You are not logged in.

#1 2019-08-12 22:49:43

cmmps
Member
From: Portugal, Lisbon
Registered: 2012-04-26
Posts: 64

AMD Ryzen7 8 core sometimes freezes or crashes

Hello there

I'm having troubles with my computer. Sometimes it crashes and reboots with CPU hardware errors or just freezes.

When it reboots I find some errors regarding the CPU microcode and I can find the following when I run "dmesg":

[    0.186674] mce: [Hardware Error]: Machine check events logged
[    0.186676] mce: [Hardware Error]: CPU 3: Machine Check: 0 Bank 5: bea0000000000108
[    0.186680] mce: [Hardware Error]: TSC 0 ADDR 1ffffb10bfd50 MISC d012000101000000 SYND 4d000000 IPID 500b000000000 
[    0.186684] mce: [Hardware Error]: PROCESSOR 2:800f11 TIME 1565648406 SOCKET 0 APIC 3 microcode 8001129

Specs:

Operating System: Arch Linux
KDE Plasma Version: 5.16.3
KDE Frameworks Version: 5.60.0
Qt Version: 5.13.0
Kernel Version: 5.2.3-arch1-1-ARCH
OS Type: 64-bit
Motherboard: MSI B350 PC MATE (MS-7A34)
Processors: AMD Ryzen 7 1700X Eight-Core Processor
Memory: 15.7 GiB of RAM (16Gb)
Graphics card: Nvidia GeForce GTX1060 3Gb/PCIe/SSE2

I don't have the package "amd-ucode" installed. Do you think this might be the problem? Or should I worry that my CPU is not OK?

Thanks

Last edited by cmmps (2019-08-12 22:57:46)


Carlos Santos

Offline

#2 2019-08-13 10:29:20

Morn
Member
Registered: 2012-09-02
Posts: 617

Re: AMD Ryzen7 8 core sometimes freezes or crashes

Are you using "processor.max_cstate=5" in your kernel boot options? If not, try it.

Offline

#3 2019-12-10 19:13:20

nerdonski
Member
Registered: 2019-12-10
Posts: 1

Re: AMD Ryzen7 8 core sometimes freezes or crashes

I got same CPU + similar board (AMD Ryzen 7 1700x + MSI B350 GAMING PRO CARBON board) and had also random reboot issues with this "Hardware Error" message. The new BIOS update from MSI (7B00v1JQ Beta version, Release Date 2019-12-03) might be a real fix. System running stable for some days now.. Fingers crossed. Running Manjaro on a 4.19 LTS kernel but I guess this is not relevant to the problem.

EDIT: Update didnt fix it.. sad

Last edited by nerdonski (2019-12-12 22:54:24)

Offline

#4 2020-02-25 21:00:54

cmmps
Member
From: Portugal, Lisbon
Registered: 2012-04-26
Posts: 64

Re: AMD Ryzen7 8 core sometimes freezes or crashes

Hello! It's been some time I didn't follow up my issue but I was trying to make sure about something. I still have this freeze / reboot error sometimes and when it happens, it happens when I'm browsing the net with firefox.
This might be a good clue.

I also decided to check the parameter processor.max_cstate=5 but I only found an example with "$ cat /sys/module/intel_idle/parameters/max_cstate" which gives me 9. But I have a AMD CPU. So, where do I find if the parameter is OK?
I changed my grub.cfg as follows:

menuentry 'Arch Linux' --class arch --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-simple-9b6b2425-331c-41da-b90f-0a69ec8d573d' {
        load_video
        set gfxpayload=keep
        insmod gzio
        insmod part_msdos
        insmod ext2
        set root='hd1,msdos1'
        if [ x$feature_platform_search_hint = xy ]; then
          search --no-floppy --fs-uuid --set=root --hint-ieee1275='ieee1275//disk@0,msdos1' --hint-bios=hd1,msdos1 --hint-efi=hd1,msdos1 --hint-baremetal=ahci1,msdos1  919d5512-9b11-4104-9cd6-0688e128fe1b
        else
          search --no-floppy --fs-uuid --set=root 919d5512-9b11-4104-9cd6-0688e128fe1b
        fi
        echo    'Loading Linux linux ...'
        linux   /vmlinuz-linux root=UUID=9b6b2425-331c-41da-b90f-0a69ec8d573d rw  loglevel=3 quiet processor.max_cstate=5
        echo    'Loading initial ramdisk ...'
        initrd  /initramfs-linux.img
}

Am I doing something wrong?

I can also try to update my BIOS.

Kind regards


Carlos Santos

Offline

#5 2020-02-26 08:26:24

Ropid
Member
Registered: 2015-03-09
Posts: 877

Re: AMD Ryzen7 8 core sometimes freezes or crashes

That intel_idle module isn't in use for you. I think the one in use on AMD Ryzen is "cpuidle". There are just two C-states visible to the OS on Ryzen CPUs. This means that "5" you try to use does nothing. If you want to try something with that max_cstate parameter, you should try "1". I don't know if this is a good idea, I don't like that the power usage and temperature will be higher.

Did you try different BIOS versions for your motherboard?

Did you still not install amd-ucode? That package is probably not needed if you use the latest BIOS for your motherboard, but if you use an old BIOS then it might do something interesting.

That 16GB of memory you are using... is that fast memory and did you enable it's "XMP" profile? You could try disabling it so that the memory runs at default speed (should be something slow like 2400MHz) and see if that changes something.

EDIT:

You can tweak that C-state stuff with the "cpupower" command from within the running system. You don't need to change the kernel command line and reboot for experimenting.

You can view the C-states on your CPU like this:

sudo cpupower idle-info

You can disable the second C-state like this:

sudo cpupower idle-set -d 2

Or like this if it's the same latency number for you as what I see here on my 2700X Ryzen:

sudo cpupower idle-set -D 400

You can switch all C-states back on like this:

sudo cpupower idle-set -E

Disabling the C-states seems to do something to the hardware. I can see the power usage going up by 5W here for me on my 2700X if I disable C2. I'm checking this with "sudo turbostat".

Last edited by Ropid (2020-02-26 08:35:57)

Offline

Board footer

Powered by FluxBB