Wayland 1.0 Stable Release Is Imminent

[Teho]

And what did Phoronix benchmarks find when running those same applications on Wayland?

According to the people who are working on Wayland the compositing causes very little overhead and it's covered in almost every presentation that I have seen. The problem with X.org compositing is also often highlighted. Benchmarking could be a bit difficult with the same software that Phoronix uses due to fact that there are changes in many parts of the system like moving from GLX to EGL that still has performance problems. If you are interested about Wayland you could of course start by watching the talk on this very article...

[89c51]

It simply sounds like you're using a blob, in which case you deserve what you get.
.

And why we shouldn't have constant user experience no matter of driver choice?? Please give me a reason that will not involve ethics of closed software or something similar.

[Rigaldo]

Yay, we got to e-peen waving! Good thing it wasn't aimed at me, but someone with a similar-sounding nick


It simply sounds like you're using a blob, in which case you deserve what you get. There is no tearing, there is no flicker here on FOSS radeon. Compositing is an extra layer, and by definition, extra layers always degrade performance.

Animations and resizing - blame your WM for those.

Let's see .. I can use FOSS radeon and overheat my laptop constantly, kill my battery life, lower my 3d performance .. And more .. Last time I tried Unigine benchmarks didn't even start under radeon. While I greatly appreciate the efforts of radeon developers, I find Catalyst river to serve me much better.
(But not really AMD's efforts as a company other than that, I don't like their strategy, I see them not having either a complete opensource or closed source driver the way they're going and fixing none... Also putting artificial limits to radeon driver, so I don't know if ever it will be at catalyst level in some sectors since AMD don't let their devs make it this way).
Sure being being opensource and everything is great, I'm even an advocate of opensource. But when it comes down to whether I want to be writing here from a laptop or a frying pan, I'll pick the first ..
From what I've gathered, AMD doesn't actually let efficient power management code to go into radeon, even though something HAD been done. So only outside of AMD can it be developed it seems .. But by whom?(Don't look at me, not a professional dev .. barely know programming)
For desktop use and if overheating is not a problem radeon is adequate, but with its own fair share of bugs, while not having some Catalyst bugs(even cairo dock has problems with Catalyst ...)

Oh, and giving up on said things was to have no tearing and flickering .. Right? I've got news for you, radeon also has these problems for me, maybe not in the exact same cases, but it does. I don't have tear free video, I can get random flickers, activating the power management it has also causes flickering all the time to the point I can't bear it and have to turn it off(and have a small fire in my room .. ok this is exaggerating), along with the rest of its problems. I'd love radeon to not have those ... But apparently it does. I even tried various configurations, but couldn't do much. And hopefully those configurations will be able to be done from a GUI soon. I've edited many .confs till now, but still find doing it from GUI better and more convenient.
But I have to admit that under radeon, my desktop experience seemed a bit smoother in regard to effects etc.

So all in all, there's reasons some people choose Catalyst over radeon. And those problems are in big part limitations of Xorg, the alien warlord that tries to run your composited desktop ..
(And it wasn't exactly meant to do what it does from the beggining ..)

[pingufunkybeat]

According to the people who are working on Wayland the compositing causes very little overhead and it's covered in almost every presentation that I have seen. The problem with X.org compositing is also often highlighted. Benchmarking could be a bit difficult with the same software that Phoronix uses due to fact that there are changes in many parts of the system like moving from GLX to EGL that still has performance problems. If you are interested about Wayland you could of course start by watching the talk on this very article...

I'll be more interested when it actually works, thx.

Like I said, interesting effort, look promising, lots of goodwill, but still far away from being any significant threat to X. I do hope that all the expectations are fulfilled and we get a better desktop, but a lot of the expectations amount to religious faith at this point.

[89c51]

I'll be more interested when it actually works, thx.

Like I said, interesting effort, look promising, lots of goodwill, but still far away from being any significant threat to X. I do hope that all the expectations are fulfilled and we get a better desktop, but a lot of the expectations amount to religious faith at this point.

FFS nobody said it is ready.

[pingufunkybeat]

Discussing here is really a waste of time.

I started by saying it's a promising project that will take years to be ready so we can see how well it lives up to the expectations, and you come here to tell me that it's not ready.

Here people are suggesting 30% performance improvement for composited environments, with no benchmarks whatsoever, and when I point out that there are no benchmarks because Wayland is not ready, you show up and tell me that it's not ready.

Fantastic.

[curaga]

I get tearing with both binary blop and Nouveau drivers. Tearing is one of the most common problems with X.org regardless of what hardware you are using (try searching it from Google).

Nouveau does not support tear-free, nor do the blobs, so it's not surprising you saw tearing. It's a driver feature to prevent it (and different from vsync, note).

So you are correct that Wayland would help there, because it forces a compositor on you. A compositor is not something you can't install on X.

X.org window managers are stuck with the problems of X.org. I haven't seen completely smooth animations (or resizing) with any window manager and to what I have understood Wayland is a big improvement on that area. The current situation where data is passed from process to process and back is bound to cause these problems.

No offense, but this sounds like expecting a magic cure.

@Rigaldo
Indeed, everyone needs to pick their poison. I used it simply as an example that it's quite possible not to tear under X.

[Rigaldo]

@pingufunkybeat Wayland is not a threat to X, there's no competition. X devs welcome and help Wayland, they're different with different purposes leading to their creation. Wayland is designed by default to have less overhead since it's designed for what it will be used in desktops. Sure, there's no benchmarks and it's not ready(no 1.0 release I can see), but I'm confident it will soon be. Aside from not being yet supported by proprietary drivers and some polish, I think there's not many significant features left. Weston lacks features, but it doesn't matter since it's more like an example compositor, other compositors/wms can implement them for themselves and likely will have to do so anyway.
Not saying it will be perfect from launch, but sometimes I think I see some "religious faith" against Wayland(or in favor of X.org? not sure .. ).
It's gonna come anyways, it's coming for a good purpose and reason and you'll still be able to use X if you want, so just welcome it, the sooner it comes the sooner it becomes "ready".

@curaga But the thing is X.org design seems to make difficult and inefficient to implement tear free. Wayland will be tear free by default, it was one of the first aims in its development and I'm sure this is good. Again I can't prove it by showing it to you, but generally Wayland is designed with such problems in account from the start of its development and conception from people who know X well enough, so it's natural to expect it will do good. And I haven't found any X driver yet able to say it's truly tear free after how much time and how many years of X.org?

[89c51]

Discussing here is really a waste of time.

I started by saying it's a promising project that will take years to be ready so we can see how well it lives up to the expectations, and you come here to tell me that it's not ready.

Here people are suggesting 30% performance improvement for composited environments, with no benchmarks whatsoever, and when I point out that there are no benchmarks because Wayland is not ready, you show up and tell me that it's not ready.

Fantastic.

Being under heavy development -and nearing a stage that DEs can built uppon- and being a research project as you said in an earlier post its worlds apart.

[pingufunkybeat]

Some of you are treating the "research project" label as some sort of grave insult, and no such insult was intended.

IMHO, it is a research project with potential to become a new standard in the future. Either in its current specification, or as a new one which learned from it.