The PC freezes now frequently instead of rarely.
The PC freezes now frequently instead of rarely.
Hello! I have quite the problem and looking for help from some hardware experts. All signs point to harware but having some perplexing stuff going on.
HARDWARE
Built Late 2021
CPU, AMD Ryzen 9 5900X
RAM, G.SKILL Trident Z Neo DDR4 3600
MOBO, GIGABYTE B550 VISION D-P AM4 AMD B550 ATX
M2, Western Digital WD BLACK SN750 NVMe M.2 2280 1TB (
Windows
)
Western Digital Blue 2TB SSD (
Linux
)
Graphics, GIGABYTE Eagle GeForce RTX 3080 Ti 12GB
PSU, Corsair HX750 model rps0074
CPU Fan, Noctua NH-U12S SE-AM4
PROBLEM
Constant freezes (seems to happen quicker using linux -- max 10~15 minutes of use), have to hard boot machine rarely the computer will just restart on its own. This even happens while using a Kubuntu
Live CD
, I've tried reinstalling linux thinking it could be OS related. OS install always fails around 50%.
HISTORY
I use two separate hard drives for linux & windows, I do not dual boot.
Linux started freezing on me about 3 months ago, but not very often and not while in use. Maybe once every 5 days that would require a hard boot. However, my children use windows to play games and never ran into freezing issues.
Hard drives
The freezes began happening more regularly and started happening while in use, it is now to the point where the freezes happen almost instantly. At first this only happened in
linux
but happened a few times while in windows while I was looking at logs (windows seems to go longer without freezing??). I figured it was a hard drive issue so I swapped my linux HD to another machine and ran SMART tools & bad. block scan (took 2 hours) and everything was fine; the fact it didn't freeze within the first 10 minutes was eye opening. I did find a failing sandisk SSD spitting out bad sectors for the last 6 months, I was able to back up important files and that SSD has now been retired.
RAM
(I re-seated this) HD seemed clear so I ran memtest on my ram for approximately 3 hours with no errors. I've also done the Windows Memory Diagnostics Tool 3x, no issues.
Drivers/Bios
My bios version was pretty outdated so I updated from F11 -> F18.5
My AMD chipset drivers were out dated so I updated from 3.x to 8.x (I think 8, maybe 5?)
GPU
I've ran furmark twice for approximately 15 minutes each time and everything looked fine.
CPU
I ran a bunch of stress-ng tests while in linux and everything looked fine (linux didn't freeze for awhile so I locked the screen and went upstairs and it was frozen 15 minutes later -- no load).
ERRORS I'VE SEEN:
Forgive me, I do not know the order in which this all took place. But here are some error codes I've seen the past few days trouble shooting after this:
After a freeze that restarted the computer, only seen this once
[0.814896] mce: [Hardware Error]: CPU 0: Machine Check: 0 Bank 27: baa000000000080b
[0.814905] mce: [Hardware Error]: TSC 0 MISC d0120001000000000 SYND 5d000000 IPID 1002e00000500
[0.814915] mce: [Hardware Error]: PROCESSOR 2:a20f10 TIME 1738635997 SOCKET 0 APIC 0 microcode a20102b
Went down this rabbit hole, nvidia proprietary drivers vs not, don't think issue.
[drm:nv_drm_master_set [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00005200] Failed to grab modeset ownership
NEXT Steps & Random thoughts
This is my work PC and having it out of comission is putting me behind. Does anyone have any ideas I test next? What the next likely culrpit might be?
My PSU fans spin, my GPU fans spin, my CPU fan spins.
I took out CMOS battery for awhile while inspecting machine.
I've reapplied thermal paste on CPU (previous job was kind of shoddy, whoops).
I can run intense stress tests while on windows and everything seems okay. I've yet to be doing anything intense and have the system freeze, seems to always happen during low usage (even on linux).
Linux consistently freezes after a max usage of 15 or so minutes.
Kubuntu Live CD freezes near 50% on install (I've tried 6x).
I'm having a hard time getting windows to freeze again.
Any ideas?
Thank you!
EDIT:
I've never overclocked. Hardly ever messed w/ bios settings. Most I've ever done in there is switch my RAM profile.
MISC:
Here is my: inxi -Fxz
System:
Kernel: 6.8.0-52-generic arch: x86_64 bits: 64 compiler: gcc v: 13.3.0
Desktop: Budgie v: 10.9.1 Distro: Budgie 24.04.1 LTS (Noble Numbat)
base: Ubuntu
Machine:
Type: Desktop System: Gigabyte product: B550 VISION D-P v: -CF
serial: <superuser required>
Mobo: Gigabyte model: B550 VISION D-P serial: <superuser required>
UEFI: American Megatrends LLC. v: F18d date: 09/02/2024
CPU:
Info: 12-core model: AMD Ryzen 9 5900X bits: 64 type: MT MCP arch: Zen 3+
rev: 0 cache: L1: 768 KiB L2: 6 MiB L3: 64 MiB
Speed (MHz): avg: 2962 high: 4584 min/max: 2200/4950 boost: enabled cores:
1: 2873 2: 3383 3: 2200 4: 2904 5: 4584 6: 3466 7: 2200 8: 2200 9: 2200
10: 2864 11: 4008 12: 3592 13: 3445 14: 2200 15: 2200 16: 2200 17: 3667
18: 2200 19: 2200 20: 3339 21: 3589 22: 2200 23: 3360 24: 4034
bogomips: 177254
Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3
Graphics:
Device-1: NVIDIA GA102 [GeForce RTX 3080 Ti] vendor: Gigabyte driver: nvidia
v: 550.120 arch: Ampere bus-ID: 52:00.0
Display: x11 server: X.Org v: 21.1.11 driver: X: loaded: nouveau
unloaded: fbdev,modesetting,vesa failed: nvidia gpu: nvidia,nvidia-nvswitch
resolution: 1: 3840x2160~60Hz 2: 2560x1440
API: EGL v: 1.5 drivers: nvidia,swrast platforms:
active: x11,surfaceless,device inactive: gbm,wayland,device-1
API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: nvidia mesa v: 550.120
glx-v: 1.4 direct-render: yes renderer: NVIDIA GeForce RTX 3080 Ti/PCIe/SSE2
Audio:
Device-1: NVIDIA GA102 High Definition Audio vendor: Gigabyte
driver: snd_hda_intel v: kernel bus-ID: 52:00.1
Device-2: AMD Starship/Matisse HD Audio vendor: Gigabyte
driver: snd_hda_intel v: kernel bus-ID: 54:00.4
API: ALSA v: k6.8.0-52-generic status: kernel-api
Server-1: PipeWire v: 1.0.5 status: active
Network:
Device-1: Realtek RTL8125 2.5GbE vendor: Gigabyte driver: r8169 v: kernel
port: f000 bus-ID: 4e:00.0
IF: enp78s0 state: up speed: 1000 Mbps duplex: full mac: <filter>
Device-2: Intel Wi-Fi 6 AX200 driver: iwlwifi v: kernel bus-ID: 4f:00.0
IF: wlp79s0 state: down mac: <filter>
Device-3: Intel Ethernet I225-V vendor: Gigabyte driver: igc v: kernel
port: N/A bus-ID: 51:00.0
IF: enp81s0 state: down mac: <filter>
IF-ID-1: docker0 state: down mac: <filter>
Bluetooth:
Device-1: Intel AX200 Bluetooth driver: btusb v: 0.8 type: USB
bus-ID: 1-6.1:6
Report: hciconfig ID: hci0 rfk-id: 0 state: up address: <filter> bt-v: 5.2
lmp-v: 11
Drives:
Local Storage: total: 8.19 TiB used: 370.73 GiB (4.4%)
ID-1: /dev/nvme0n1 vendor: Western Digital model: WDS100T3X0C-00SJG0
size: 931.51 GiB temp: 53.9 C
ID-2: /dev/sda vendor: Western Digital model: WD Blue SA510 2.5 2TB
size: 1.82 TiB
ID-3: /dev/sdb vendor: Western Digital model: WD6003FZBX-00K5WB0
size: 5.46 TiB
Partition:
ID-1: / size: 78.19 GiB used: 24.74 GiB (31.6%) fs: ext4 dev: /dev/sda3
ID-2: /boot/efi size: 1.05 GiB used: 6.1 MiB (0.6%) fs: vfat
dev: /dev/sda1
ID-3: /home size: 1.56 TiB used: 334.12 GiB (20.9%) fs: ext4
dev: /dev/sda5
ID-4: /opt size: 124.93 GiB used: 11.86 GiB (9.5%) fs: ext4 dev: /dev/sda4
Swap:
ID-1: swap-1 type: partition size: 32 GiB used: 0 KiB (0.0%) dev: /dev/sda2
Sensors:
System Temperatures: cpu: 47.6 C mobo: N/A gpu: nvidia temp: 54 C
Fan Speeds (rpm): N/A gpu: nvidia fan: 0%
Info:
Memory: total: 32 GiB available: 31.24 GiB used: 4.14 GiB (13.2%)
Processes: 461 Uptime: 13m Init: systemd target: graphical (5)
Packages: 2457 Compilers: gcc: 13.3.0 Shell: Zsh v: 5.9 inxi: 3.3.34
I'm lucky enough to have gathered a set of PCs over the years, which means I won't have to deal with any single unit being out of action.
This also allows me to experiment freely by swapping parts and trying to pinpoint the issue, even if it's not immediately clear what's causing the problem.
The most challenging part to replace is the motherboard, given its significant cost.
If the device is high-end, the CPU comes next—especially if a high-quality one is available. It's simpler to swap out RAM, GPU, or SSD components. Even a basic GPU can serve as a useful test.
I'd look for the most affordable compatible CPU, possibly second-hand on platforms like eBay, and use it in place of the 5900X.
You've already verified your RAM with MemTest (86 or 86+?), and it passed. However, if you're pushing it to overclock with XMP/EXPO settings near 3600MT/s, consider lowering it back to JEDEC 2133 or 2300MT/s (XMP disabled) for better stability. Sometimes multiple MemTest runs are necessary to detect hidden issues.
The final step is swapping the motherboard. I recommend searching eBay for the most affordable microATX board with a proven working history, checking seller ratings. I've purchased numerous used boards without major problems. You also have eBay's warranty if the listing is misleading. My most cost-effective setup (AMD FM2) cost me $7 and is now in a server.
A used board/CPU combo is likely the best option. This way, you can keep a spare CPU for testing your B550 Vision motherboard. Assuming the eBay deal is valid, you could try running your 5900X on the second board if its BIOS supports it.
Considering it's your primary work machine and only a few years old, if you're facing major losses or client complaints, now could be a good time to upgrade—either with a new CPU, a DDR5 motherboard, or an upgraded RAM setup. If neither makes sense financially, a used office PC might be the safest choice while you wait.
We're assuming your HX750 PSU is still functional, but the Corsair website indicates it's unavailable in the US, EU, and UK. Did you purchase the HX750 new in 2021, or did you upgrade from an older model? If it's over five years old, it might be worth questioning its reliability.
Power supplies are another component that eventually fails, even with warranties. You may need to replace one for a short period, borrowing a spare from a friend or local repair shop if possible.
In summary, swap parts gradually until the problem disappears. Good luck.
Thank for your well thought out reply!
I do have a few computers around here but unfortunately all Intel.
RAM
I used memtest86+, I am not over clocking it -- I'm so inept in that I don't even know what any of those settings mean, ha!
"My cheapest working mobo/CPU combo (AMD FM2) cost me $7. It's now in a server." Wow, this sounds kind of appealing as I've wanted to get a reliable always on postgres server locally.
"We're assuming here that your HX750 PSU is still OK, but the Corsair.com web site shows it (or a similar model) is out of stock in the US?, EU and UK. Did you buy the HX750 brand new in 2021, or bring it over from an older build? If it's more than 5 years old, treat it as suspect."
(
US
)
Good call, I finally remembered to search it up in my email and I RMA'ed a defective one back in July 2018. So, I've been using this replacement since.
"Power supplies don't last forever, even if they have a 7 or 10-year warranty. It's yet another component you need to swap out for a few hours, if you can "borrow" a spare PSU from a friend or local repair shop."
My wife has a janky 550watt something I got off amazon... I can throw in and do light tasks like programming (day job) and see what happens.
Just so perplexing. I've had windows up for 1hr30 minutes so far, streaming you tube, twitch, music and randomly running GPU benchmarks. No issues.
I wonder if there is something in the linux kernel that isn't playing nice with my PSU? That would explain the quick freezes in linux and USB Live CD...
My next steps are to leave windows running w/ a bunch of tasks overnight and see if it ever freezes.
If no, I'll try using linux (daily driver) again and see if I have freezes.
If yes, I'll swap out the PSU with my wife's PSU and report back.
If I still have issues w/ swapped PSU I'll start looking at mobo/cpu combo replacements, ugh!
3600 on memory is treated as an overclock even with XMP enabled. Reduce it to 3200 for RAM speed, maintaining XMP but lowering it back to 3200 from 3600. Anything above 3200 is considered an overclock. Your CPU cooler isn't sufficient for the Ryzen 5900x; I'm using a 7 heat pipe cooler with 3 fans active. Set a BIOS temperature target at 85°C, with a maximum of 90°C. Lowering the clock when it reaches 85°C will help extend the CPU's lifespan.
Gotcha. I turned XMP off and set it back to 3200. I managed to install Kbuntu, which gave me some hope. But after about ten minutes of the fresh install, I'm freezing again.
Also, your CPU cooler isn't sufficient for the Ryzen 5900x. Thanks, I'll check that out.
"Also, set a temperature target in BIOS to 85°C."
I searched for this in BIOS but couldn't find it. Not sure if my BIOS supports that setting.
I came across this link and thought it might be relevant. It discusses system freezes with a Ryzen CPU on Linux, noting that the user experienced issues after a year of use. They mention trying different kernel versions and found success with "Power Supply Idle Control" settings. The post also shares ongoing stress tests to check for freezes.
A minor adjustment in the kernel might lead to significant issues, yet it's interesting that you noted Windows also has unstable behavior, and now you bring up this older (legacy) setting for PSUs in the BIOS... Could this be connected to the PSU perhaps? Maybe I overlooked the information—what's the age of your PSU?
Yeah, everything seems a bit unclear due to stress (working on the computer and meeting deadlines, haha). Let me try to remember the Windows freezes... I think I experienced two while doing light tasks in Windows, probably just checking system events for hardware issues. It’s possible Windows is more demanding, so it might not have triggered a 'Low' power state often.
I own my PSU since July 2018 and had to return the previous one through RMA. I found the original support ticket with Corsair.
"Early this year (2018) on January 13th, I bought a CORSAIR HX Series HX750 powersupply from newegg.com. Three days ago, my computer wouldn’t boot fully. After some troubleshooting with a friend and swapping the power supply for one in his machine, the problem was resolved. I’m worried the power supply might have failed after about five months of use."
So I’ve been using the same PSU since July 2018.
mmm... roughly six and a half years old plus the last update when they provided a new one since the original wasn't functioning properly.
Probably trying a different PSU would help, and make sure to revert the PSU setting back to the default in the BIOS.
By my understanding, each time I perform a BIOS update I always save the default/optimized configuration once completed. Then immediately return to the BIOS and adjust XMP, FTP, FPTm, PBO, and any other parameters before starting Windows.
I own a M.2 that functions as an external hard-drive.
When I link it through USB and it becomes warm, the entire system freezes. The computer thaws when I disconnect the external device.
Thus, any unstable drive linked to the machine, even if idle, can cause freezing.