Random system crashes that seemingly follow even across reinstalls

seth · 2024-09-08 20:10:46

I tried, and it still failed, on auto, The new memory only works on auto

What does that mean? You cannot control the memory clocks/timings or any configuration but "auto" fails immediately?

Run memtest86+ (at least over night, 16+ h - or until you hit errors)
If this reveals memory issues w/ different DIMMs, it's very likely related to the board, I'd then first check for firmware updates.

Other than that:

seth wrote:

Is this all you have or could you save more journal leading up to this?

theRedCyclops · 2024-09-08 21:27:43

That is the output of journalctl -eb -1, last boot being the one that crashed, if there are any additional logs I do not know where they are.
I used memtest86+ and mprime to test the memory when I was seeing some failures related to it, it did reveal memory issues with different DIMMs, but only when using the EXPO profiles, with the auto profile the new memory did not give errors although I didn't test it continuously for that long, I think I did ~6hr, although I'm not really sure.

seth · 2024-09-08 21:37:13

-e jumps to the end, did you directly feed that into 0x0.st or copy it somehow?
Does "sudo journalctl -b -1" not start with some early boot messages?

Manual RAM configuration isn't available (so you can just set it to the lowest values)?

theRedCyclops · 2024-09-08 22:41:21

I piped it into the 0x0 curl command, I just jumped to the end to check the logs manually and reused the command when uploading, I only know how to set the memory settings in the UEFI, What do you propose I do exactly? I'm almost sure the memory is no longer the issue.

Last edited by theRedCyclops (2024-09-08 22:42:38)

seth · 2024-09-09 05:53:24

I only know how to set the memory settings in the UEFI

And the question is what degree of control the board allows you over the RAM clocks and timings - this can range from "not" to fidlling w/ RAM voltages and highly depends on the board, you'll have to consult its manual for that - are you trying to overclock the system in any way?

I piped it into the 0x0 curl command

Please double check whether the journal for that boot does indeed not start in a regular way but cuts srtaight into the middle of a kernel oops.

theRedCyclops · 2024-09-09 08:59:23

The board does let me set everything manually, but I have not tried to. I have not overclocked the system now or ever, nor have I attempted to. The most I have tried, as I said, is using an EXPO profile for the memory.

I think it started in a regular way, the crash occurred several hours after the system started and I don't see anything relating to the crash in the boot stage

theRedCyclops · 2024-09-09 09:02:52

The system also fails to poweroff since the kernel parameters you proposed were added, it reaches the poweroff or reboot target but never actually shuts off

theRedCyclops · 2024-09-09 09:09:22

Ok, it ended up rebooting, but not cleanly, it took several minutes and processes had to be killed by the kernel

cryptearth · 2024-09-09 10:01:26

about the expo profile: can you give us more information?
what's the current jedec speed the ram works with and what's the expo speed it should go up to but doesn't work?

theRedCyclops · 2024-09-09 10:19:47

the memory is currently working at 4800 MHz, the EXPO profiles are for 6400 MHz, but I think my CPU doesn't support them

cryptearth · 2024-09-09 11:23:47

both the cpu and board should support such speeds
but I question the overclock: the 7600x is designed for a base clock of 4.7ghz with a boost up to 5.3ghz
yet your initial post shows a current speed of 5.45ghz
have you enabled something called "a.i. tweaker" or similar? its ASUS branding for automatic overclocking and usually enabled by default
could be the board pushing the thermals higher than what your cooler can handle

Last edited by cryptearth (2024-09-09 11:24:28)

theRedCyclops · 2024-09-09 11:30:58

Everything is currently set to auto, should I manually Disable core performance boost?

Last edited by theRedCyclops (2024-09-09 11:31:07)

andyturfer · 2024-09-09 12:01:16

May not be related, but I have a Ryzen 7 3900x that used to run on an ASUS board (I think it was an X570p). I was getting random crashes. The system would just "freeze", so that not even the mouse cursor was working. These freezes would happen randomly - sometimes once or twice a day, sometimes it wouldn't happen for 3 or more days.

While I never figured out exactly what the issue was, the fix was to increase the CPU voltage (by a very small amount, I think it was something like 0.02V, although I can't remember exactly).

I no longer have the ASUS motherboard, and now have an MSI board (still running the 3900x). I haven't had a single crash since, and haven't had to change any of the BIOS settings.

I've read many posts from others experiencing the same issue with ASUS boards and Ryzen CPUs. See here for example:

https://www.reddit.com/r/buildapc/comme … _voltages/

Might be worth a try (just be careful)...

Last edited by andyturfer (2024-09-09 12:01:38)

theRedCyclops · 2024-09-09 13:49:15

I don't think I'm comfortable changing the voltage of the CPU, I have far too little experience in the field.

seth · 2024-09-09 13:54:29

Does the UEFI provide the curve optimizer?
https://wiki.archlinux.org/title/Ryzen#Random_reboots

the memory is currently working at 4800 MHz, the EXPO profiles are for 6400 MHz, but I think my CPU doesn't support them

And can you set a lower frequency? What about the timings?
In doubt make a photo of the RAM and CPU configs and link them here so we know what your options actually are.

theRedCyclops · 2024-09-09 14:31:02

I will do that, but really, I don't understand why you're so interested in the RAM, I thought it had been discarded?
Shouldn't I test with core performance boost disabled for at least a few days? so far I had no crashes since I disabled it and the system is shuting down correctly

Last edited by theRedCyclops (2024-09-09 14:34:37)

seth · 2024-09-09 14:35:30

The memory is not only the DIMMs it's also the sockets you plug them into, your board is what the information in the RAM travels through.
From all you've provided this very much looks like a memory corruption, so it's either the CPU or the RAM (since the subject indicates it's regarless of the OS)

Driving the latter at lower speeds and higher latencies is the easier thing to check. So we check that first.

theRedCyclops · 2024-09-09 14:57:08

Ok, I have managed to get the timings for the EXPO profiles, but not for the auto profile which is the one I am currently using.
The EXPO profiles are the following:
DOCP I:
DDR5-6400 32-40-40-84-2N-1.40V-1.40V
DOCP II:
DDR5-6400 32-40-40-84-2N-1.40V-1.40V

andyturfer · 2024-09-09 15:13:20

theRedCyclops wrote:

I don't think I'm comfortable changing the voltage of the CPU, I have far too little experience in the field.

Then you could try swapping out the motherboard to rule it out. Unfortunately, this means buying (or having access to) another motherboard. As mentioned in post, two things worked for me:

1. Upping the CPU voltage on my ASUS motherboard.
2. Replacing the ASUS motherboard with an MSI motherboard.

Do a search on the web, there are hundreds of posts about people with random crashes with AMD CPU + ASUS motherboard + Linux, and the "solution" (workaround) is to increase the CPU voltage (or go with another motherboard vendor).

Stability was very important to me as I need my system for work, so I splashed out and bought a new mobo.

theRedCyclops · 2024-09-09 15:22:26

I'm guessing that increasing the voltage would be overclocking?
I was thinking of just taking it to a pc repair shop, but if the issue only occurs on linux they probably wouldn't even detect it

Last edited by theRedCyclops (2024-09-09 15:32:41)

seth · 2024-09-09 15:29:07

seth wrote:

In doubt make a photo of the RAM and CPU configs and link them here so we know what your options actually are.

theRedCyclops · 2024-09-09 15:36:39

https://0x0.st/XxH4.jpg

agapito · 2024-09-09 19:09:13

Assuming your memory is fine, you should test your CPU using CoreCycler on Windows. You can find hundreds of tutorials on the web on how to do it.

seth · 2024-09-09 19:37:18

theRedCyclops wrote:

https://0x0.st/XxH4.jpg

Ok, but what are your options besides "Auto", can you actually disable the entire "AI tweaker" thing? Or is it "auto" XOR "EXPO"?

theRedCyclops · 2024-09-09 19:38:50

I don't have a windows install, how do you propose I use it? install media? a vm? wine?

Arch Linux

#26 2024-09-08 20:10:46

Re: Random system crashes that seemingly follow even across reinstalls

#27 2024-09-08 21:27:43

Re: Random system crashes that seemingly follow even across reinstalls

#28 2024-09-08 21:37:13

Re: Random system crashes that seemingly follow even across reinstalls

#29 2024-09-08 22:41:21

Re: Random system crashes that seemingly follow even across reinstalls

#30 2024-09-09 05:53:24

Re: Random system crashes that seemingly follow even across reinstalls

#31 2024-09-09 08:59:23

Re: Random system crashes that seemingly follow even across reinstalls

#32 2024-09-09 09:02:52

Re: Random system crashes that seemingly follow even across reinstalls

#33 2024-09-09 09:09:22

Re: Random system crashes that seemingly follow even across reinstalls

#34 2024-09-09 10:01:26

Re: Random system crashes that seemingly follow even across reinstalls

#35 2024-09-09 10:19:47

Re: Random system crashes that seemingly follow even across reinstalls

#36 2024-09-09 11:23:47

Re: Random system crashes that seemingly follow even across reinstalls

#37 2024-09-09 11:30:58

Re: Random system crashes that seemingly follow even across reinstalls

#38 2024-09-09 12:01:16

Re: Random system crashes that seemingly follow even across reinstalls

#39 2024-09-09 13:49:15

Re: Random system crashes that seemingly follow even across reinstalls

#40 2024-09-09 13:54:29

Re: Random system crashes that seemingly follow even across reinstalls

#41 2024-09-09 14:31:02

Re: Random system crashes that seemingly follow even across reinstalls

#42 2024-09-09 14:35:30

Re: Random system crashes that seemingly follow even across reinstalls

#43 2024-09-09 14:57:08

Re: Random system crashes that seemingly follow even across reinstalls

#44 2024-09-09 15:13:20

Re: Random system crashes that seemingly follow even across reinstalls

#45 2024-09-09 15:22:26

Re: Random system crashes that seemingly follow even across reinstalls

#46 2024-09-09 15:29:07

Re: Random system crashes that seemingly follow even across reinstalls

#47 2024-09-09 15:36:39

Re: Random system crashes that seemingly follow even across reinstalls

#48 2024-09-09 19:09:13

Re: Random system crashes that seemingly follow even across reinstalls

#49 2024-09-09 19:37:18

Re: Random system crashes that seemingly follow even across reinstalls

#50 2024-09-09 19:38:50

Re: Random system crashes that seemingly follow even across reinstalls

Board footer