Bad RAM can cause CPU issues like instability, crashes, or slow performance.
Bad RAM can cause CPU issues like instability, crashes, or slow performance.
System details: Fedora KDE MSI B550 Gaming Plus equipped with the newest BIOS, Ryzen 9 5950X, 64 GB Mushkin Redline CL18, Radeon RX 6750XT 10GbE card, PCIE 1X to SATA adapter (providing two additional SATA ports). Crucial NMVE 4TB, two 12TB WD Red Pro spinning drives in RAID1 configuration prior to encountering issues: PBO enabled, core settings adjusted, XMP profile set to 1. The situation began after acquiring a 5800X3D for my gaming rig. Since then, the machine has operated smoothly on Kubuntu, Debian, Nobara, and Fedora KDE, with minor problems only. Below are the events:
I observed sporadic hardlocks. The system would become unresponsive—no input devices, complete shutdown. I had to restart by disconnecting and reconnecting or using the reset function. Lockups occurred roughly monthly.
When Firefox was open on one display while another window appeared elsewhere, the machine froze in a few minutes. Initially, keeping Firefox open helped for months.
After switching my gaming PC to Linux, I removed the 4TB drive and connected it to the Fedora system after powering both down. I transferred approximately 2.9TB of video recordings (mostly off-road footage for YouTube) from the old drive to a 12TB RAID1 partition. Half transferred successfully; the rest locked up completely.
I reset the system, attempted again, but continued locking. Eventually, 2.5TB was copied before total failure. The machine would hardlock immediately after reaching the desktop and remained unresponsive for several more tries.
During troubleshooting, I disconnected the SATA adapter, all extra drives, leaving only CPU, GPU, and NVME connected. It displayed a blank screen with garbled characters. Removing the B2 RAM stick resolved the issue—system booted normally.
After resetting, I tried again, but it hardlocked after a few minutes. Eventually, it wouldn’t even start up. I replaced the drive with a spare DDR4 3600 CL22 module and powered it up. It ran for about ten minutes before locking again.
I reset the BIOS, disabled XMP and PBO, performed MEMTEST for hours, rebooted, and successfully logged into KDE. A stress test on the CPU lasted an hour without issues. I then re-enabled XMP and PBO. The system eventually hardlocked after a few minutes.
I disabled XMP and PBO, ran MEMTEST, rebooted, and the system remained stable for 24 hours. I left the timings at JDEC defaults and increased RAM from 3000 to 3600. After MEMTEST, I re-enabled XMP without issues.
However, during a stress test, the system locked up briefly while running MEMTEST. I ordered a replacement memory kit but opted for a 32GB dual-channel 3600 CL16 instead due to cost concerns. I’m concerned about potential damage to the 5950X.
What checks should I perform before declaring the CPU faulty? I confirmed memory issues, as another AM4 system with RAM also failed to boot.
Additional advice:
- Verify BIOS settings and ensure no conflicting profiles are active.
- Check for overheating risks—consider using ECO mode during peak loads.
- Monitor temperatures closely; prolonged high temps can accelerate wear.
- If possible, run diagnostics with a known-good system to confirm hardware health.
- Keep spare parts on hand for quick replacements.
Please let me know if you need further guidance or specific test procedures.
Initially, RAM problems shouldn't really harm the CPU. The high room temperature may actually cause more damage. Regarding the RAM, it could be due to a faulty slot on the board or the CPU being over-tightened. Have you looked for any bent pins or re-seated the CPU? Another suggestion is trying to disable fast boot, MCR, and DRAM power down. But after doing this, your PC will perform memory training every time it boots. The DRAM power down merely halts the RAM from entering low-power mode.
AM4 isn't an LGA design; the pins are on the processor and fit into a slot on the motherboard. I'm concerned about potential damage from a shorted RAM, which could affect the CPU. The CPU has been in good condition—running Linux for a while—and I haven't worked on any video projects recently. I use an NHD15S cooler, keeping the CPU temperature stable. The RAM is completely faulty; this stick would stop any system from starting. Although it's no longer present, it remains unstable when PBO is enabled. This raises my concern about possible CPU damage.
It could simply be the CPU nearing end of life if RAM fails, making it impossible to boot Windows. Don’t jump to conclusions unless you know the technical details inside out. If you suspect a random failure, consider checking the warranty and adjusting the voltage slightly lower. Save up for an upgrade—used units aren’t ideal. The 3600C16 is overpriced; 3600C18 might be better value. Some GSKILL kits offer high-performance options at lower costs, especially if you need the 16Gbit revision for ballistix 4000/4400 chips. Otherwise, you risk wasting money on a less-than-optimal RAM package.
Tldr: no? I don't really understand why you think the CPU is faulty, any CPU benchmarks? Buy Samsung bdie or at least gskill (preferably gskill bdies that's what I did, I bought 3200 cause I knew they'd OC easily, and they do! ) Or those if you can! Well, I mean pbo is basically an "overclock" meaning not at spec... tell you what I can run pbo -30 and no tests fail whatsoever, even games run fine, except overwatch... even pbo -10 it'll crash ... (mind you not the PC, just the game...) So I'm just not running pbo, not an issue I have good cooling! Never have I thought cpu is faulty... it's running above specs even at default in cinebench, no crashes ever whatsoever (without "pbo")
Uncertain about your reasoning. A faulty RAM might stop the final submission.
The issue appears to be causing corruption before the machine attempts to write to bad RAM, and this happens across all systems I've tested. I chose g.skill with a CL16 3600 processor, and since every PC I own runs Linux, I'm mainly concerned about stability rather than overclocking. DOCP/XMP remains functional at its current settings for years. I intend to adjust PBO later if needed to improve stability. I've experienced faulty CPUs before, especially those with AM4 sockets, but not the 3900XT. If PBO fails altogether, I plan to replace it with a 3900XT and upgrade to a 5950X when finances allow. The POST process finishes without issues, but the screen output becomes corrupted before any successful start.
The system has recovered after the update, thanks to the new RAM. I ran a memory test afterward.