Diagnosing a PC problem can be challenging.
Diagnosing a PC problem can be challenging.
I have a setup that suddenly stopped freezing after a POST. It uses an MSI X570 Gaming Edge WiFi board with a Ryzen 5800X processor. At first, I assumed the CPU was dead because the CPU debug light appeared on the motherboard. Replacing it with an older 3600 resolved the issue, but then the 5800X worked properly on another machine. After that, I replaced it back and it booted normally.
I tested a full set of benchmarks (like cinebench, 3dmark, heaven) and everything performed well afterward. The following day, the same freezing issue came back this time, but this time the DRAM debug light was active. I had two sets of Corsair Vengeance Pro 16GB RAM and tried swapping them in different configurations. Each stick worked independently, leading me to suspect a faulty slot. I moved the RAM to slots 1 and 3 again, but the problem returned the next day with the memory in channel B.
Now the CPU debug light is back on. At this stage, I’m considering replacing the X570, though I’ve managed to get the BIOS updated successfully during these intermittent restarts and can still run NVME integrity tests. However, I’m unable to consistently reproduce or pinpoint the issue. I’ve changed all cables and removed the UPS, but it’s hard to determine what’s causing the problem reliably.
Any advice would help me avoid wasting money on a new board that might not fix the matter?
Did XMP get activated to 4000 before these symptoms appeared? I'm checking if memory speed is causing issues with the CPU's memory controller. Running RAM at 3600 should stabilize performance. See the Quora thread for more details.
As mentioned before, I have already updated the BIOS and also attempted a CMOS reset. I’ve mostly tested swapping individual BIOS parameters and performed a complete factory reset. Previously, XMP was enabled, but I’m very unlikely to experience system issues. Although it was among the settings I considered turning off, I ran several years with a CPU and GPU that were tuned for stability, resulting in no crashes or shutdowns over the past year. This machine has been used for daily tasks, production work, and occasional gaming, so it had some wear—but only from parts that are only a few years old.
I also wanted to try using a different power outlet and connected a separate system to see if it would run properly. Both seemed to operate normally, which didn’t give clear results. I value the advice on testing options but also want input on possible upgrades. I’m stuck because I haven’t decided yet to switch to AM5, and Intel is facing tough times right now. I’d prefer to keep this rig functional for another year or so before making a major change, especially since mobile options have become more expensive and some CPUs like the 5800X3D are becoming harder to find. Thank you for your feedback so far.
You likely managed to keep the PC functioning well for some time using the settings previously suggested, especially the RAM speed adjustments. However, excessive strain might cause a slight misalignment in the memory controller. The problems you're experiencing could appear similarly to what's being observed. As mentioned earlier, 3600 represents the optimal performance range for your Ryzen processor. Exceeding this often leads to failure or adverse effects linked to the Infinity Fabric. If you run the RAM at its standard specifications without XMP activated, do you encounter the same difficulties? Consider resetting the CMOS and BIOS to their default settings first. What BIOS version are you currently using?
The current version is 7C37v1P (released on August 9th), which I modified during this process. I've updated the BIOS twice before, though I'm unsure which version I was using when the issue started. It's clear the update was completed prior to my 3600 swap, so by the time I reinstalled the 5800x and executed R23, it was already current. My testing lasted roughly two hours before shutdown, followed by the problem reappearing about 8 to 10 hours later. There were no overclocks at any point, and XMP was disabled throughout.
Unfortunately, the situation isn't what I expected. It's now a NAS and was given to a friend. I borrowed it for testing before, but I can't get immediate access anytime. During my initial RAM swapping, I tested those sticks, yet when I saw mine working in the other system after swapping back, they functioned again as mentioned in my previous post. At that stage, I only tested two of my RAM slots with other memory. Still, each attempt resulted in a complete boot. It almost feels like whenever hardware changes, it suddenly works again, but once it's off, the issues return.
New update, after discussing this, I decided to try turning it on without making any changes. The debug lights don't activate at all anymore, and it still fails to progress past the splash screen. As a reminder, considering how often it boots up again after hardware changes, it might not be worth pinpointing my problem if I make an adjustment that gets it running and then run some tests. As before, I performed benchmarks, performance checks, thermals, fan curves—everything looked normal at first glance. Unfortunately, I don't have enough knowledge to interpret the logs while it's running.
Really, no. I reset it after clearing CMOS, but I don't have a battery tester or any other tool that comes to mind for CR-2025 (2032? whatever the size). I'll attempt to install a new one, thank you for the warning.