Sajin
Which power supplies are you running?

iamWalk
I ordered a Corsair AX850 titanium cert, which I read in another thread is on the unofficial "this is working" list that one user took the time to put together, just because if this is a power supply issue I don't want to risk anymore bs. 

iamWalk
Hello, Recently got a 3090 FTW3 ULTRA.
 
I plugged it into my PC, and it worked fantastic for 2.5 hours. Then, I received a crash while playing Halo(very low utilization as it's a potato game, and was frame locked to 165), followed by "no video signal". Restarting did not fix, and when I put in my old graphics card, there was no signal either - the motherboards PCIEx16 slot was straight up dead.
 
Lucky, I had a backup PC, and put the 3090 in here, it still worked. However, I just received the exact same issue - a crash followed by no video output from GPU. This time, the computer restarted and output resumed. But I'm worried that there will be a next time, and it will also kill my second PC. Two crashes, all within 2.5 hours of use on two different PCs, on a game that barely stresses the GPU, one which resulted in a straight up destroyed mobo seems to be a major problem.
 
To clear up any questions I know will be asked - both PC's had their nvidia drivers cleanly installed for the 3090 prior to use. Both had different PSU which met the power requirements, each is 750W. Both were seated/installed correctly. Both PCs had run w/ a 1080(bought shortly after it was released) prior to the 3090 and have never had any sort of hardware failure related to the GPU during they're use. Both PCs were top of the line, first one was using a 10900k, second a 7700k, with rest of hardware matching quality. Temperatures were closely monitored and never got close to anything I would worry about, for both CPU and GPU. I don't have graphs, unfortunately.
 
edit: I ran a bit of a benchmark stress test - maximum temperature never exceeded 75 under full stress load(99%), which is the same as my ingame experience when I first got the card. As said earlier, the crashes described occurred instead under very light load, under 10% due to Halo being a potato game. I will run a continuous test overnight at 10% GPU load and update this post in the AM. 
 
What I'm looking for:
 
1) An explanation as to what is causing these crashes, as I see a ton of issues that appear similar when I google it, hopefully from EVGA.
 
2) A solution, if one exists, which I have not found in my googling.
 
3) Some advice on how to test power draw / temperature in a controlled way(haven't done much benchmarking in the past so would like some advice on which software is best for this)
 
4) Some reassurance that this graphics card will not fry PC #2, especially after I just experienced the same crash which fried my extremely expensive, relatively new 10900k machine. I already ordered a replacement mobo, and would very much like to know that I can put my 3090 back into my computer and not end up just frying another $250 dollar motherboard because of some absurd crash. 

brendan.alyssa

iamWalk
Hello, Recently got a 3090 FTW3 ULTRA.
 
I plugged it into my PC, and it worked fantastic for 2.5 hours. Then, I received a crash while playing Halo(very low utilization as it's a potato game, and was frame locked to 165), followed by "no video signal". Restarting did not fix, and when I put in my old graphics card, there was no signal either - the motherboards PCIEx16 slot was straight up dead.
 
Lucky, I had a backup PC, and put the 3090 in here, it still worked. However, I just received the exact same issue - a crash followed by no video output from GPU. This time, the computer restarted and output resumed. But I'm worried that there will be a next time, and it will also kill my second PC. Two crashes, all within 2.5 hours of use on two different PCs, on a game that barely stresses the GPU, one which resulted in a straight up destroyed mobo seems to be a major problem.
 
To clear up any questions I know will be asked - both PC's had their nvidia drivers cleanly installed for the 3090 prior to use. Both had different PSU which met the power requirements, each is 750W. Both were seated/installed correctly. Both PCs had run w/ a 1080(bought shortly after it was released) prior to the 3090 and have never had any sort of hardware failure related to the GPU during they're use. Both PCs were top of the line, first one was using a 10900k, second a 7700k, with rest of hardware matching quality. Temperatures were closely monitored and never got close to anything I would worry about, for both CPU and GPU. I don't have graphs, unfortunately.
 
edit: I ran a bit of a benchmark stress test - maximum temperature never exceeded 75 under full stress load(99%), which is the same as my ingame experience when I first got the card. As said earlier, the crashes described occurred instead under very light load, under 10% due to Halo being a potato game. I will run a continuous test overnight at 10% GPU load and update this post in the AM. 
 
What I'm looking for:
 
1) An explanation as to what is causing these crashes, as I see a ton of issues that appear similar when I google it, hopefully from EVGA.
 
2) A solution, if one exists, which I have not found in my googling.
 
3) Some advice on how to test power draw / temperature in a controlled way(haven't done much benchmarking in the past so would like some advice on which software is best for this)
 
4) Some reassurance that this graphics card will not fry PC #2, especially after I just experienced the same crash which fried my extremely expensive, relatively new 10900k machine. I already ordered a replacement mobo, and would very much like to know that I can put my 3090 back into my computer and not end up just frying another $250 dollar motherboard because of some absurd crash. 

How did it fry your PCIE?

Kylearan
BTW there is something worth noting.
 
On all drivers AFTER 456.98 hotfix, there seems to be a memory temp "blip" which causes 6500C memory temp for a split second, triggering a "thermal" flag, which doesn't seem to happen at full load--only at idle or during a load change from heavy load to light load.  This was discussed in the main 3090 FTW3 vbios thread by that person with that girl smiling avatar picture, a few pages back in that thread.  I do NOT know if that's related to these cards dying in Halo or not though!!
 

lemmein

Kylearan
BTW there is something worth noting.
 
On all drivers AFTER 456.98 hotfix, there seems to be a memory temp "blip" which causes 6500C memory temp for a split second, triggering a "thermal" flag, which doesn't seem to happen at full load--only at idle or during a load change from heavy load to light load.  This was discussed in the main 3090 FTW3 vbios thread by that person with that girl smiling avatar picture, a few pages back in that thread.  I do NOT know if that's related to these cards dying in Halo or not though!!
 

The memory and pwr temp blip was always been there for me since November 4th is when I first got my 3080 ftw3 ultra. It seems to happen to all the mem and pwr in HWINFO and only when its hot enough it shows weird spikes as such. Till this day I'm still confused at why this is happening I hope it's not faulty sensors.