I'm in a need of some desperate assistance as I have exhausted all possibilities known to me. My PC randomly crashes, usually several times per day (on average every few hours). There are no noticeable patterns, it happens during idle, browsing, gaming, anything. Lately it seems that it sometimes generates a dump in Windows' LiveKernelReports folder. Analyzing the dump reports a failure bucket id: LKD_0x141_IMAGE_amdkmdag.sys with a failure hash 48b738dd-5a92-7ff8-63d0-f075fc680fe0.
PC specs are the following:
After various troubleshooting, the prime suspect appears to be something in this configuration combined with the monitor (Asus MG 278). The reason for this suspicion is that every single piece of hardware has either been replaced over the span of last few months (for as long as the PC has been crashing) or tested in a different PC (where it worked flawlessly for days). As long as the specific monitor was plugged in, the PC would eventually crash within 24h (various components have been swapped 1 by 1 with components from the secondary PC to try to determine if any individual piece of hardware is the culprit). However, when that monitor is plugged into another computer, everything worked fine (on both the configuration that causes problems and the "test" one that validated if the hardware is at fault). It would seem like there's some connection between what the monitor is doing and the AMD components (or their drivers), but the strange thing is that the other test configuration also has AMD components (Ryzen 2700X, RX580, Adrenalin Driver Version: 21.4.1, Windows 10 Home) and there are no issues whatsoever. Before the Ryzen 3800X, the configuration in question had an Intel processor and it was causing issues back then too (it still had Vega 56 inside so AMD drivers were installed there as well). For what's it worth, the problematic configuration at this moment has 21.5.1 drivers, while the test one has 21.4.1.
Other than replacing every piece of hardware, various cables have been replaced (DisplayPort/HDMI for the monitor, SATA cables for HDDs, whole PSU and its cables, PC casing and its cables). Various AMD drivers version were tested, none of them ever fixed the issue. Windows were also freshly installed few times, without any improvement to the issue. All drivers, chipset, bios, Windows are up to date. The PC usually has dual monitor setup, but both scenarios (dual and single screen) were tested and as long as the specific monitor is plugged, it always caused problems. Maybe relevant, but this is the only 1440p monitor on either of the 2 computers mentioned, the rest are all 1080p with one of them being 144hz as well. Temperatures of both configurations are superb, so that was never an issue. Since CPUs were swapped few times, fresh thermal paste was also reapplied. Various benchmarks were ran, memory test was ran, repair tools like DISM, sfcscan, anything that's easily found online as a possible solution has probably been tried by now.
Any help or hint is greatly appreciate because I am slowly losing my mind as it has been going on for months now.
I noticed there were some overclock settings automatically on, when i updated bios. Double check that all of those are disabled!
My pc crashes alot when there's some sort of pbo and docp on simultaniously. Asus B550 and 5600x.
Also do not use docp memory profile. If those help, it's something to do with not very good bios. You could also try to use older stable bios version.