I built my desktop a few years back and upgraded my GPU from a 4090 to an MSI Ventus 5090 about 4-5 months ago. Up until recently, everything was running smoothly. However, when I tried playing Arc Raiders for the first time over the weekend, my game crashed repeatedly after just a few minutes, and I experienced wild frame rate fluctuations. Initially, I thought the game itself was to blame, but after running a 3DMark test, I was shocked to see it crash after only 5 seconds. Even when it did work, my score was less than half of the expected performance for a 5090.
I attempted to troubleshoot by rolling back my Nvidia drivers, clean installing Windows on a spare drive, and reseating all components — yet nothing seemed to help. While monitoring everything with HWInfo, I noticed that when I put the card under load, it would initially spike to over 570W but then drop to around 300W after just a couple of seconds, leading to stuttering and crashes.
Thinking my power supply might be the issue, I replaced my old Thermaltake 1350W PSU with a new Seasonic 1200W. After installing the new PSU, I was able to run 3DMark multiple times successfully, and Arc Raiders played smoothly without any crashes. However, after a fresh Windows install, I started experiencing the same power drop issue again. My 3DMark scores plummeted, and although Furmark should push the card to maximum, I saw similar wattage drops with GPU usage halving.
Everything else seems fine — the card doesn't overheat, and the new PSU should handle the load without problems. Now, I'm wondering if the card might be damaged from the old PSU. Before I go down the RMA route with MSI, does anyone have suggestions? I've already tried various Nvidia driver versions, different loads, reseating everything, and a fresh Windows install, ruling out issues with the install or thermal throttling.
3 Answers
Have you tried forcing the PCI-E1 into Gen4 mode? Also, check how many lanes your PCI-E slot is currently using; that might affect the card's performance too.
It sounds like your MSI Ventus might be suffering from overheating FETs, an issue that’s been reported before with that model. Unfortunately, they’ve had a reputation for overheating problems across several generations. Have you checked for any sensors that could help confirm this?
It might be possible that your outlet can't handle the power draw. Have you tried moving it to a different outlet? Sometimes, even if it used to work fine, there could be issues with the current one you're using.
Good point! I did switch outlets just to test and it ran fine for a while with the 4090, but I'll definitely keep that in mind.
Interesting! This is the first I've heard about it. I was actually wondering if the VRMs were overheating. I tried running Furmark rapidly to see if the card would throttle quickly, but it managed to hit 570+ before any problems showed up.