better thermal range the card operates in.

Simply turning up the power target doesn’t mean more performance. If you ONLY do that, your byproduct is heat. You need a combination of higher clock speeds and higher power targets to better use that excess power, as well as cooling capabilities. In my testing, temperature was the biggest factor in maintaining a high boost clock.