Bug 106363 - Powerplay support for SI asics
Summary: Powerplay support for SI asics
Status: RESOLVED MOVED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/AMDgpu (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium enhancement
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-05-02 20:47 UTC by lemaitre.dulotus
Modified: 2019-11-19 08:37 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:


Attachments

Description lemaitre.dulotus 2018-05-02 20:47:38 UTC
This is a feature request for powerplay support on SI asics.
I can understand that they are not a priority but since they are (experimentally) supported by AMDgpu, I think they should have it.
Comment 1 Gregor Münch 2018-06-19 14:26:27 UTC
Im on Tahiti and I just append amdgpu.dpm=1 to my kernel config and things started working.
DPMS, Suspend to Disk all is working fine.
Comment 2 Michel Dänzer 2018-06-19 14:30:38 UTC
(In reply to Gregor Münch from comment #1)
> DPMS, Suspend to Disk all is working fine.

Neither DPMS nor suspend to disk has anything to do with PowerPlay (or DPM) though. :)
Comment 3 Alex Deucher 2018-06-19 14:42:13 UTC
dpm is enabled by default on SI parts so setting dpm=1 doesn't change anything.
Comment 4 Gregor Münch 2018-06-20 09:31:39 UTC
(In reply to Alex Deucher from comment #3)
> dpm is enabled by default on SI parts so setting dpm=1 doesn't change
> anything.

This depends if you use amdgpu instead of radeon on SI. In that case amdgpu.dpm seems to be disabled.
I tried to enable DPMS and read through a bunch of tutorials configuring Xorg but nothing did work.
Then I stumbled about a newsentry on phoronix mentioned this patch:
https://lists.freedesktop.org/archives/amd-gfx/2018-May/022771.html
There you change the behavior for CIK parts to use PP by default. The logic of this patch changes amdgpu.dpm=-1 (auto) to switch from whatever to Powerplay.


I started my Kernel with amdgpu.dpm=1 and everything suddenly worked like magic.

Since I dont what Powerplay actually does in the Linux world and a quick google search came up with nothing and modinfo amdgpu just reveils
dpm:DPM support (1 = enable, 0 = disable, -1 = auto) (int)
but nothing about Powerplay I was really under the impression that DPM also enables Powerplay.
Im sorry for the confusion.
Comment 5 Alex Deucher 2018-06-20 14:17:42 UTC
(In reply to Gregor Münch from comment #4)
> (In reply to Alex Deucher from comment #3)
> > dpm is enabled by default on SI parts so setting dpm=1 doesn't change
> > anything.
> 
> This depends if you use amdgpu instead of radeon on SI. In that case
> amdgpu.dpm seems to be disabled.

nope.  dpm is enabled by default on both radeon and amdgpu.

> I tried to enable DPMS and read through a bunch of tutorials configuring
> Xorg but nothing did work.

DPMS is not related to dpm at all.

> Then I stumbled about a newsentry on phoronix mentioned this patch:
> https://lists.freedesktop.org/archives/amd-gfx/2018-May/022771.html
> There you change the behavior for CIK parts to use PP by default. The logic
> of this patch changes amdgpu.dpm=-1 (auto) to switch from whatever to
> Powerplay.

This change applies to CIK only not SI.

> 
> 
> I started my Kernel with amdgpu.dpm=1 and everything suddenly worked like
> magic.
> 
> Since I dont what Powerplay actually does in the Linux world and a quick
> google search came up with nothing and modinfo amdgpu just reveils
> dpm:DPM support (1 = enable, 0 = disable, -1 = auto) (int)
> but nothing about Powerplay I was really under the impression that DPM also
> enables Powerplay.
> Im sorry for the confusion.

dpm and powerplay are just two sw implementations of the same hardware functionality.  CIK and VI share the same SMU IP so they can share most of the power code.  SI uses an old SMU block and can't share the code so it still uses the dpm implementation originally ported from radeon.
Comment 6 Gregor Münch 2018-06-23 11:30:20 UTC
Thanks for your explanations Alex.

Maybe useful for other:
It turned out that Steam client has probably a bug preventing screensaver getting active.

https://github.com/ValveSoftware/steam-for-linux/issues/5532

One more problem seems to be that Deepin or whatever still cant run the screensaver even after Steam was closed.

https://github.com/linuxdeepin/developer-center/issues/567

Im sorry again for occupying this bug report.
Comment 7 Martin Peres 2019-11-19 08:37:24 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/amd/issues/371.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.