AMD's Raven Ridge Botchy Linux Support Appears Worse With Some Motherboards/BIOS

[Sonadow]

waiting on 4.16 to release ... I am a Gentoo person :-)

Kernel 4.16 is not going to solve any problems since the fault lies in the raven ridge portions of the amdgpu stack, and 4.16 is already in its rc stages.

In other words, it has got completely nothing to do with the CPU portion, so you can safely ignore every single last word blabbered by the person you quoted.

[devius]

Sorry, but Linux doesn't have 85% of the smartphone business, period.

https://www.idc.com/promo/smartphone-market-share/os

[bridgman]

1. There's a laptop APU driver for ALL the APUs on AMD support site except Ryzen mobile ones, even the desktop APUs are accounted for
2. The driver on HP web if marked for Windows 10 1703, we have 1709 as the latest, Intel Kaby Lake driver is 1709
3. User experiments that successfuly installed the Adrenaline 18.2.1/2 saw a 15-30% GPU performance increase (try TechEpiphany on youtube f.e.) - this did not work for me I guess because of the display panel, but I am really not sure

So ... I am denied performance by HP and AMD ... either through incompetence or ignorance. And not one of them does seem to care ... No drivers on AMD site, nothing new on HP site ...

At first glance it appears the "performance increase" with the Adrenaline drivers is coming from forcing a different power profile which may or may not fit within the cooling solution in the Envy. It's not as simple as "the newer drivers are faster".

[haplo602]

At first glance it appears the "performance increase" with the Adrenaline drivers is coming from forcing a different power profile which may or may not fit within the cooling solution in the Envy. It's not as simple as "the newer drivers are faster".

That might be so, but still, I doubt I can force the APU over the 25W cTDP by just drivers ... at least not long term ... however the benchmarks indicate that it is not just a short-term increase .... Also the same has no effect on the 2700u ... that one seems to regress ... but that might be something else as those are different laptops ...

but 30% performance is a LOT in a mobile chip ...

[bridgman]

That might be so, but still, I doubt I can force the APU over the 25W cTDP by just drivers ... at least not long term ...

The Windows drivers also carry a copy of power management microcode so potentially they might be able to over-ride power limits. Not sure if the microcode is different from OEM drivers though.

[agd5f]

The Windows drivers also carry a copy of power management microcode so potentially they might be able to over-ride power limits. Not sure if the microcode is different from OEM drivers though.

The SMU microcode is delivered via the sbios on APUs. There's no way for the driver to override it, regardless of the OS.

[agd5f]

This is exactly what happenes. Has been discussed in https://bugs.freedesktop.org/show_bug.cgi?id=102820 - and I guess for the next few years people using amdgpu will need to revert the commit named "drm/amd/display: Block 6Ghz timing if SBIOS set HDMI_6G_en to 0" from their local kernels in order to have working HDMI 2.0 support. Because neither the GPU board manufacturers nor MicroSoft will bother to change their "not spec compliant" implementations.

The Windows driver enforces this as well. If there are inconsistencies, it's a bug.

[bridgman]

The SMU microcode is delivered via the sbios on APUs. There's no way for the driver to override it, regardless of the OS.

D'oh !! ... right

[dwagner]

The Windows driver enforces this as well. If there are inconsistencies, it's a bug.

Maybe. The latest patch Harry just attached to https://bugs.freedesktop.org/show_bug.cgi?id=102820 does indeed fix the symptom for me.

(If only the other bugs that haunt my amdgpu driver were as easy to fix by either some official or some locally-applied-by-myself patch...)

[starshipeleven]

Sorry, but Linux doesn't have 85% of the smartphone business, period.

Globally, it has 88%.

https://qz.com/826672/android-goog-just-hit-a-record-88-market-share-of-all-smartphones/

Locally the percentages differ. In the US and a few other countries Apple has more penetration than its less-than-20% world average, but Android still has a large lead.