AMD Has Massive Radeon Patch Set - Power Management!

[zanny]

Thank you AMD and thank you all the people working on this!

If my case could fit a second Radeon 7970, I would buy one

They still don't support crossfire, though.

[DeepDayze]

That's absolutely awesome. Is fan speed management also included?

As an NVIDIA blob user I almost envy AMD users.

That sounds great for sure for the AMD camp, and as I'm currently an nvidia user as well so now I'd like to see full reclocking and PM in the nouveau driver but as nvidia is not actually supporting the nouveau devs in any way this will surely take some time.

Perhaps it may be time for me to finally try out a Radeon again after my old ATI Radeon 9600SE card died 7 yrs ago

[PuckPoltergeist]

OK, cloned that Git, checked out drm-next-3.11-wip, copied over my config file (followed by make oldconfig), started the compile and this is all I get:

ERROR: "__udivdi3" [drivers/gpu/drm/radeon/radeon.ko] undefined!
ERROR: "__divdi3" [drivers/gpu/drm/radeon/radeon.ko] undefined!

What am I doing wrong.

Looks like only tested on 64bit systems. Can you paste the whole error message?

[mazumoto]

First thank a lot to all devs involved ! It's great to finally get this fundamental feature working.

works with 1 monitor , it doesnt with dual monitor (6870 here dual dvi 1080p monitors)

Can someone confirm this ? What's the status of vsync with dual monitors ?

[bobwya]

Fucking A!!

Finally my notebook, with a Radeon 4650M part, might be usable under the radeon driver (without going into thermonuclear meltdown!! ) It's amazing to see how fast this driver has been developed...
Hopefully (eventually) Catalyst for Linux can be put down (and out of my misery)...

[mr_marmalade]

Thanks & well done AMD!

I had been considering getting nVidia for my next GFX card because support for my apparently 'legacy' Radeon 4650 AGP in fglrx is a mess & has made me stay on Debian 6 rather than changing to 7, however, if AMD are able to make the opensource drivers as good as closed ones really aught to be, then I'd be happy. I'd also prefer to get a modern AMD CPU on my next build & positive opensource moves like this encourage me to do this - it'd also be good if this was a forward-looking step towards getting their future-gen hUMA stuff to work in Linux, I'd not like it if Linux was left behind in that regard.

[agd5f]

First thank a lot to all devs involved ! It's great to finally get this fundamental feature working.



Can someone confirm this ? What's the status of vsync with dual monitors ?

DPM works fine with multiple monitors.

[Marc Driftmeyer]

New kernels are available in Debian experimental and I did build everything from git a few months ago. It worked but I'd rather not do it again since it's a big time sink to maintain that kind of installation. I just think that it's time for radeonsi support to land in Debian experimental too. There hasn't been any progress on http://packages.qa.debian.org/x/xser...video-ati.html although the ubuntu branch has had support for radeonsi for quite some time.

Linux 3.9.1 [3.9.6 Kernel.org version #] is in Sid/Unstable. Runs just fine. 3.10-rc5 is in Experimental.

http://ftp.debian.org/debian/pool/main/l/linux/

After Debian 7.0 was released the updates to Sid/Unstable have been constant.

We should get 3.11 in Experimental two to three weeks after release to allow The Kernel Team to patch as needed, while moving 3.10 to Sid.

[pingufunkybeat]

Linux 3.9.1 [3.9.6 Kernel.org version #] is in Sid/Unstable. Runs just fine. 3.10-rc5 is in Experimental.

http://ftp.debian.org/debian/pool/main/l/linux/

After Debian 7.0 was released the updates to Sid/Unstable have been constant.

We should get 3.11 in Experimental two to three weeks after release to allow The Kernel Team to patch as needed, while moving 3.10 to Sid.

Mesa is very recent in Sid too, but it seems like xorg-video-ati is the problem.

Sid was stuck for a long time because Wheezy was locked, but now they are quite recent -- with most packages.

[Nille_kungen]

AMD really got things going lately, legacy catalyst instead of totally dropped support for older HD hardware (HD2000-4000) and the radeon driver advancing a lot.
I like how AMD have listen to users and made things better, Bridgeman you where right when you said "I think you will find that the changes are in line with what folks here have been requesting / suggesting".
UVD, DPM and ASPM well done AMD, only last year i was worried that things would get real bad on Linux but i has not.

The old phoronix article "2013: A Good Year For Open-Source AMD?" was right 2013 is good for Open-Source AMD.

Thanks to AMD open-source team and even the legal department (for letting this code through rater fast i think) and all the other (independent) developers that helped with the radeon driver.
How many people works with OSS at AMD now? Does Tim Writer (twriter), John Bridgman (now works with HSA i think) (bridgman), Alex Deucher (agd5f), Christian K?nig (Deathsimple), Michel D?nzer and Tom (Thomas) Stellard all work with OSS?