New Firmware update Reportedly solves Ryzen 3000 boot issues Linux

Published by

Click here to post a comment for New Firmware update Reportedly solves Ryzen 3000 boot issues Linux on our message forum
https://forums.guru3d.com/data/avatars/m/189/189980.jpg
Booting Linux on the new Ryzen platform was a hit and miss. For me this is what bothered me since the launch. Phoronix explained in detail how it is an RdRand instruction issue on these newer processors manifested by systemd. Now, having this issue solved, my new computer can be built and the upgrade from Skylake finally done.
https://forums.guru3d.com/data/avatars/m/247/247876.jpg
Of course anticupidon was the first to comment here. =)
https://forums.guru3d.com/data/avatars/m/189/189980.jpg
^ Why is that??
data/avatar/default/avatar11.webp
Warnings "Event 17, WHEA-Logger" eliminated in Windows Event Log - THIS IS NOT FIXED
https://forums.guru3d.com/data/avatars/m/258/258664.jpg
anticupidon:

^ Why is that??
Not sure, but I too wasn't all surprised to see you first posting on Ryzen + Linux tbh, not in a bad way just... first 😀
https://forums.guru3d.com/data/avatars/m/247/247876.jpg
anticupidon:

^ Why is that??
Because of Linux.
Paulo Narciso:

Warnings "Event 17, WHEA-Logger" eliminated in Windows Event Log - THIS IS NOT FIXED
What, there is Windows Event Log in Linux?
https://forums.guru3d.com/data/avatars/m/189/189980.jpg
Sheesh, that makes me one of few Gnu/Linux G3D gurus... My avatar kinda gave me a leeway into this. But I'm ok with it, feels more natural as times passes to use Linux, rather than Windows. And Ryzen/Threadripper works better under Linux kernel, it's just a hidden truth.
mbk1969:

What, there is Windows Event Log in Linux?
https://kde.org/applications/system/org.kde.ksystemlog
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
I find it interesting how a lot of news sources seem to acknowledge that Destiny 2 and Linux had Ryzen 3000 problems, where rdrand was known to be the culprit, yet nobody seems to connect the two. By fixing Destiny 2, Linux in turn was also fixed.
anticupidon:

Sheesh, that makes me one of few Gnu/Linux G3D gurus... My avatar kinda gave me a leeway into this.
Same here. Haha the avatar only gives it away to Linux users (I'm also an Arch user, BTW). I don't think the logo means anything to most people.
data/avatar/default/avatar37.webp
i think that games apart, everything runs better on linux. I got a macbook for work, got used to it, and now if i have to choose between windows and linux i would take linux 100% or hackintosh.
https://forums.guru3d.com/data/avatars/m/270/270792.jpg
I had no problem till now. I'm using my old Debian 9 (kernel 4.9) installation (i5+Z77) on my R7 3700X flawless. 😀
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
asturur:

i think that games apart, everything runs better on linux. I got a macbook for work, got used to it, and now if i have to choose between windows and linux i would take linux 100% or hackintosh.
Actually if the game has a native port and didn't use a lazy DX to OGL translation, it's pretty common for games to run better on Linux. Really gets you to question why Windows drivers are hundreds of MB, when all GPU drivers for Linux combined make up less than 100MB (excluding Nvidia's proprietary drivers).
data/avatar/default/avatar27.webp
i think drivers are big becaue they may contain specific code for specific games. So called optimizations. I may be wrong tho.
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
asturur:

i think drivers are big becaue they may contain specific code for specific games. So called optimizations. I may be wrong tho.
That is definitely some of it, though as I alluded to before, you can often get the same performance without any of those application-specific optimizations. A lot of it is just bloatware. I do give somewhat of a pass for stuff like CUDA, since Nvidia is basically providing the entire API and ABI and a lot of the resources necessary to use it.
https://forums.guru3d.com/data/avatars/m/189/189980.jpg
asturur:

i think drivers are big becaue they may contain specific code for specific games. So called optimizations. I may be wrong tho.
Once, just for the kicks, I opened a Synaptics driver which was about 200 Mbs in size. Well, most of it was a how to guide and lots pictures and short movies/animations about how to scroll. The actual driver was around 10 mega. Now, really one can appreciate the simpler approach of the Linux kernel having the drivers baked in. No bloatware, no nonsense. But I digress, sorry
data/avatar/default/avatar23.webp
When I read such news I appreciate more distros like Gentoo that do not jump on the latest (broken) software as soon as it’s available. First it gives me the possibility of not using systemd, and the default kernel is still 4.9 too. (That’s not to say that Gentoo is perfect; it isn’t.)
https://forums.guru3d.com/data/avatars/m/163/163032.jpg
Hilbert Hagedoorn:

We don't talk about Linus a lot,
Speaking if Linus....this article was brought to you by.... Guru3d!
data/avatar/default/avatar16.webp
I just finished my build with a Ryzen 3900X and a MSI MPG gaming mainboard and there is definatly some strange behavior when i start the PC on first boot it will get stuck in windows load or if i go to th BIOS it will lock up, then after 1-2 power off it will boot normally, other than that the CPU is running very smooth and i already love.
https://forums.guru3d.com/data/avatars/m/271/271612.jpg
Yes, AMD did indeed somehow fix the RDRAND problem with the 1.0.0.3ABB AGESA version. The only confusion there I guess is that motherboard manufacturers have picked different names for the RDRAND problem when writing changelogs etc, many choosing to refer to it as the "Destiny 2" problem.
https://forums.guru3d.com/data/avatars/m/273/273678.jpg
just a reminder you need to power off the system after applying this patch for the chipset driver fix to no longer disable the RD instructions.
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
hawk7000:

Yes, AMD did indeed somehow fix the RDRAND problem with the 1.0.0.3ABB AGESA version.
To my recollection, the "fix" was to just simply turn it off, where anything depending on it will default to a different random number generator.