Question AMD CPUs with Linux OS.

Terabyte 11

Junior Member
Oct 26, 2019
19
0
6
I'm working on selecting parts for a new build and was hoping to get a Ryzen 9. However the PC has to be able to run Linux OS, and I've recently read a couple articles saying that AMD CPUs have issues with Linux. Has anybody here had any issues with Linux using AMD CPUs, or is this issue just being overstated? If it is an issue then is there any way to fix it or do I just need to use an Intel CPU?

Thanks for all the help.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,551
14,509
136
I'm working on selecting parts for a new build and was hoping to get a Ryzen 9. However the PC has to be able to run Linux OS, and I've recently read a couple articles saying that AMD CPUs have issues with Linux. Has anybody here had any issues with Linux using AMD CPUs, or is this issue just being overstated? If it is an issue then is there any way to fix it or do I just need to use an Intel CPU?

Thanks for all the help.
So far, I have had an 1800x, 2 2700x's, and 3 3900x's, and 8 threadrippers(5 1950x's, 1 2970wx, and 2 2990wx's), all running linux and not a ONE has had any problems.

I think that answers your concerns.

Edit: first I used cinnamon mint 18, now 19.2
 
Last edited:

moinmoin

Diamond Member
Jun 1, 2017
4,949
7,658
136
There are no general issues with Zen processors under Linux. What you need to keep in mind though is that AMD keeps adding/changing tech in each CPU/(especially) APU gen so to have everything supported on the newest chips you best run the latest kernel version available for the best possible support.

What Linux distribution do you intend to use?
 

Topweasel

Diamond Member
Oct 19, 2000
5,436
1,654
136
Original Ryzen had a bug that showed up during Linux compiles. AMD would swap those out under warranty and was only for the first 3 months of production. Recently they had a microcode issue where the random number generator was outputting the same number. This apparently used quite a bit in linux code, and would cause some kind of loop. Anyways that was on launch bios's. So any motherboard that has gotten updated since should work fine.
 
  • Like
Reactions: lightmanek

moinmoin

Diamond Member
Jun 1, 2017
4,949
7,658
136
Recently they had a microcode issue where the random number generator was outputting the same number. This apparently used quite a bit in linux code, and would cause some kind of loop.
That was "only" the init code of systemd actually. Since then both systemd has been updated to check for such repeats of numbers as well as the Linux kernel black listed the instruction for AMD chips. (Only systemd directly used the instruction to begin with since its original purpose is offering better entropy for cryptography early on in the boot up phase. There are more commonly used better interfaces for random data at the later point, like /dev/random.)
 

Shamrock

Golden Member
Oct 11, 1999
1,438
558
136
This guy (Chris Titus) has ONLY AMD products in his linuxes. Has a Zen 1 1700x and a Zen+ 2600, both with AMD video cards.

 

Charlie22911

Senior member
Mar 19, 2005
614
228
116
I had a 1950x, the only issue I ever experienced was that virtualization hardware pass through would lockup my unraid box. Those was quite some time ago and is probably fixed now, just use a recent up to date distro and you’ll be fine.
 

yeshua

Member
Aug 7, 2019
166
134
86
I'm working on selecting parts for a new build and was hoping to get a Ryzen 9. However the PC has to be able to run Linux OS, and I've recently read a couple articles saying that AMD CPUs have issues with Linux. Has anybody here had any issues with Linux using AMD CPUs, or is this issue just being overstated? If it is an issue then is there any way to fix it or do I just need to use an Intel CPU?

Thanks for all the help.

The only issue is that there's no HWiNFO64 for Linux, so your monitoring abilities will be severely limited (most people don't care about that anyways). Aside from that Linux works perfectly with X570 mobos/Ryzen 3000 CPUs.

I currently run Fedora 31 with Ryzen 7 3700X/ASUS X570 and there's nothing to report really.

Here are all the bad things kernel 5.3 reports:

Code:
efi: memattr: Entry attributes invalid: RO and XP bits both cleared
pci 0000:00:03.1: ASPM: current common clock configuration is broken, reconfiguring
Expanded resource Reserved due to conflict with PCI Bus 0000:00
PPR X2APIC NX GT IA GA PC GA_vAPIC
Hangcheck: starting hangcheck timer 0.9.1 (tick is 180 seconds, margin is 60 seconds).
i8042: PNP: PS/2 appears to have AUX port disabled, if this is incorrect please boot with i8042.nopnp
acpi PNP0C14:01: duplicate WMI GUID 05901221-D566-11D1-B2F0-00A0C9062910 (first instance was on PNP0C14:00)
acpi PNP0C14:02: duplicate WMI GUID 05901221-D566-11D1-B2F0-00A0C9062910 (first instance was on PNP0C14:00)
hid-generic 0003:046D:C042.0002: device has no listeners, quitting
usb 3-6: config 1 has an invalid interface number: 2 but max is 1
usb 3-6: config 1 has no interface number 1
hid-generic 0003:0B05:18F3.0003: device has no listeners, quitting
ACPI Warning: SystemIO range 0x0000000000000295-0x0000000000000296 conflicts with OpRegion 0x0000000000000290-0x0000000000000299 (\AMW0.SHWM) (20190703/utaddress-204)

As far as I can see they don't affect anything.