AMD hardware support in Manjaro

I may suck it up and go with the acer, I could always pick back up some of my older games and run them under wine heh I don’t run any new new games, everything I did play ran fine on my gtx 960m so the rx580 would run them a bit better.

And another great news:
Google Stadia will run on Linux with AMD GPU’s using Vulkan and Unity is already fully on the train. The GPU’s are said to have 10,7 TFlops and 16GB HBM2 each, so it is either a more powerful custom Vega56 or a Vega VII that is heavily optimized on efficiency (with lower clocks and power targets).

This means that now, in addition to Valve, Redhat and others, Google and Unity will also contribute to the AMD driver stack ^^

3 Likes

That would be the best case, but since this is google i think they will have an own fork for this own Linux branch of them mixed with the proprietary AMD vulkan stack.

It’s not like google has a fetish for forks… If your mind goes back to thinking about android, that just simply happen because ARM world is a clusterduck lagging a decade (if not more) behind x86 standardization.

google has a history of contributing to the mainline kernel, at least in the last years. They are also a Platinum member of the Linux Foundation since middle of 2018: https://www.linuxfoundation.org/press-release/2018/06/google-becomes-platinum-member-of-linux-foundation-demonstrating-its-commitment-to-the-open-source-community/

Also, which “proprietary Vulkan stack” are you talking about? Is there another driver besides AMDVLK and Radv? Because AMDVLK is open source for quite some time now, and Radv always was.

And thats probably not a good thing … :sweat_smile: https://youtu.be/TVHcdgrqbHE?t=1002

3 Likes

He’ll mean that since it’s MIT (not much dissimilarly from android’s BSD), they’ll be able to keep closed the parts they want.
Which, I mean, after all is actually the reason they still don’t buy into radv imo.

But, fun fact, mesa is too. So gg for tinfoiling.

@anon35400795 lunduke went total bonkers with that talk

1 Like

Oh, so you mean insane, smarter-than-you and everyone else because I say so, must market myself with even more insanity, Lunduke? That one? No way. :laughing:

I used to watch Lunduke a lot, but he went from facts to very un-objective meanings way too much.
I don’t see it as a bad thing that Google joined the Linux foundation. And I also don’t care that much about what license a project is opensourced with, as long as it is any open source license.

Reporting that linux 5.0.3-1 kernel gave me system freezes after certain periods of time. Didn’t have the time yet to investigate, back at 4.19 and all good.

Not that there's anything wrong with standard normal open drivers of course.
But I'd just like to report how latest AMDGPU-PRO has been made compatible with Arch
(I'm not sure all packages could work out of the box here, without some mhwd tinkering, but still)

Well, we had the warnings for ages now within the compiled drivers. Still have to figure out why they switch to errors now. Also with 5.3 release to build extramodules will change. All in all it is getting harder to support proprietary drivers in general.

1 Like

I feel like you replied in the wrong thread, anyway please don't waste your time in these menial tasks.
Simply don't offer the modules for newer kernels and we are square. As I was saying, it will still be like almost 10 years before 4.4 goes out of support :slight_smile:
And in the meantime perhaps somebody else will have done the work for you.

my opinion: Don't offer the -pro driver. It has no performance gains over the free driver and is getting more and more obsolete. Reason: More and more stuff is opensourced and thus available with the free driver package anyway.

6 Likes

Pro would be a nice "meta-package" superset of both ROCm and opencl-amd though.
Also, I have never tried it, but IIRC it let you choose about what kind of opengl (and vulkan?) ICD you wanted to use.

Perhaps the re-enable wait in pipelock, but add timeout patch can be added for Navi users while it makes its way into the mainline? I tried to make the commit myself in https://gitlab.manjaro.org/packages/core/linux54/ but I don't seem to be able to unfortunately

Good news everyone.
NIR for r600 is getting tesselation. Which means.. I think with the exception of compute shaders (and performance) it should give you the whole opengl 4.5?

In a different twist, I also believe to have finally understood why people were reporting wildly different experiences with fglrx (which doesn't compile anymore since 5.4).
It wasn't even the DE. It was the display manager. LightDM is perfect, but try sddm and possibly gdm, and you get the famous dead black screen.

Hi everyone! After last mesa-driver update I have troubles with games. I'm playing Elder Scrolls Online through steam and with mesa 20.0.4-1 I have 50-60 fps, but after installing update (mesa 20.0.4-2 and it dependencies) my fps in ESO drops down to 10fps. I've loaded last system image from Timeshift and now I'm again at mesa 20.0.4-1 and everything works great. How can I find out what is the reason of this problems ?

System:    Host: madmagnetic-lp Kernel: 5.6.2-pf x86_64 bits: 64 compiler: gcc v: 9.3.0 Desktop: KDE Plasma 5.18.4 
           tk: Qt 5.14.2 info: latte-dock wm: kwin_x11 dm: SDDM Distro: Manjaro Linux 
Machine:   Type: Laptop System: ASUSTeK product: TUF Gaming FX705DY_FX705DY v: 1.0 serial: <filter> 
           Mobo: ASUSTeK model: FX705DY v: 1.0 serial: <filter> UEFI: American Megatrends v: FX705DY.315 date: 03/09/2020 
Battery:   ID-1: BAT0 charge: 59.5 Wh condition: 59.8/64.0 Wh (93%) volts: 15.6/15.6 model: ASUSTeK ASUS Battery type: Li-ion 
           serial: <filter> status: Not charging 
CPU:       Topology: Quad Core model: AMD Ryzen 5 3550H with Radeon Vega Mobile Gfx bits: 64 type: MT MCP arch: Zen+ rev: 1 
           L2 cache: 2048 KiB 
           flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm bogomips: 33535 
           Speed: 3270 MHz min/max: 1400/2100 MHz boost: enabled Core speeds (MHz): 1: 3270 2: 3436 3: 3239 4: 3374 5: 3015 
           6: 3226 7: 2959 8: 3442 
Graphics:  Device-1: AMD Baffin [Radeon RX 460/560D / Pro 450/455/460/555/555X/560/560X] vendor: ASUSTeK driver: amdgpu 
           v: kernel bus ID: 01:00.0 chip ID: 1002:67ef 
           Device-2: Advanced Micro Devices [AMD/ATI] Picasso vendor: ASUSTeK driver: amdgpu v: kernel bus ID: 04:00.0 
           chip ID: 1002:15d8 
           Display: x11 server: X.Org 1.20.8 driver: amdgpu FAILED: ati unloaded: modesetting alternate: fbdev,vesa 
           compositor: kwin_x11 resolution: 1920x1080~60Hz 
           OpenGL: renderer: AMD RAVEN (DRM 3.36.0 5.6.2-pf LLVM 9.0.1) v: 4.6 Mesa 20.0.4 direct render: Yes 
Audio:     Device-1: Advanced Micro Devices [AMD/ATI] Raven/Raven2/Fenghuang HDMI/DP Audio vendor: ASUSTeK 
           driver: snd_hda_intel v: kernel bus ID: 04:00.1 chip ID: 1002:15de 
           Device-2: Advanced Micro Devices [AMD] Family 17h HD Audio vendor: ASUSTeK driver: snd_hda_intel v: kernel 
           bus ID: 04:00.6 chip ID: 1022:15e3 
           Sound Server: ALSA v: k5.6.2-pf 
Network:   Device-1: Realtek RTL8821CE 802.11ac PCIe Wireless Network Adapter vendor: AzureWave driver: rtl8821ce v: N/A 
           port: e000 bus ID: 02:00.0 chip ID: 10ec:c821 
           IF: wlp2s0 state: down mac: <filter> 
           Device-2: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet vendor: ASUSTeK driver: r8169 v: kernel port: d000 
           bus ID: 03:00.0 chip ID: 10ec:8168 
           IF: enp3s0 state: up speed: 1000 Mbps duplex: full mac: <filter> 
Drives:    Local Storage: total: 931.51 GiB used: 301.49 GiB (32.4%) 
           ID-1: /dev/sda vendor: Toshiba model: MQ04ABF100 size: 931.51 GiB speed: 6.0 Gb/s rotation: 5400 rpm 
           serial: <filter> rev: 0J scheme: GPT 
Partition: ID-1: / size: 388.11 GiB used: 301.45 GiB (77.7%) fs: ext4 dev: /dev/sda6 
Sensors:   System Temperatures: cpu: 61.4 C mobo: N/A 
           Fan Speeds (RPM): cpu: 2900 
           GPU: device: amdgpu temp: 61 C device: amdgpu temp: 60 C 
Info:      Processes: 301 Uptime: 2h 14m Memory: 15.52 GiB used: 6.79 GiB (43.8%) Init: systemd v: 244 Compilers: gcc: 9.3.0 
           clang: 9.0.1 Shell: fish v: 3.1.0 running in: yakuake inxi: 3.0.37

Upd.
Today after updating all packages everything works fine.

Quick question, I have not read the whole post here, very new to this forum.

I would like to setup a Window 10 PC and Manjaro Cinnamon as a Virtual box (computer)
A. Window 10 pro and Manjaro Cinnamon (windows only apps & Blender 3D)
B. Two graphic cards ( RX 5700 Xt & RX 5500 XT)
C. Each OS use a separate graphic cards
D. Shared a CPU (AMD) 24 cores
E. 32GB of memory

The reason is to be able to switch back and forth easily without constrain the resources; any suggestion? have anyone did this before? Is it possible to do it?

Nice Topic

Forum kindly sponsored by