You are not logged in.

#1 2019-06-07 18:40:56

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Optimising OBS VA-API usage with AMD and Intel GPUs

Hi!

I'm new to Arch, hence my posting here. I installed it last night/this morning with the intention of testing D9VK, a d3d9-->vulkan wrapper made by a friend, for use with Windows games.
I want to screen record it, so I'm trying with OBS, and since I don't have a beastly CPU, I'd like to use GPU accelerated encoding for the stream.
I realised that OBS supports VA-API, and trying it with my AMD card led to significantly degraded performance in games (which wouldn't be good for performance comparisons). I thought I could perhaps use my Intel iGPU, and though it shows up as a VA-API input device, I get an error trying to use it.
I thought I'd installed the relevant drivers, though perhaps I'm wrong.

Curiously, vainfo doesn't seem to show the Intel GPU at all.


For lack of a better means of sharing info, have some screenshots:
https://imgur.com/cnj9dUb.png
https://imgur.com/Q8fcOjz.png
Working (with AMD GPU):
https://imgur.com/xeVhUZX.png
Not working (with Intel iGPU):
https://imgur.com/h2aoq9B.png

Error I get in the terminal from OBS with the Intel config:
https://imgur.com/01Eh9Xv.png
And in OBS itself:
https://imgur.com/B8bYvNx.png

Paste of my installed packages:
https://hastebin.com/hoqemamopu.css


I hope this is...decent info to work from, and posted in the right place. If you need anything more, please tell me how to provide it.
Advise me, please (:

Mod note: Replaced oversized images with urls. Please read the Code of Conduct and adhere to the image posting guidelines.

Last edited by WorMzy (2019-06-11 16:23:36)

Offline

#2 2019-06-07 19:00:21

ugjka
Member
From: Latvia
Registered: 2014-04-01
Posts: 1,794
Website

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

It looks like your card only supports mpeg2?


https://ugjka.net
paru > yay | webcord > discord
pacman -S spotify-launcher
mount /dev/disk/by-...

Offline

#3 2019-06-07 20:13:21

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

ugjka wrote:

It looks like your card only supports mpeg2?

Thanks for the reply!
At present, perhaps, but it should support what I'm after, according to this:
https://01.org/linuxgraphics/community/vaapi

Most concerning is, I guess, that it's not showing up under vainfo.

Offline

#4 2019-06-08 07:53:31

ugjka
Member
From: Latvia
Registered: 2014-04-01
Posts: 1,794
Website

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

If I understand correctly it is picking up your AMD card not IntelHD


https://ugjka.net
paru > yay | webcord > discord
pacman -S spotify-launcher
mount /dev/disk/by-...

Offline

#5 2019-06-08 12:27:17

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

First, don't post huge pictures, see https://wiki.archlinux.org/index.php/Co … s_and_code .

Your system appears to use hybrid graphics / PRIME , how did you set it up ?

post glxinfo -B and xorg log .


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#6 2019-06-08 23:50:58

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

First, don't post huge pictures, see https://wiki.archlinux.org/index.php/Co … s_and_code .

Your system appears to use hybrid graphics / PRIME , how did you set it up ?

post glxinfo -B and xorg log .

I see, apologies.

I didn't notice there were code tags available to me.

glxinfo -B:
http://sprunge.us/ZH2idH

For xorg...I apparently don't have it as an available command. Have I been solely using Wayland without realising? Am I missing something?

[tedster@BlackYeti ~]$ loginctl show-session 2 -p Type
Type=x11

Evidently I am using X server, but I don't know what I'd actually call to get the log you'd want.

In the case it's helpful, here's my Xorg.0.log:
http://sprunge.us/q35VTH

And...in terms of graphics setup, I'm not sure.
I installed the base system, then Budgie and LightDM, then GNOME, and then relevant Vulkan drivers and loaders.
All I've manually touched (apart from telling GRUB to load the correct AMD driver) would be represented in my list of packages, I think.

My grub.cnf:
http://sprunge.us/MTbxn2

When I experienced bad performance, I enabled my iGPU in the BIOS and installed whichever Intel Mesa packages, which should be listed in the list of packages linked above.

Thanks,
Ted

Offline

#7 2019-06-10 13:46:09

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

THe glxinfo output clearly shows you'rre using the amd videocard, but xorg logs indicate you're using modesetting.

Install xf86-video-amdgpu , reboot or restart X and post fresh xorg logs.
Also post full lpsci -k output please.

One thing to keep in mind :
I'm not looking at OBS or VA-API for now. Your graphics setup is less-then-optimal and we need to get things setup as best as possible first.


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#8 2019-06-10 22:09:19

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

THe glxinfo output clearly shows you'rre using the amd videocard, but xorg logs indicate you're using modesetting.

Install xf86-video-amdgpu , reboot or restart X and post fresh xorg logs.
Also post full lpsci -k output please.

One thing to keep in mind :
I'm not looking at OBS or VA-API for now. Your graphics setup is less-then-optimal and we need to get things setup as best as possible first.

Makes sense!

Installed that package; the info you wanted:

Xorg:
http://sprunge.us/O4AsOy

Lspci:
http://sprunge.us/g9pUyz

Again, thanks

Offline

#9 2019-06-11 11:14:51

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

lspci -k output is as expected, xorg logs look much better.


[     4.884] (II) AMDGPU(0): Output DisplayPort-0 connected
[     4.884] (II) AMDGPU(0): Output DisplayPort-1 disconnected
[     4.884] (II) AMDGPU(0): Output HDMI-A-2 connected
[     4.884] (II) AMDGPU(0): Output DVI-I-0 connected
[     4.884] (II) AMDGPU(0): Using spanning desktop for initial modes
[     4.884] (II) AMDGPU(0): Output DisplayPort-0 using initial mode 1920x1080 +0+0
[     4.884] (II) AMDGPU(0): Output HDMI-A-2 using initial mode 1920x1080 +1920+0
[     4.884] (II) AMDGPU(0): Output DVI-I-0 using initial mode 1920x1080 +3840+0

- You appear to have 3 monitors of the same type but using different connectors : DisplayPort, HDMI and DVI .
Is this intentional ?

- Those monitors are currently setup to form one big virtual screen, is that how you want things ?


[    44.704] (II) event3  - Logitech Gaming Mouse G402: SYN_DROPPED event - some input events have been lost.

Not sure what this means, but if may be a symptom of a problem . It should be looked at some point in a separate thread I think.


Post xrandr --listproviders please.


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#10 2019-06-11 15:30:52

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

lspci -k output is as expected, xorg logs look much better.


[     4.884] (II) AMDGPU(0): Output DisplayPort-0 connected
[     4.884] (II) AMDGPU(0): Output DisplayPort-1 disconnected
[     4.884] (II) AMDGPU(0): Output HDMI-A-2 connected
[     4.884] (II) AMDGPU(0): Output DVI-I-0 connected
[     4.884] (II) AMDGPU(0): Using spanning desktop for initial modes
[     4.884] (II) AMDGPU(0): Output DisplayPort-0 using initial mode 1920x1080 +0+0
[     4.884] (II) AMDGPU(0): Output HDMI-A-2 using initial mode 1920x1080 +1920+0
[     4.884] (II) AMDGPU(0): Output DVI-I-0 using initial mode 1920x1080 +3840+0

- You appear to have 3 monitors of the same type but using different connectors : DisplayPort, HDMI and DVI .
Is this intentional ?

- Those monitors are currently setup to form one big virtual screen, is that how you want things ?


[    44.704] (II) event3  - Logitech Gaming Mouse G402: SYN_DROPPED event - some input events have been lost.

Not sure what this means, but if may be a symptom of a problem . It should be looked at some point in a separate thread I think.


Post xrandr --listproviders please.

The display setup is as I want it in each case, yes.
For the mouse, it's probably just that the system couldn't keep up with the mouse's update rate at one point.

That xrandr command:

[tedster@BlackYeti ~]$ xrandr --listproviders
Providers: number : 2
Provider 0: id: 0x8f cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 6 outputs: 4 associated providers: 1 name:AMD Radeon HD 7900 Series @ pci:0000:01:00.0
Provider 1: id: 0x45 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 3 outputs: 3 associated providers: 1 name:modesetting

Offline

#11 2019-06-11 16:05:56

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Xrandr providers output is exactly what it should look like for a Hybrid Graphics system, check https://wiki.archlinux.org/index.php/PRIME for details .

run (and post outputs)

$ vainfo
$ xrandr --setprovideroffloadsink 1 0
$ DRI_PRIME=1 glxinfo -B
$ DRI_PRIME=1 vainfo

Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#12 2019-06-11 16:34:03

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

Xrandr providers output is exactly what it should look like for a Hybrid Graphics system, check https://wiki.archlinux.org/index.php/PRIME for details .

run (and post outputs)

$ vainfo
$ xrandr --setprovideroffloadsink 1 0
$ DRI_PRIME=1 glxinfo -B
$ DRI_PRIME=1 vainfo

Respectively:

vainfo: VA-API version: 1.4 (libva 2.4.1)
vainfo: Driver version: Mesa Gallium driver 19.0.6 for AMD Radeon HD 7900 Series (TAHITI, DRM 3.30.0, 5.1.8-arch1-1-ARCH, LLVM 8.0.0)
vainfo: Supported profile and entrypoints
      VAProfileMPEG2Simple            :	VAEntrypointVLD
      VAProfileMPEG2Main              :	VAEntrypointVLD
      VAProfileNone                   :	VAEntrypointVideoProc
name of display: :0
libGL error: Different GPU, but blitImage not implemented for this driver
libGL error: failed to load driver: i965
display: :0  screen: 0
direct rendering: Yes
Extended renderer info (GLX_MESA_query_renderer):
    Vendor: Intel Open Source Technology Center (0x8086)
    Device: Mesa DRI Intel(R) Haswell Desktop  (0x412)
    Version: 19.0.6
    Accelerated: yes
    Video memory: 1536MB
    Unified memory: yes
    Preferred profile: core (0x1)
    Max core profile version: 4.5
    Max compat profile version: 3.0
    Max GLES1 profile version: 1.1
    Max GLES[23] profile version: 3.1
OpenGL vendor string: Intel Open Source Technology Center
OpenGL renderer string: Mesa DRI Intel(R) Haswell Desktop 
OpenGL core profile version string: 4.5 (Core Profile) Mesa 19.0.6
OpenGL core profile shading language version string: 4.50
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile

OpenGL version string: 3.0 Mesa 19.0.6
OpenGL shading language version string: 1.30
OpenGL context flags: (none)

OpenGL ES profile version string: OpenGL ES 3.1 Mesa 19.0.6
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.10
kmsro: driver missing
kmsro: driver missing
libva error: /usr/lib/dri/radeonsi_drv_video.so init failed
vaInitialize failed with error code 2 (resource allocation failed),exit

Offline

#13 2019-06-12 09:00:34

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Not what i hoped to see, especially the libGL errors.

I'm beginning to think the motherboard/firmware/chipset don't support current config very well, but let's look more in-depth for now.
Please post dmesg and/or journalctl -b  .


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#14 2019-06-13 03:27:53

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

Not what i hoped to see, especially the libGL errors.

I'm beginning to think the motherboard/firmware/chipset don't support current config very well, but let's look more in-depth for now.
Please post dmesg and/or journalctl -b  .

Quite verbose!
http://sprunge.us/ksnOhD

I'm...removing the majority of the journalctl output because it's 31MB, hopefully this is what you're after:
http://sprunge.us/EceDtT

Offline

#15 2019-06-13 08:00:43

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Several messages indicate low-level problems.

Jun 10 23:03:22 BlackYeti kernel: [Firmware Bug]: TSC_DEADLINE disabled due to Errata; please update microcode to version: 0x22 (or later)

Jun 10 23:03:22 BlackYeti kernel: Your BIOS is broken; DMAR reported at address 0!
                                  BIOS vendor: American Megatrends Inc.; Ver: F2; Product Version: To be filled by O.E.M.

Jun 10 23:03:23 BlackYeti kernel: resource sanity check: requesting [mem 0x000c0000-0x000dffff], which spans more than PCI Bus 0000:00 [mem 0x000d0000-0x000d3fff window]
Jun 10 23:03:23 BlackYeti kernel: caller pci_map_rom+0x6a/0x17d mapping multiple BARs
Jun 10 23:03:23 BlackYeti kernel: amdgpu 0000:01:00.0: No more image in the PCI ROM

Configure intel microcode updating, see https://wiki.archlinux.org/index.php/Microcode
post new dmesg after that.

Your motherboard is a Gigabyte Z87X-SLI , you're using bios version F2 .
https://www.gigabyte.com/Motherboard/GA … rt-dl-bios lists only 3 bios versions.
The latest, F3c is listed as a beta version.

Try to find out what experiences other GA-Z87X-SLI owners have with that bios.

N.B.
These issues are often not linux/OS-specific, linux just makes it easier to find them.


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#16 2019-06-13 18:01:17

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

Several messages indicate low-level problems.

Jun 10 23:03:22 BlackYeti kernel: [Firmware Bug]: TSC_DEADLINE disabled due to Errata; please update microcode to version: 0x22 (or later)

Jun 10 23:03:22 BlackYeti kernel: Your BIOS is broken; DMAR reported at address 0!
                                  BIOS vendor: American Megatrends Inc.; Ver: F2; Product Version: To be filled by O.E.M.

Jun 10 23:03:23 BlackYeti kernel: resource sanity check: requesting [mem 0x000c0000-0x000dffff], which spans more than PCI Bus 0000:00 [mem 0x000d0000-0x000d3fff window]
Jun 10 23:03:23 BlackYeti kernel: caller pci_map_rom+0x6a/0x17d mapping multiple BARs
Jun 10 23:03:23 BlackYeti kernel: amdgpu 0000:01:00.0: No more image in the PCI ROM

Configure intel microcode updating, see https://wiki.archlinux.org/index.php/Microcode
post new dmesg after that.

Your motherboard is a Gigabyte Z87X-SLI , you're using bios version F2 .
https://www.gigabyte.com/Motherboard/GA … rt-dl-bios lists only 3 bios versions.
The latest, F3c is listed as a beta version.

Try to find out what experiences other GA-Z87X-SLI owners have with that bios.

N.B.
<strong>These issues are often not linux/OS-specific, linux just makes it easier to find them</strong>.

Reminds me of the good old 'my code is legitimately fine, your compiler is broken, it works with the other ones' situations I hear about every now and then!

New dmesg after installing microcode package, regenerating GRUB config and rebooting:
http://sprunge.us/MLrxlE

Judging by this thread on OCUK, not sure I really want to try the beta BIOS:
https://forums.overclockers.co.uk/threa … .18617041/

Again, thanks for your help (:

Offline

#17 2019-06-16 12:43:13

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Understood.

There are some options in the firmware that could influence behaviour.
From "bios features' screen post values of
Fast Boot
VGA Support
Windows 8 Features

from Peripherals screen post
Initial Display Output


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#18 2019-06-16 21:29:10

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

Understood.

There are some options in the firmware that could influence behaviour.
From "bios features' screen post values of
Fast Boot
VGA Support
Windows 8 Features

from Peripherals screen post
Initial Display Output

Alright; I bit the bullet.
In the case it causes issues and I can't downgrade, I figure can just (hopefully) torture the system enough with forcibly failed boots to take me to the backup BIOS.
...it also took me far too long to figure out how to fix my GRUB UEFI setup so I could boot again (:

Tons of stuff again for thoroughness:
vainfo:
http://sprunge.us/Nj71RT
DRI_PRIME=1 glxinfo -B:
http://sprunge.us/3GviBz

DRI_PRIME=1 vainfo:

kmsro: driver missing
kmsro: driver missing
libva error: /usr/lib/dri/radeonsi_drv_video.so init failed
vaInitialize failed with error code 2 (resource allocation failed),exit

Interesting!
dmesg:
http://sprunge.us/u1AqIg


Regardless of the above, the various options you're asking for:

Fastboot:
Enabled

VGA Support:
EFI Driver

Windows 8 Features:
Other OS

For the inital display output, I have it set to the dedicated card instead of the iGPU (in this case, the first slot)

Offline

#19 2019-06-17 14:42:44

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

firmware settings appear to be as they should be, the bios update has solved the DMAR issue and the libgl error in
DRI_PRIME=1 glxinfo -B.

I've looked into va-api with multiple cards and found https://askubuntu.com/questions/892726/ … ideo-cards .
That post indicates libva needs to be told which driver to use for the 2nd card.

Your processor is an Intel(R) Core(TM) i5-4670K from the haswell family .
https://github.com/intel/media-driver/ doesn't list haswell as supported, so you have to use libva-intel-driver instead.

Install libva-intel-driver (if not present) , then run

$ DRI_PRIME=1 LIBVA_DRIVER_NAME=i965 vainfo

and post output.


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#20 2019-06-17 15:33:07

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

firmware settings appear to be as they should be, the bios update has solved the DMAR issue and the libgl error in
DRI_PRIME=1 glxinfo -B.

I've looked into va-api with multiple cards and found https://askubuntu.com/questions/892726/ … ideo-cards .
That post indicates libva needs to be told which driver to use for the 2nd card.

Your processor is an Intel(R) Core(TM) i5-4670K from the haswell family .
https://github.com/intel/media-driver/ doesn't list haswell as supported, so you have to use libva-intel-driver instead.

Install libva-intel-driver (if not present) , then run

$ DRI_PRIME=1 LIBVA_DRIVER_NAME=i965 vainfo

and post output.

[tedster@BlackYeti ~]$ yay -S libva-intel-driver
[sudo] password for tedster: 
warning: libva-intel-driver-2.3.0-2 is up to date -- reinstalling
resolving dependencies...
looking for conflicting packages...

Packages (1) libva-intel-driver-2.3.0-2

Total Installed Size:  7.82 MiB
Net Upgrade Size:      0.00 MiB

:: Proceed with installation? [Y/n] 
(1/1) checking keys in keyring                     [######################] 100%
(1/1) checking package integrity                   [######################] 100%
(1/1) loading package files                        [######################] 100%
(1/1) checking for file conflicts                  [######################] 100%
(1/1) checking available disk space                [######################] 100%
:: Processing package changes...
(1/1) reinstalling libva-intel-driver              [######################] 100%
:: Running post-transaction hooks...
(1/1) Arming ConditionNeedsUpdate...
[tedster@BlackYeti ~]$ DRI_PRIME=1 LIBVA_DRIVER_NAME=i965 vainfo
DRM_IOCTL_I915_GEM_APERTURE failed: Invalid argument
Assuming 131072kB available aperture size.
May lead to reduced performance or incorrect rendering.
get chip id failed: -1 [2]
param: 4, val: 0
vainfo: ../intel-vaapi-driver/src/intel_memman.c:38: intel_memman_init: Assertion `intel->bufmgr' failed.
Aborted (core dumped)

Guess it was already there; interesting failure

Offline

#21 2019-06-27 12:44:45

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

One thing that may influence results is the xorg driver.

I think xf86-video-intel is not installed and the intel card uses xorg modesetting ?

If so, try installing xf86-video-intel.


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#22 2019-06-27 14:42:25

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

One thing that may influence results is the xorg driver.

I think xf86-video-intel is not installed and the intel card uses xorg modesetting ?

If so, try installing xf86-video-intel.

You know plenty more than me (:

Installed, rebooted, no dice as far as I can tell.
Want whatever initialization output again?

Also:
Got this message while installing, worth a shot?

>>> This driver now uses DRI3 as the default Direct Rendering
    Infrastructure. You can try falling back to DRI2 if you run
    into trouble. To do so, save a file with the following 
    content as /etc/X11/xorg.conf.d/20-intel.conf :
      Section "Device"
        Identifier  "Intel Graphics"
        Driver      "intel"
        Option      "DRI" "2"             # DRI3 is now default 
        #Option      "AccelMethod"  "sna" # default
        #Option      "AccelMethod"  "uxa" # fallback
      EndSection

Thanks,
Ted

Offline

#23 2019-06-28 09:58:33

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

hmm, haven't looked at xorg logs in a while, and if xf86-video-intel makes  a difference it should show there.

post xorg logs.


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#24 2019-06-28 14:07:44

Tedster
Member
From: Edinburgh, UK
Registered: 2019-06-07
Posts: 15

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Lone_Wolf wrote:

hmm, haven't looked at xorg logs in a while, and if xf86-video-intel makes  a difference it should show there.

post xorg logs.

http://sprunge.us/OiQX9a

Offline

#25 2019-06-30 13:16:17

Lone_Wolf
Member
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,868

Re: Optimising OBS VA-API usage with AMD and Intel GPUs

Nothing weird in that log except the last 2 lines, and those appear to be after a considerable amount of time has passed.

DRI3 appears to be working fine, so I doubt switcihng to dri2 does anything.
Let's try to figure out if va-api for intel is working at all.

Change "Initial Display Output" to use the iGPU .
boot, start X and post

vainfo

output .


Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

Board footer

Powered by FluxBB