GPU and SSD issues leading to Windows freezes—might the motherboard or CPU be the culprit?
GPU and SSD issues leading to Windows freezes—might the motherboard or CPU be the culprit?
Windows 10 would become unresponsive during: gameplay after quitting, read/write tasks on secondary M.2 SSD. System freezes usually occur until shutdown or restart. Occasionally, logs appear such as: "My specs are..." followed by detailed hardware info. The motherboard is an AMD 5600X CPU model, and the BIOS version is 3607 (updated April 2024). It features two M.2 slots (PCIe 4 x4 and PCIe 3 x4) plus two PCIe x16 slots (PCIe 4 x16 and PCIe 3 x4). The second slots for both M.2 and x16 share bandwidth with unused slots. This board is no longer under warranty. Troubleshooting includes installing RTX 3080 in the second x16 slot, emptying the first x16 slot, swapping RTX 3080 for MSI GTX 680 Twin FROZR III, inserting a 512GB Kingston OEM NVMe SSD in the same position as the 2TB drive, and using a Samsung NVME storage drive. Firmware updates, clean installs of Steam, NVIDIA, and Hulu were performed. SSDs ran at 98%+ usage, no warnings detected, and a Secure Erase was completed. SSDs were reflashed by Samsung after being tested. The issue appeared to resolve when using a different SSD and running specific drivers.
Maybe thanks, though I assumed the BIOS battery only held BIOS settings and a clock—nothing seemed wrong with that. I doubt PCIe timings would be small enough to rely on the battery either. Also, I’m puzzled as to why it wouldn’t impact all my M.2 SSDs or GPUs. The problem appears only when my RTX 3080 uses the PCIe 4 x16 slot (it functions in a PCIe 3 x16 slot at PCIe 3 x4), and my Samsung SSD in the second M.2 slot (PCIe 3 x4). ChatGPT can’t confirm the idea either.
I experienced more freezes on my PC with the RTX 3080. I replaced my old CR2032 with another from my drawer, but the issue continued. It took some time to locate my multimeter; otherwise I’d have responded earlier. The original battery was showing 3.17V, which is higher than the expected 3V and the ASUS requirement of 2.9V, while the new battery from the drawer was below 1V. This suggests my motherboard’s original CMOS cell has been fine, and I discarded the spare one. I plan further testing, but if freezing remains even with the PCIe 3 x16 @x4 setup, it likely points to a motherboard problem unless another explanation exists.
Have you attempted to reinstall the CPU, inspected for bent pins, verified RAM installation? Consider running Memtest86+ or a Passmark Memtest86 test? What information does Hard Disk Sentinel provide regarding your storage devices? Does this issue persist with Windows 11? Have you tried an older BIOS version? Updated chipset drivers? Monitor SSD temperatures during operation? Ensure proper SSD cooling.
They changed it a while back, and the problem only appeared recently. It seems I have limited options if I want to avoid bent pins during installation over the last couple of years... Maybe I should give it another shot. I haven’t seen any DRAM-related error lights or event logs about this issue. The regular setup didn’t boot from a USB, but I haven’t tried the passmark version yet. I’ll check that out. I’m still seeing occasional updates that cause problems, so I might take a picture of my boot drive. What details does the CDI system provide that the Event Log doesn’t? When I put a non-boot drive in an M.2 case, I no longer get error messages. So I think I’ve eliminated that one. To summarize my past SSD troubleshooting: I definitely upgraded the BIOS—don’t recall if I also went back to an older version. I could double-check that again... SSDs usually have aluminum heat sinks from the motherboard or case. It’s likely their temperatures stay between 50-60°C under load.
I've repositioned it and now only has two (2x8GB) instead of four. Despite this, occasional freezes still occur, and the thumb drive disconnected during my attempts to transfer files for passmark's bootable Memtest86. Once it started, it completed 75% with no errors. I'm planning to let it finish but suspect the issue might lie with the motherboard itself. I remember not placing as many standoffs underneath, possibly missing a back/pressure plate on the CPU. There might have been a small screwdriver dropped recently, though no visible damage was found. Many circuits are hidden and not immediately obvious, even if I don't notice any sudden system problems afterward. Before or after installing these components, my onboard NIC disappeared from the hardware list, and even trying to install ASUS' driver didn't resolve it. Around Thanksgiving I might switch to Windows 11 and, if needed, replace the motherboard since there could be deeper problems than just a single part.