Difference between revisions of "Talk:ATI"

From ArchWiki
Jump to: navigation, search
(Questionable Statements: strikethrough resolved dialogue for deletion)
m (Adding xf86-video-amdgpu: remove closed discussion)
 
(112 intermediate revisions by 23 users not shown)
Line 1: Line 1:
== Questionable Statements ==
+
== The generic modesetting driver ==
<s>"Radeons from HD 2xxx to HD 6xxx ... (for example, '''powersaving is still in a testing phase''')."
 
I do believe this is now functional, from using my HD6770 with power_profile and the featureMatrix represented on X.org.  Any objections to removal?</s> [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 18:22, 6 February 2013 (UTC)
 
: <s>Is there any offical release note /commit log about powersaving support exit testing phase ?</s> -- [[User:Fengchao|Fengchao]] ([[User talk:Fengchao|talk]]) 09:31, 7 February 2013 (UTC)
 
:: <s>Excellent point.  I think I can investigate that later today. Although the table on X.org showed most features functional across the board. This may have happened months ago. </s> [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 19:03, 7 February 2013 (UTC)
 
::: <s>I don't have time to cherry-pick through the [http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/ commit logs].  I skimmed through 2012.  However, this [http://wiki.x.org/wiki/RadeonFeature#KMS_Power_Management_Options page] shows the options for powersaving, and the table clearly shows functions are complete for all but the oldest cards.  On the flip side, where is the reference ''proving'' the original statement?</s> [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 20:57, 8 February 2013 (UTC)
 
::::<s> Then the testing phase statement could be removed. </s> --[[User:Fengchao|Fengchao]] ([[User talk:Fengchao|talk]]) 12:54, 11 February 2013 (UTC)
 
  
== KMS -- do we have to do anything at all to enable it? ==
+
We should also mention the generic modesetting driver. For my card (HD6570), this driver outperforms the ATI driver (I have tested it with glxgears, gtkperf and Unigene Valley) and some benchmarks available on the web seems to go in the same direction. The generic driver gives 2D acceleration via glamor and the 3D mesa libraries; 3D acceleration is given by the Mesa libraries. It is not clear that the modesetting driver is better under all circumstances but it is a "real" driver that is worth considering.
The KMS sections says it's on by default, should we even have to put radeon in MODULES any longer? If not, the intro to that section should probably mention that "you can skip the rest of this section if 'dmesg|grep modesetting' shows 'radeon kernel modesetting enabled'.
 
:: (rc.conf/MODULES are deprecated with systemd) Thinking about it, I don't explicitly recall having to enable it for the default ARCH kernel. The ''late start'' section may indeed be rendered redunant. Can anyone confirm this on a fresh install? [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 18:45, 6 February 2013 (UTC)
 
:: From [http://wiki.x.org/wiki/RadeonFeature#Linux_kernel_parameters this page], it is stated that the current behavior is to autodetect.  So as long as the module is available and the hardware is there, it will load without intervention. It only remains to be stated when this started -- for now we can simply say: currently. Leave a link to xorg? [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 21:06, 8 February 2013 (UTC)
 
  
=== KMS Early ===
+
{{unsigned|07:08, 6 October 2016‎|Olive}}
I don't know how others feel about this, but my preferred setup is to put <tt>options radeon modeset=1</tt> in <tt>modprobe.conf</tt> and then to include <tt>modprobe.con</tt> along with the <tt>radeon</tt> module in <tt>mkinitcpio.conf</tt>. This way KMS is enabled once for both early and late start. Also, in the event of problems, <tt>radeon.modeset=1</tt> can still be added as a kernel option to the bootloader. --[[User:piezoelectric|piezoelectric]]
 
:: The bigger question here is: How early is Early?  If I'm not mistaken: '''boot options''' come first, then '''initramfs''', then '''modprobe.conf''' once the final root filesystem is mounted.  If it is enabled at boot, any other toggles may be rendered pointless.  This may only be applicable if firmware needs to be compiled into the kernel (newer cards).  On the other hand, see parent section here. [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 18:45, 6 February 2013 (UTC)
 
  
== Enabling video acceleration ==
+
== Turn vsync off ==
  
<tt>~/.bashrc</tt> is the wrong place for setting this environment variables. This file is not consulted when starting i.e. firefox from the gnome shell. As it is a hardware dependent setting that is reasonable for every user on this box, i created a file in /etc/profile.d/ to accomplish this.  
+
driver="dri2" works both with DRI2 and DRI3. Also I would suggest changing vblank_mode to 1, because 0 is "force disable" and 1 is "disable by default", which seems more correct to me.
  
Setting this variables makes a difference when using flashplayer for youtube videos. When set, flashplayer reports accelerated rendering enabled at home with my ati hd3200 onboard graphics and rendering is faster in full screen mode.
+
{{unsigned|19:55, 12 December 2016‎|Equeim}}
  
It works now no more worse than under windows.
+
:You are completely right, this section should be edited, there's no need for "Reason: Is this still valid with DRI3? (which is default)", I will be free to remove it since it works on all drivers with DRI3 (nouveau, radeon etc.). You are also right about vblank_mode, but I've tested it just now, and it should stay 0 because some applications may refuse to disable it when it's "by default" (such as Chromium/Chrome).  
  
== Radeon vs. Catalyst ==
+
:However, there is something with mutter/clutter on Gnome-Shell/Cinnamon/Budgie etc. that makes browsers behave weird if it's set to 0, with 1, behaviour is as expected, so maybe we should add that as an note or added section? Here is bug explained: https://bugs.freedesktop.org/show_bug.cgi?id=99418
  
;Catalyst
+
:I would add notice for Gnome-Shell (mutter/clutter based WM), but I need second opinion. [[User:Lpr|Lpr]] ([[User talk:Lpr|talk]]) 22:36, 1 February 2017 (UTC)
The ATI and Catalyst pages state that Catalyst performs better for 3D. As most of you may have experienced, the Catalyst driver is clunky and very badly optimized for Linux. Gaming performance is usually worse than on Windows. Catalyst and Nvidia drivers do not uses Unix interfaces properly (Linux kernel and Xorg), a big part of it is just a bunch of hacks. See what [http://nouveau.freedesktop.org/wiki/FAQ#So_go_to_a_country_where_it.2BIBk-s_not_illegal_and_produce_specs.21 the nouveau folks think]. It's probably a matter of syncing development between different platforms, so adapting the driver as less as possible from Windows is probably more profitable. The lack of respect for interfaces has one major issue: the compatibility with Xorg and Kernels is extremely fragile. Besides, a lot of features or simply not reliable (or at least a real pain to setup) with these drivers, such as dual head / external output, custom kernels, hibernation...
 
  
It would be a great relief for the Unix communities to finally have a free, full-featured and top class 3D graphic driver.
+
== Restructuring ==
  
;Radeon
+
This is still called "ATI", after the company that has been acquired by AMD eleven years ago.
AMD did a very good thing when they decided to release the specs of their cards. FOSS radeon driver is now lightyears beyond nouveau in term of OpenGL implementation. The FOSS radeon driver seems like the ideal future of Unix graphics to me.
+
I think this is rather confusing because the article is covering the radeon DRM driver, Mesa stuff etc. and nothing really carries the name ATI still inside it, besides the old ddx driver.
  
Since version 9.0.* from late 2012 and January 2013, OpenGL implementation has moved several steps forward as you can see on the [http://www.x.org/wiki/RadeonFeature feature matrix]. In fact, I've tested several games (native or Wine) and radeon performs much better than catalyst most of the times.
+
Also, the old stuff about catalyst and fglrx could get into a separate, legacy article as imho most Arch users should not be concerned about that stuff.
  
For performance and Unix-frendliness, I suggest we should encourage Unix users to choose radeon over catalyst. Too many people are held up in their switching-to-Unix because of the terrible, infamous catalyst, even if all they need is average 3D graphics.
+
The "quick start guide" that comes after the installation guide (https://wiki.archlinux.org/index.php/General_recommendations#Display_drivers) also only refers to [[ATI]], [[Intel]] and [[NVIDIA]].
  
Since this revolution is brand new, it requires support from the community. We also need concrete data that states the progress, the support and the performance of radeon compared to catalyst.
+
I would like to restructure this whole stuff and add some more knowledge about the actual graphics stacks.
  
I suggest we create a benchmark page for that. Should we create a new page? I'll begin it below for now, but we shall move it in the future.
+
In a separate, e.g. "graphics drivers" article we could
 +
* note that there are free and proprietary stacks
 +
* explain that there isn't an actual "graphics driver" but a graphics stack that consists of DRM/kernel drivers, the libdrm and userspace drivers for 3D, Video, GPGPU and X
 +
* present the different stacks in a table, say which stack supports which hardware and link to separate articles for each: Intel, free AMD stack, AMDGPU-PRO, nouveau, Nvidia, and maybe one legacy ATI/AMD stacks.
 +
* recommend to use free stacks in general. Also recommend generic modesetting w/ glamor instead of old ddx drivers
  
--[[User:Ambrevar|Ambrevar]] ([[User talk:Ambrevar|talk]]) 12:07, 6 February 2013 (UTC)
+
Then, in each article with the more meaningful titles keep the same info for each of the stacks (tips, troubleshooting, ...).
  
:: A benchmark page could be informative, especially for newcomers that are having a difficult time making up their minds.  AMD is actually shifting support from Catalyst to radeon drivers.  They had hired two bodies to work on the radeon driver.  But then I came across this: http://www.h-online.com/open/news/item/AMD-dismisses-numerous-open-source-developers-1745131.html
 
:: I just found this: http://www.theregister.co.uk/2013/02/05/amd_open_source_hd_8000_drivers/print.html
 
:: Regardless, there has been a lot of headway in the last four years.  I recently tested some games on vanilla Wine and performance was considerable.
 
:: When I came to Arch, I was using Catalyst.  I was a pain to setup and maintain.  I was since persuaded by a Wine developer to switch to the radeon driver, as it is much easier to debug and submit patches upstream for either Wine or Radeon.  Not to mention Catalyst only ever "officially" supported Ubuntu, SUSE, and RedHat.
 
:: Drop a line to the Catalyst talk page?  [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 17:27, 6 February 2013 (UTC)
 
::: I think the message in the introduction is clear: '''If unsure, try the open source driver first, it will suit most needs and is generally less problematic'''. -- 09:37, 7 February 2013 (UTC)
 
::::Sure, but the previous paragraph clearly tells 'if you want 3D performance, go for Catalyst', which is not very true anymore. --[[User:Ambrevar|Ambrevar]] ([[User talk:Ambrevar|talk]]) 09:43, 7 February 2013 (UTC)
 
:::::I agree.  So we just need a clear statement from a benchmarking website that supports that.  Preferably something recent.
 
:::::I will not question the removal of dubious statements, so long as that point is made clear in the edit summary. [[User:T1nk3r3r|T1nk3r3r]] ([[User talk:T1nk3r3r|talk]]) 19:00, 7 February 2013 (UTC)
 
=== Benchmark ===
 
  
To be continued.
+
How do you guys think about that proposal?
 +
 
 +
[[User:Iuno|Iuno]] ([[User talk:Iuno|talk]]) 22:50, 7 April 2017 (UTC)
 +
 
 +
:* What are the "more meaningful titles" that you propose?
 +
:* There is no catalyst/fglrx info on this page.
 +
:* The "quick start guide" should point to [[Xorg#Driver_installation]], where most of your proposal is already covered. If not, feel free to prepare a draft on your userpage so that we can look at it.
 +
:-- [[User:Lahwaacz|Lahwaacz]] ([[User talk:Lahwaacz|talk]]) 07:27, 8 April 2017 (UTC)
 +
 
 +
:: Thanks for your answer. For "more meaningful titles" I'd suggest some consistency. E.g. there is [[Intel graphics]] (which seems fine on it's own), [[NVIDIA]] (the brand name that should also be fine as they build GPUs and SoCs, the latter might not be significant here), [[Nouveau]] (driver stack name). And then there is [[ATI]], as mentioned a brand name that died many years ago and the relevant article covers quite recent hardware. And of course [[AMDGPU]], the name of the driver stack that newbies might not now. [[Radeon]] also redirects to [[ATI]], as does [[Amd]]. If a user searches one term for AMD graphics, they land in [[ATI]] 99% and that's confusing imho. Sure, there is the introduction mentioning amdgpu and catalyst but should this not be on a short and clear overview page and then link to relevant article?
 +
:: I like the table at [[Xorg#Driver_installation]] very much. But here you can also see there is catalyst mentioned instead of amdgpu-pro. Also it covers OpenGL but not VA-API, VDPAU, Vulkan or OpenCL. Also, it lists ddx drivers and make them look like a requirement but they don't really offer advantages over the generic modesetting driver at all anymore. The opposite is true for some users.
 +
:: I also dont' get what the GCN table is doing there: [[Xorg#AMD]]. I added the info about amdgpu compatibility a while ago on [[ATI]] or [[AMDGPU]] and it got removed. Also, this section is not tied to X, as it applies to Wayland too.
 +
:: -- [[User:Iuno|Iuno]] ([[User talk:Iuno|talk]]) 16:29, 8 April 2017 (UTC)
 +
 
 +
:::Since you don't have anything better straight away, I think that "ATI" is a good enough title as it is sufficiently different from both [[AMDGPU]] and [[AMD Catalyst]]. There is already different driver for newer hardware, so we might as well just let the page perish following the lead of the brand. -- [[User:Lahwaacz|Lahwaacz]] ([[User talk:Lahwaacz|talk]]) 20:09, 8 April 2017 (UTC)

Latest revision as of 10:15, 11 April 2017

The generic modesetting driver

We should also mention the generic modesetting driver. For my card (HD6570), this driver outperforms the ATI driver (I have tested it with glxgears, gtkperf and Unigene Valley) and some benchmarks available on the web seems to go in the same direction. The generic driver gives 2D acceleration via glamor and the 3D mesa libraries; 3D acceleration is given by the Mesa libraries. It is not clear that the modesetting driver is better under all circumstances but it is a "real" driver that is worth considering.

—This unsigned comment is by Olive (talk) 07:08, 6 October 2016‎. Please sign your posts with ~~~~!

Turn vsync off

driver="dri2" works both with DRI2 and DRI3. Also I would suggest changing vblank_mode to 1, because 0 is "force disable" and 1 is "disable by default", which seems more correct to me.

—This unsigned comment is by Equeim (talk) 19:55, 12 December 2016‎. Please sign your posts with ~~~~!

You are completely right, this section should be edited, there's no need for "Reason: Is this still valid with DRI3? (which is default)", I will be free to remove it since it works on all drivers with DRI3 (nouveau, radeon etc.). You are also right about vblank_mode, but I've tested it just now, and it should stay 0 because some applications may refuse to disable it when it's "by default" (such as Chromium/Chrome).
However, there is something with mutter/clutter on Gnome-Shell/Cinnamon/Budgie etc. that makes browsers behave weird if it's set to 0, with 1, behaviour is as expected, so maybe we should add that as an note or added section? Here is bug explained: https://bugs.freedesktop.org/show_bug.cgi?id=99418
I would add notice for Gnome-Shell (mutter/clutter based WM), but I need second opinion. Lpr (talk) 22:36, 1 February 2017 (UTC)

Restructuring

This is still called "ATI", after the company that has been acquired by AMD eleven years ago. I think this is rather confusing because the article is covering the radeon DRM driver, Mesa stuff etc. and nothing really carries the name ATI still inside it, besides the old ddx driver.

Also, the old stuff about catalyst and fglrx could get into a separate, legacy article as imho most Arch users should not be concerned about that stuff.

The "quick start guide" that comes after the installation guide (https://wiki.archlinux.org/index.php/General_recommendations#Display_drivers) also only refers to ATI, Intel and NVIDIA.

I would like to restructure this whole stuff and add some more knowledge about the actual graphics stacks.

In a separate, e.g. "graphics drivers" article we could

  • note that there are free and proprietary stacks
  • explain that there isn't an actual "graphics driver" but a graphics stack that consists of DRM/kernel drivers, the libdrm and userspace drivers for 3D, Video, GPGPU and X
  • present the different stacks in a table, say which stack supports which hardware and link to separate articles for each: Intel, free AMD stack, AMDGPU-PRO, nouveau, Nvidia, and maybe one legacy ATI/AMD stacks.
  • recommend to use free stacks in general. Also recommend generic modesetting w/ glamor instead of old ddx drivers

Then, in each article with the more meaningful titles keep the same info for each of the stacks (tips, troubleshooting, ...).


How do you guys think about that proposal?

Iuno (talk) 22:50, 7 April 2017 (UTC)

  • What are the "more meaningful titles" that you propose?
  • There is no catalyst/fglrx info on this page.
  • The "quick start guide" should point to Xorg#Driver_installation, where most of your proposal is already covered. If not, feel free to prepare a draft on your userpage so that we can look at it.
-- Lahwaacz (talk) 07:27, 8 April 2017 (UTC)
Thanks for your answer. For "more meaningful titles" I'd suggest some consistency. E.g. there is Intel graphics (which seems fine on it's own), NVIDIA (the brand name that should also be fine as they build GPUs and SoCs, the latter might not be significant here), Nouveau (driver stack name). And then there is ATI, as mentioned a brand name that died many years ago and the relevant article covers quite recent hardware. And of course AMDGPU, the name of the driver stack that newbies might not now. Radeon also redirects to ATI, as does Amd. If a user searches one term for AMD graphics, they land in ATI 99% and that's confusing imho. Sure, there is the introduction mentioning amdgpu and catalyst but should this not be on a short and clear overview page and then link to relevant article?
I like the table at Xorg#Driver_installation very much. But here you can also see there is catalyst mentioned instead of amdgpu-pro. Also it covers OpenGL but not VA-API, VDPAU, Vulkan or OpenCL. Also, it lists ddx drivers and make them look like a requirement but they don't really offer advantages over the generic modesetting driver at all anymore. The opposite is true for some users.
I also dont' get what the GCN table is doing there: Xorg#AMD. I added the info about amdgpu compatibility a while ago on ATI or AMDGPU and it got removed. Also, this section is not tied to X, as it applies to Wayland too.
-- Iuno (talk) 16:29, 8 April 2017 (UTC)
Since you don't have anything better straight away, I think that "ATI" is a good enough title as it is sufficiently different from both AMDGPU and AMD Catalyst. There is already different driver for newer hardware, so we might as well just let the page perish following the lead of the brand. -- Lahwaacz (talk) 20:09, 8 April 2017 (UTC)