Chrome/Chromium's Ozone X11 Code Now Fully Enabled, Old Legacy X11 Code To Be Removed

[Azrael5]

They already did. Their Tegra drivers have been using it for awhile and they already have the desktop GBM driver mostly finished. An Nvidia dev even said that Sway runs on it without any modifications.

Release date?

[kpedersen]

So, you think you are the center of the universe and other people do not exist. Gotcha.

What a silly thing to say. I'm not even sure it relates to any of my posts.

Just use Xorg like everyone else does, deal with software decoding. Get on with your life and stop whining is *actually* what I think.

In short Linux users are probably the only computer users who continue to "enjoy" software video decoding and encoding.

Deal with it. The rest of us manage. Not to mention us BSD users never got the Flash plugin in the first place

Of course, Xwayland is the natural successor of X11 in order to run X11 applications on Wayland.

Xwayland isn't really a successor to X11. It is more like an implementation. For example the main ones are: XQuartz, Xming, VcXsrv, Xorg, Xwayland.

What is slightly odd is that *all* of them are actually stripped down forks of Xorg. It seems that writing one from scratch is quite tricky for "modern" developers. The most recent from scratch rewrite was Xephyr (KDrive based) which is sill fairly old.

[avem]

What a silly thing to say. I'm not even sure it relates to any of my posts.

But what I *do* think is that people should should be happy with what they have and not try to break what is already working *before* they have a serious improvement.

The whole world is still very much reliant on Xorg. Whether "special" people exist or not doesn't matter. They are insignificant or "phone consumers" that are absolutely bottom priority to open-source.

Again you don't care about laptop Linux users because your laptop is always plugged in (IOW you're not using it as a laptop) and just like I said people with different needs and means don't exist for you. And you put yourself above everyone else and even go as far as to call other people's needs as "bottom priority". There are a lot more Linux laptop users who absolutely could use HW video decoding/encoding than those who don't care about it. Please continue to argue with someone else.

[avem]

Many articles in phoronix highlights how no relevant updates are provided on Xorg.

• What relevant updates do you personally need for Xorg, which are present in Wayland and without which you cannot exist?
• What's the percentage of Linux users out there that shares the same needs?
• Has X.org stopped working? Maybe it inadequately supports modern HW or even damages it? Maybe it's a pain to use and everyone suffers?

[evil_core]

Laptop users are looking at you in disbelief, confusion and horror.

Jokes aside, it looks like you totally don't understand the essence of HW video encoding/encoding. In terms of energy it can be hugely more efficient than doing the same on a generic CPU no matter the uArch.

Here's some recent data:



And Apple M1 is currently one of the most efficient and performant CPUs on this planet. The difference could be much bigger for less performant CPUs.

More info here: http://web.archive.org/web/202107071...sumption-pt-2/

My Ryzen 7 5800X under Linux/Chrome 92 consumes around 25W while playing

at 4K VP9. My 1660 Ti consumes less than 10W doing the same. 15W must be insignificant for you I guess.

It's stupid and irrelevant to post some web browser benchmarks, on such obscure and niche architecture (nobody, except Apple and few other companies is optimizing for Apple M1 silicon. It takes years for ffmpeg to get optimization for mainstream x86/x86_64 architectures new extensions)

Show us benchmarks with mpv/mplayer/ffmpeg playing local files(with different codecs: h264, h265 and av1 at least), because web browsers are so complicated nowadays, that they can have greater impact themselves on power usage, than decoding backends on x86_64)

[Myownfriend]

Release date?

I don't know. That would be internal to Nvidia. The changes to Mesa that would allow Nvidia's proprietary drivers to slot in as an alternate GBM backend are already part the last Mesa release though.

[Candy]

I've been using Linux for over two decades. In terms of using Xorg:

• Never fiddled with Xorg conf files to get 144Hz on my current display - it worked out of the box

Two decades ago there was XFree86 (and some commercial alternatives). Remembering back to these time, you'd had to fiddle the hell out of an XFree86.conf to get an XServer running the way you wanted. Mouse, Keyboard, Display, Modules, Resolution, Frequency and so on. Must be an alternate universe then.

[avem]

Two decades ago there was XFree86 (and some commercial alternatives). Remembering back to these time, you'd had to fiddle the hell out of an XFree86.conf to get an XServer running the way you wanted. Mouse, Keyboard, Display, Modules, Resolution, Frequency and so on. Must be an alternate universe then.

X.org has been autoconfiguring itself for more than a decade now. Wayland is not that old for you not to know that which indicates that either you haven't used Linux recently or you're making things up. People in this discussion are talking solely about X.org, not about its alternatives or predecessors - I'm not sure why you're bringing XFree86 to the mix.

[avem]

It's stupid and irrelevant to post some web browser benchmarks, on such obscure and niche architecture (nobody, except Apple and few other companies is optimizing for Apple M1 silicon. It takes years for ffmpeg to get optimization for mainstream x86/x86_64 architectures new extensions)

Show us benchmarks with mpv/mplayer/ffmpeg playing local files(with different codecs: h264, h265 and av1 at least), because web browsers are so complicated nowadays, that they can have greater impact themselves on power usage, than decoding backends on x86_64)

You seem not to have read my entire message which is painful to see. Maybe read it again: I've compared a modern system with a modern CPU and a relatively modern GPU. Only I was wrong about my 1660 Ti power consumption, it's a lot lower when playing AV1 4K clips. At most I see extra few watts being consumed vs. 24-35W for my CPU (under Linux it's higher than under Windows 10 for some reasons), in other words playing video with my GPU is ten times more efficient if we're talking about extra watts being consumed. The overall system power consumption is of course a lot higher but I don't have access to a modern laptop to test 4K VP9 videos. My old one can only accelerate VP9 decoding at 1080p or below.

Since we must have tons of hardcore Linux fans who are rocking the latest Intel CPUs, they could test it on their systems instead using like you said mpv which allows to switch between HW and software decoding.

[Candy]

X.org has been autoconfiguring itself for more than a decade now.

That's right. Most of that started after the modularization of XOrg has begun. But you wrote that you've been using Linux for nearly two decades now. You should've known that there used to be times, where configuring the XServer was a pain in the ass (besides the rest of the system).