AGESA 1.0.0.6b Might Fix The Ryzen Linux Performance Marginality Problem

[chithanh]

Is there any brand that openly supports linux on consumer hardware and will actually care about linux-related issues on their hardware?

None of the big mobo vendors officially support Linux. Their first-level support will either refuse to help you, or will forward your support ticket to some engineer who may or may not understand what Linux is.

Your best bet is letting others make the experience first, and then buy what is proven to work. Some Phoronix reviews of AM4 mobos:
ASRock AB350 Pro4: A Decent, Linux-Friendly Ryzen Motherboard
Gigabyte AB350N-GAMING WiFi: An Ideal Mini-ITX Ryzen Motherboard For Linux
MSI B350 TOMAHAWK: A Capable AMD Ryzen Motherboard

Some specialized vendors like Fujitsu at least acknowledge that Linux exists and that you can run it on their mobos, but don't guarantee that it will work well.

[RavFX]

I would recommend shying away from gigabyte motherboards for linux support. Gigabyte doesn't seem to care about linux and the linux community had to work around an gpio issue with some of their ryzen motherboards.

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1671360

Basically all they could do was make the code more robust so that when this stuff blew up, the kernel was able to recover.

https://lkml.org/lkml/2017/5/23/1099

Gigabyte control some of there thing using GPIO...

Also GPIO should not be enabled in a kernel destined for a desktop computer... People does not use GPIO on desktop, they use GPIO on dev board, embedded stuff... thing like that that does not run Ubuntu anyway. It was Ubuntu fault.

[mat69]

I've just red a post where the problem is addressed to some Ryzen CPus realized before 25th weeks of the year. Is it confirmed?

This is the post: https://community.amd.com/message/2821277

I have a Ryzen 7 1700 that was built in week 28 ("UA 1728PGT" is on the chip) and have crashes too when running kill_ryzen.sh.
Unfortunately there isn't a bios update existing for my board yet.

[starshipeleven]

Gigabyte control some of there thing using GPIO...

Also GPIO should not be enabled in a kernel destined for a desktop computer... People does not use GPIO on desktop, they use GPIO on dev board, embedded stuff... thing like that that does not run Ubuntu anyway. It was Ubuntu fault.

I thought GPIO was used for buttons, leds or other oddball functions also on desktop.

[BillBroadley]

I haven't bought a Ryzen yet. I don't care about size (uATX, MATX, or ATX), pci-e slots (1 for a GPU is fine), or overclocking. I just want a stable Ryzen desktop, ECC working would be a bonus I'd pay extra for.

Any recommendations?

Have the fixed (week 25 or later) Ryzen CPUs made it through the pipeline at Amazon and/or Newegg?

[Anty]

I have a Ryzen 7 1700 that was built in week 28 ("UA 1728PGT" is on the chip) and have crashes too when running kill_ryzen.sh.

WOW
This is pretty bad news because all believe batches after 1725 are the good ones
Could you share it on AMD forum in this thread:

https://community.amd.com/thread/215773?start=0&tstart=0

[audir8]

As others have said in other threads, Newegg and Amazon don't have past week-25 stock yet. I built a system using the labor day sale on newegg, and got a week 21 part. DDR4 prices are something to keep an eye on if you're going to build a system soon. You might actually want to buy RAM now since prices on the TridentZ stick I bought have gone up by $50 in two weeks and aren't forecast to go down. As far as boards go, manufacturers can use the PCI-E lanes for M.2 or PCI-E or even USB. I would say in general, Asrock has boards with 2x M.2, MSI/ASUS have more PCI-E 16x..1x slots and Gigabyte has more USB 3.1 Gen2/1 ports. There are exceptions offcourse, but get the I/O connectivity you want/need. Keep in mind a lot of board makers disable SATA or reduce M.2 speed if you have PCI-E/M.2 slots filled or an APU/Ryzen processor.

The AGESA 1.0.0.6b update might have affected the bug if AMD mandated the option to turn opcode cache off in BIOSes. My Gigabyte AB350M-D3H had the option before and after the AGESA 1.0.0.6b update (I received the board with BIOS version F2).

[audir8]

WOW
This is pretty bad news because all believe batches after 1725 are the good ones
Could you share it on AMD forum in this thread:

https://community.amd.com/thread/215...art=0&tstart=0

Yep, makes me want to get A12-9800 and return my Ryzen. Waiting until next year wouldn't be a bad idea.

[LinuxID10T]

WOW
This is pretty bad news because all believe batches after 1725 are the good ones
Could you share it on AMD forum in this thread:

https://community.amd.com/thread/215...art=0&tstart=0

I was under the impression it was the batches after 1730 that were good. I could again be wrong though.

[rk17]

Well they actually do ask you to go that high.
For me it was "up to a maximum of 1.425V (but not going over)".
Still did not help so RMA it is.

Vcore voltages can be adjusted in A320M boards?
If not possible then adjusting this voltages to solve this bug(even if it works) on B350 and X370 boards is just a paper on the crack.
I had ryzen 5 1600 and asus A320M mobo and i sold it before this bug become wildly famous.

On another note, If AMD know that AGESA 1.0.0.6b will fix this bug then they should have issued a public statement about the bios microcode update solution rather than approving thousands of RMAs which is a PITA and corrodes their reputation a little.