GPU memory errors after power outage - is my card dying?

Did the errors start *right after* the power cut, or did they crop up gradually? Power outages can corrupt GPU memory state without actually damaging the hardware - I've seen similar issues where a clean driver reinstall or even just a full shutdown and restart cleared things up. If you're still getting memory errors after a reboot and your card's been stable before, it's probably just scrambled data rather than hardware failure. That said, if the errors persist across multiple reboots and stress tests, yeah, then something might've gotten physically damaged, but that's less common than people think.

Power surges during outages can mess with GPU firmware or VRAM state without killing the actual chip - try a full driver reinstall or BIOS reset on your card before assuming it's hardware failure. If the errors persist after that, run a memory test utility like MemtestG80 or GFXBench's stress test to see if it's reproducible or a one-off corruption issue.

Before you panic about hardware failure, try pulling the power cable completely, waiting a couple minutes, then booting fresh. Sounds silly but power outages can leave your GPU in a funky state where it's holding onto corrupted memory addresses, and a full power drain clears that out way better than a software reboot.

The thing people miss is that if your card is actually dying from surge damage, you'd usually see the errors *immediately and consistently* after the outage - every single time you stress test or run the same workload. But if you're getting intermittent errors, or they happened once and now it's working fine, that's usually corrupt state not hardware failure. Run something like MemTest86+ or your GPU's built-in memory diagnostic a few times to get a real sense of whether it's repeatable. One flaky test in a dozen runs points to state corruption; consistent failures across all runs means you've got a real problem.

If the errors keep happening, yeah, do a clean driver install like the earlier responses suggested, but also check your PSU. Power outages sometimes destabilize cheaper power supplies even after they're back online - they can deliver slightly dirty power for hours afterward without totally failing. If you've got another PC nearby you can borrow a known-good PSU from, swap it in for a test. That tells you pretty quick whether it's the GPU or something feeding it bad power.

Memory errors right after a power event are usually firmware or state corruption rather than actual die failure - the good news is that's often reversible. What folks sometimes miss is checking your PSU itself; if the outage caused a surge that got through, your power supply might be struggling to deliver clean voltage to the card, which would trigger exactly these kinds of errors even if the GPU is fine. Worth testing with a different PSU if you have one available, or at minimum check that all power connectors are seated firmly before you do any software troubleshooting.