Frequent crashes happen while playing (I can resolve them, but they keep occurring).
Frequent crashes happen while playing (I can resolve them, but they keep occurring).
Hey everyone, I'm reaching out for advice on a persistent issue with my PC that's been bothering me for months. From what the title suggests, during gaming my system sometimes freezes and stops responding, though the audio usually stays active. After a few seconds it crashes and restarts without any BSOD. This happens consistently whenever I launch a game. I've tried resetting the system, using DISM, running sfc, and checking DDU, but nothing seems to resolve the problem except for one temporary fix—completely unplugging the GPU and holding down the power button to discharge the capacitors. Once I reconnect it, the crashes stop, though within a week or less the issue returns again.
The crashes seem tied to GPU activity or asset loading, especially when my character moves around or interacts with menu items. The event viewer shows only two suspicious ACPI entries: one labeled "critical" related to kernel power creation at the crash moment. Temperatures look normal, so a hardware thermal problem doesn't seem likely.
My specs are: CPU i9-9900K, GPU AMD RX 6800 XT (mounted on a riser for over a year), motherboard BIOS updated, ASUS ROG Strix z370-e gaming PC, Corsair RM850x PSU, 32 GB T-force RAM, Samsung 970 EVO 500 GB SSD, and TEAMGROUP T-force Cardea Zero Z440 2 TB SSD. I'm using Windows 11 Pro.
Anyone else notice this pattern? Does anyone have similar experiences or suggestions?
Ensure PCIE_1 slot operates at version 3.0 and set bifurcation parameters to 16x in the UEFI.
The PCIe edition is already configured at version 3.0 (z370 cannot handle 4.0) and this board lacks bifurcation support except for the ASUS M.2 slot card.
If you don't manually adjust 3.0, it's handled automatically. Sometimes PCIe will settle for smaller versions to conserve energy, but the GPU may have trouble managing them or reverting correctly, leading to crashes. I've noticed this problem multiple times—forcing a specific version can resolve it in certain cases. You might also want to turn off power-saving options to see if that helps.
I'll verify the PCIE version later. What confuses me is whether this configuration or something else can account for my PC's behavior after unplugging and reinstalling the GPU. To temporarily resolve the issue, I usually disconnect both PCIe power cords from the card and the card from the riser (or sometimes the riser from the motherboard), but it doesn't always work.
The PCIe version needs constant checking. I've observed this pattern with an older serial PCIe card in a custom setup, where only removing and reinserting it works (not applicable for systems with VGA aux power like big GPUs). To explore further, it might relate to a negotiation problem: IEEE Xplore Full-Text PDF suggests UEFI settings could disable power-saving modes. There could be a mix of PCIe negotiation and UEFI that forces version 3.0 or mishandling L1.1/L1.2 states. It's also conceivable the hardware itself is faulty, which might be resolved by adjusting these settings. If you lack a test system, reducing variables is your only option here.
Here's an update on the situation, unfortunately manually setting the pcie interface at 3.0 speed didn't solve the issue. While the interval between the state of "working fine" and "pc crashing" was longer this time (about 3/4 weeks) my pc started crashing again