Driver missing from AMD GPU and frequent system failures despite fixing RAM issues and reinstalling OS
Driver missing from AMD GPU and frequent system failures despite fixing RAM issues and reinstalling OS
Hey there! As the guide mentions, the GPU driver seems to vanish each time I restart. Even though it was properly installed on a brand new Windows system and functioned well initially, it disappears after a reboot. The control panel still shows up in Windows Search, but launching it brings me right back to the same error.
Here’s what happened: I acquired some used RAM from a trusted seller. After installing it, it didn’t work at all with XMP 3200. I lowered it to 2933, which was the other profile, but still couldn’t boot. I kept adjusting the frequency until it stabilized around 2800/2666.
Over the past month, I experienced several random BSODs related to memory issues like MEMORY_MANAGEMENT and IRQL_NOT_LESS_OR_EQUAL. To try fixing things, I changed things a bit: increased the SoC voltage to 1.1V, set ProcODT to 60 Ohm, adjusted DRAM frequency to 2400 MHz with a higher voltage, swapped memory sticks after realizing I had mixed them up (the slots weren’t what I thought), and even upgraded my BIOS.
Eventually, I managed to boot at 2933 MHz (XMP) but the system remained unstable. Yesterday it crashed completely, making Windows fail to start. I had to reinstall it. After that, I upgraded my BIOS to the latest version, which worked fine initially. I ran `sfc /scannow`, removed the new sticks, and kept the old ones, but the problems persisted with the same frequency issues.
I’ve installed several memory kits over the years—some worked, others didn’t. My HDD shows 0% health in HD Sentinel, though it seems to function normally without any problems. The GPU itself isn’t failing, and there are no random errors.
I did note that my old Windows version (2004) had no driver issues, unlike the current one. The new RAM is the same type but uses different memory chips (Hynix vs Kingston). My specs include: a Chieftec 750W power supply, Ryzen 5 1600AF, Tomahawk MAX BIOS updated, two 8GB HyperX Predator cards (paired), Sapphire RX 570 GPU, and drives like the WD Green 240GB and Seagate 4TB.
Before installing the new RAM, everything was fine. The only thing that changed was the memory configuration. Thanks for your help!
Have you confirmed the driver was downloaded from the Amdus website?
Windows usually sets up the drivers automatically. Until now I just downloaded the WHQL version and used the newest one too, but the same problems kept happening. I assumed it might be because I hadn't restarted after logging in first, so I did restart then installed the drivers. The same went for DDU—each time I encountered a BSOD with code `0x21`. (After the welcome screen, I couldn’t access Windows at all; it showed a black screen. The computer didn’t freeze, but pressing Ctrl+Alt+Delete still worked, yet TaskMgr wouldn’t open.)
Sorry, I haven't viewed your description yet before replying. Please check my previous comment.
I noticed something new—my air cooler came with a homemade mount (like the old AMD clip design). Could the pressure during installation have harmed the CPU or motherboard? Or are all these problems linked to improper mounting force? Once I installed the stock cooler, I had to flip it around so the clip faced the opposite side. I plan to re-seat the CPU and see if that helps. A few days ago I also mounted the stock cooler, and since my RAM was in good condition, it might mean the CPU got damaged or wasn’t seated correctly.
I resolved the driver problem by following these steps: First, I ran an AIDA stress test which failed, so I increased the SoC voltage to 1.1V and waited 25 minutes before restarting. After that, I removed the NVMe SSD and started installing Windows on a SATA drive. During memory management setup, I took out one RAM stick (now only an 8GB module remains). I restarted several times to check if the GPU driver remained stable. Eventually, it worked.
Next, I updated the GPU driver. Some lingering questions remain: Was the NVMe SSD always the culprit? Did the BSODs begin after installing the new RAM? Could the faulty RAM have also harmed the NVMe drive? I suspect the SSD was the main issue.
I raised the SoC voltage to 1.1V and set the PCIe controller to 60 ohms, but neither adjustment seemed to affect the CPU. It appears the RAM might not be functioning properly unless the SoC voltage is adjusted further—possibly up to 1.35V. I also considered the possibility of physical damage from improper mounting pressure during installation.
I re-installed the cooler after replacing the RAM, which blocked the slots, and doubted that was the cause. The issue seems more related to the SSD. To confirm, I installed all memory sticks and ran a memory test for 8 hours while studying. I also used HD Sentinel on the NVMe drive; sometimes the GPU driver would vanish after running DDU.
It’s possible the CPU or memory module suffered damage from incorrect pressure during installation. Removing the cooler helped, suggesting the problem was with the cooling system rather than the RAM itself.
In summary, my NVMe SSD was likely responsible for the driver issues. Switching to a SATA drive resolved the problem. The faulty RAM probably contributed to the failure, but I’m still unsure why it worked on the SSD after changes.
I resolved the problem by identifying the "known-good" memory as the cause. Interestingly, using only one module from the kit works perfectly, but combining both triggers driver errors and a BSOD 0x21. I experimented with modules in different channels, including A1 B1 instead of the suggested A2 B2, but it didn’t help. Shutting down the PC resolved everything, yet restarting still caused loading issues. It’s quite an odd situation overall.