Difference between revisions of "AMDGPU"

From ArchWiki
Jump to: navigation, search
m (title - move to #Selecting the right driver)
m (Selecting the right driver: rw)
Line 19: Line 19:
 
At the moment there is support for [https://www.x.org/wiki/RadeonFeature/ Volcanic Islands (VI)] (and newer) and experimental support for [https://www.phoronix.com/scan.php?page=news_item&px=AMD-AMDGPU-Released Sea Islands (CI)] and [https://www.phoronix.com/scan.php?page=news_item&px=AMDGPU-SI-Experimental-Code Southern Islands (SI)] cards. AMD has no plans to support pre-GCN GPUs.
 
At the moment there is support for [https://www.x.org/wiki/RadeonFeature/ Volcanic Islands (VI)] (and newer) and experimental support for [https://www.phoronix.com/scan.php?page=news_item&px=AMD-AMDGPU-Released Sea Islands (CI)] and [https://www.phoronix.com/scan.php?page=news_item&px=AMDGPU-SI-Experimental-Code Southern Islands (SI)] cards. AMD has no plans to support pre-GCN GPUs.
  
Owners of unsupported AMD/ATI video cards may use the [[ATI|Radeon open source]] or [[AMD Catalyst|AMD's proprietary]] driver instead.
+
Owners of unsupported GPUs may use the open source [[ATI|radeon]] or the [[AMD Catalyst]] driver instead.
  
 
== Installation ==
 
== Installation ==

Revision as of 11:21, 16 October 2018

amdgpu is the open source graphics driver for the latest AMD Radeon graphics cards.

Selecting the right driver

Depending on the card you have, find the right driver in Xorg#AMD. This page has instructions for AMDGPU and AMDGPU PRO.

At the moment there is support for Volcanic Islands (VI) (and newer) and experimental support for Sea Islands (CI) and Southern Islands (SI) cards. AMD has no plans to support pre-GCN GPUs.

Owners of unsupported GPUs may use the open source radeon or the AMD Catalyst driver instead.

Installation

Note: If coming from the proprietary Catalyst driver, see AMD Catalyst#Uninstallation first.

Install the mesa package, which provides the DRI driver for 3D acceleration.

Support for accelerated video decoding is provided by libva-mesa-driver and lib32-libva-mesa-driver for VA-API and mesa-vdpau and lib32-mesa-vdpau packages for VDPAU.

Enable Southern Islands (SI) and Sea Islands (CIK) support

The linux package enables AMDGPU support for cards of the Southern Islands (SI) and Sea Islands (CIK). When building or compiling a kernel, CONFIG_DRM_AMDGPU_SI=Y and/or CONFIG_DRM_AMDGPU_CIK=Y should be be set in the config.

Even when AMDGPU support for SI/CIK has been enabled by the kernel, the radeon driver may be used instead of the AMDGPU driver.

To make sure the amdgpu is loaded first use the following Mkinitcpio#MODULES array, e.g. MODULES=(amdgpu radeon).

Set required module parameters

To enable full support for SI/CIK when using the amdgpu, set the following kernel parameters to prevent the radeon module from being used [1]:

$ dmesg
[..] amdgpu 0000:01:00.0: CIK support provided by radeon.
[..] amdgpu 0000:01:00.0: Use radeon.cik_support=0 amdgpu.cik_support=1 to override.

The flags depend on the cards GCN version:

  • Southern Islands (SI): radeon.si_support=0 amdgpu.si_support=1
  • Sea Islands (CIK): radeon.cik_support=0 amdgpu.cik_support=1

AMD DC

AMD DC (display code), introduced in linux 4.15-4.17, is a new display stack that brings support for atomic mode-setting and HDMI/DP audio. For more info about AMDGPU-DC, see this article.

If you are on GCN 1.1 or newer with AMDGPU and want to use DC, set amdgpu.dc=1 as kernel parameter or module option when using KMS.

AMDGPU PRO

Warning: Arch Linux is officially not supported.
Note:
  • To use the proprietary OpenCL component without AMDGPU PRO, install opencl-amdAUR instead.
  • A downgrade of the linux (4.9) and Xorg (1.18) packages is required to use AMDGPU PRO 17.10.

AMD provides a proprietary, binary userland driver called AMDGPU PRO, which works on top of the open-source AMDGPU kernel driver. The driver provides OpenGL, OpenCL, Vulkan, VA-API and VDPAU support (although this is also supported by the open-source driver). For some workloads it provides better performance than the open-source driver (example benchmark), while for others the contrary is true (example benchmark).

See the release notes and the announcement at the Phoronix forum for more information.

A patched version of the official AMDGPU PRO driver is available as amdgpu-proAUR in AUR.

Loading

The amdgpu kernel module should load fine automatically on system boot.

If it does not happen, then:

Enable early KMS

Tip: If you have problems with the resolution, Kernel mode setting#Forcing modes and EDID may help.

Kernel mode setting (KMS) is supported by the amdgpu driver, is mandatory and enabled by default.

KMS is typically initialized after the initramfs stage. It is possible, however, to enable KMS during the initramfs stage. To do this, add the amdgpu module to the MODULES line in /etc/mkinitcpio.conf:

MODULES=(amdgpu radeon)

Now, regenerate the initramfs:

# mkinitcpio -p linux

The change takes effect at the next reboot.

Xorg configuration

Xorg will automatically load the driver and it will use your monitor's EDID to set the native resolution. Configuration is only required for tuning the driver.

If you want manual configuration, create /etc/X11/xorg.conf.d/20-amdgpu.conf, and add the following:

/etc/X11/xorg.conf.d/20-amdgpu.conf
Section "Device"
     Identifier "AMD"
     Driver "amdgpu"
 EndSection

Using this section, you can enable features and tweak the driver settings, see amdgpu(4) first before setting driver options.

Tear Free Rendering

TearFree controls tearing prevention using the hardware page flipping mechanism. If this option is set, the default value of the property is 'on' or 'off' accordingly. If this option isn't set, the default value of the property is auto, which means that TearFree is on for rotated outputs, outputs with RandR transforms applied and for RandR 1.4 slave outputs, otherwise off:

Option "TearFree" "true"

DRI level

DRI sets the maximum level of DRI to enable. Valid values are 2 for DRI2 or 3 for DRI3. The default is 3 for DRI3 if the Xorg version is >= 1.18.3, otherwise DRI2 is used:

Option "DRI" "3" 

Features

Video acceleration

See Hardware video acceleration.

Overclocking

Since Linux 4.17, it is possible to adjust clocks and voltages of the graphics card via /sys/class/drm/card0/device/pp_od_clk_voltage. It is however required to unlock access to it in sysfs by appending the boot parameter amdgpu.ppfeaturemask=0xffffffff.

After this, the range of allowed values must be increased to allow higher clocks than used by default. To allow the maximum GPU clock to be increased by e.g. up to 2%, run:

echo "2" > /sys/class/drm/card0/device/pp_sclk_od
Note: Running cat /sys/class/drm/card0/device/pp_sclk_od does always return either 1 or 0, no matter the value added to it via echo.

Unlike with previous kernel versions, this alone doesn't lead to a higher clock. The values in pp_od_clk_voltage for the pstates have to be adjusted as well. It's a good idea to get the default values as a guidance by simply reading the file. In this example, the default clock is 1196MHz and a range increase by 2% allows up to 1219MHz.

To set the GPU clock for the maximum pstate 7 on a Polaris GPU to 1209MHz and 900mV voltage, run:

# echo "s 7 1209 900" > /sys/class/drm/card0/device/pp_od_clk_voltage
Warning: Double check the entered values, as mistakes might instantly cause fatal hardware damage!

To apply, run

# echo "c" > /sys/class/drm/card0/device/pp_od_clk_voltage

To check if it worked out, read out clocks and voltage under 3D load:

# watch -n 0.5  cat /sys/kernel/debug/dri/0/amdgpu_pm_info

You can reset to the default values using this:

# echo "r" > /sys/class/drm/card0/device/pp_od_clk_voltage

To set the allowed maximum power consumption of the GPU to e.g. 50 Watts, run

# echo 50000000 > /sys/class/drm/card0/device/hwmon/hwmon0/power1_cap

If the video card bios doesn't provide a maximum value above the default setting, you can only decrease the power limit, but not increase.

Note: The above procedure was tested with a Polaris RX 560 card. There may be different behavior or bugs with different GPUs.

Enable GPU display scaling

Tango-go-next.pngThis article or section is a candidate for moving to xrandr.Tango-go-next.png

To avoid the usage of the scaler which is built in the display, and use the GPU own scaler instead, when not using the native resolution of the monitor, execute:

$ xrandr --output "<output>" --set "scaling mode" "<scaling mode>"

Possible values for "scaling mode" are: None, Full, Center, Full aspect

  • To show the available outputs and settings, execute:
$ xrandr --prop
  • To set scaling mode = Full aspect for just every available output, execute:
$ for output in $(xrandr --prop | grep -E -o -i "^[A-Z\-]+-[0-9]+"); do xrandr --output "$output" --set "scaling mode" "Full aspect"; done

Troubleshooting

Xorg or applications won't start

  • "(EE) AMDGPU(0): [DRI2] DRI2SwapBuffers: drawable has no back or front?" error after opening glxgears, can open Xorg server but OpenGL apps crash.
  • "(EE) AMDGPU(0): Given depth (32) is not supported by amdgpu driver" error, Xorg won't start.

Setting the screen's depth under Xorg to 16 or 32 will cause problems/crash. To avoid that, you should use a standard screen depth of 24 by adding this to your "screen" section:

/etc/X11/xorg.conf.d/10-screen.conf
Section "Screen"
       Identifier     "Screen"
       DefaultDepth    24
       SubSection      "Display"
               Depth   24
       EndSubSection
EndSection

Screen artifacts and frequency problem

Dynamic power management may cause screen artifacts to appear when displaying to monitors at higher frequencies (120+Hz) due to issues in the way GPU clock speeds are managed[2][3].

A workaround [4] is saving high or low in /sys/class/drm/card0/device/power_dpm_force_performance_level.

There is also a GUI solution [5] where you can manage the "power_dpm" with radeon-profile-gitAUR and radeon-profile-daemon-gitAUR.

Screen flickering

If you experience flickering [6] add amdgpu.dc=0 to your kernel parameters.

R9 390 series Poor Performance and/or Instability

If you experience issues [7] with a AMD R9 390 series graphics card, set radeon.cik_support=0 amdgpu.cik_support=1 amdgpu.dpm=1 amdgpu.dc=1 as kernel parameters to force DPM support.

Freezes with "[drm] IP block:gmc_v8_0 is hung!" kernel error

If you experience freezes and kernel crashes during a GPU intensive task with the kernel error " [drm] IP block:gmc_v8_0 is hung!" [8], a workaround is to set amggpu.vm_update_mode=3 as kernel parameters to force the GPUVM page tables update to be done using the CPU. Downsides are listed here [9].