Black screen detected along with maximum fan operation.
Black screen detected along with maximum fan operation.
I received a functional 3080 and after upgrading, it crashes when fully loaded. Details: Ryzen 5 5600X with MSI B450I AC Corsair Vengeance 2x16Gb, 3200Mhz GPU. New PSU and GPU. Performance: Pure Power 13 850W, 80+ Gold, running Win11 24H2. The graphics card fades out completely under load, causing the whole system to shut down. Fans reach maximum speed and stay at 100%. Crash timing is inconsistent—sometimes nearly five minutes without a failure, other times after about one minute. I’ve followed troubleshooting steps: updated Windows, BIOS, chipset, and all drivers to latest versions; installed graphics drivers via Nvidia site; limited RAM speed to 1750Mhz, switched RAM slots; tested with two 16Gb modules from separate PCIe cables. The crash occurred when using the old PSU. The new PSU worked without issues. I spent nearly five hours trying to fix it. Any advice would be greatly appreciated.
Which GPU is it? (You can update the firmware of your MSI 3070 using the MSI Center tool) Install HWinfo64, run its "sensor only" feature. Can you observe the GPU temperature, hotspot and VRAM under load? A repass and repaste could be useful. Edited October 12, 2025 by leclod
Micron indicates the GDDR6X max operating temperature reaches 110°C, meaning it should function properly
To clear doubts about your memory, try a brief Testmem5 or a thorough Memtest86. You might reinsert the GPU and reconnect the two PCIe power cables (also on the PSU side). You could reduce VRAM speed using Afterburner (though I still think VRAM is involved). Lowering the GPU voltage (and capping frequency) is straightforward—it cuts power use without hurting performance. Edited October 12, 2025 by leclod
Several 3080 and 3090 models experience VRAM temperatures near 100°C. This is generally expected for these graphics cards. It doesn't necessarily mean something is wrong, just typical behavior.
It's quite confusing. The only thing I can suggest is testing a different PSU with the old graphics card to see if the new one works. If it does, it might mean the graphics card is the issue. Reducing its temperature could improve stability, and you can try adjusting the fan curve more aggressively if needed.
It's unclear what's going wrong. I tried the new PSU with a 3060 and everything worked fine. Maybe the issue lies with the 3080 controller, leading to crashes during voltage spikes. Locking the max voltage to 875mV still caused a crash once when under load in a game. It seems stable at 850mV, but performance isn't much better than the OC 3060. Temperatures stayed below 78°C on the chip when at 875mV.
It could be a driver problem or a conflict somewhere. Or the GPU might be defective. You might try installing another Windows version and see if it works on a clean system. If the issue persists, it likely points to a GPU problem. Sometimes Windows resists adding a new GPU for some reason, even if it's the same brand.