Half-Life 2 Benchmarks On Linux Are Imminent

[phoronix]

Phoronix: Half-Life 2 Benchmarks On Linux Are Imminent

Pushed publicly yesterday was the test profile to run benchmarks of the popular Half-Life 2 game under Linux. As a result, coming out soon will be benchmarks of Half-Life 2 on Linux with an assortment of graphics cards and drivers...

http://www.phoronix.com/vr.php?view=MTA3MDU

[crazycheese]

As feature of PTS, its a good one!


And... lets hope AMD improves their fglrx for wine, or there won?t be much tests.

And WINE is NOT a correct way for linux applications.

[M1kkko]

I'm surprised how this article didn't have any word on when we should expect to see the native Linux version coming out.

[allquixotic]

I'm surprised how this article didn't have any word on when we should expect to see the native Linux version coming out.

"Soon", in Michael/Valve time -- it'll be released alongside Steam on Linux, Ubuntu on Wayland by default, the open source Unigine engine, UVD support for Radeon, OpenCL on Gallium3d, and the native Unreal Tournament 3 Linux port.

Or not.

[Nobu]

Ubuntu on Wayland by default

I don't see why you'd think this one is so far fetched...sure, it may not happen within the next year or two, but I'm sure it'll happen eventually--at the very least on Ubuntu.

[allquixotic]

I don't see why you'd think this one is so far fetched...sure, it may not happen within the next year or two, but I'm sure it'll happen eventually--at the very least on Ubuntu.

The biggest problem is that the proprietary drivers don't seem to want to support Wayland, and probably won't support it until years after it's been in one of the AAA distros. As long as mesa is stuck at OpenGL 2.1 (and slow, and not supporting the latest hardware) for everyone but Intel chips, any distro shipping Wayland by default is going to fear user backlash. With fear of user backlash due to the inability to use proprietary drivers, few distros would be brave enough to jump to Wayland.

It's a chicken and the egg problem, or something similar.

You can think of it like this: imagine that nobody wants to be on a boat without their friends. But nobody can physically see one another until they are ON the boat. Being on the boat without ALL of your friends on board would be very, very, VERY bad and you would instantly get nervous and wish you could jump ship. But since every friend in the group isn't sure whether all their other friends are going to be on the boat when they step foot on it, they're going to wait, and wait, and wait, until they get up the courage to get on the boat, or until they think that all their friends are on the boat. And they'll just have to hope the hell that all their friends are boarding at exactly the same time, or Bad Things will happen. Problem being, since every other friend is thinking the same thing about wanting to be on the boat at the same time as everyone else, no one is going to board the boat first.

In our case, it's just a matter of time until there is someone who boards the boat first; i.e., a distro that doesn't give a rat's ass whether they're the only one on the boat or not. Either they don't have any friends (haha) or they just don't care if their friends are there or not.

But that first distro will most definitely not be a popular one. The popular kids hang with all their friends.

My prediction is that, rather than having the big AAA distros adopt Wayland first, it'll first get adopted in very locked-down environments, such as smartphones, tablets, and embedded devices. From there, it may spread to specialist or niche distros such as Gentoo, and I predict that entire new distros will be created (small ones, with a small development team) with the sole purpose of deploying a usable Wayland desktop.

Only after seeing the success of these distros will any of the larger players consider flipping Wayland on by default. And we all know that the default experience is really the most important, because few people can be arsed to change such a fundamental part of the stack. I mean, shoot, there was an ENORMOUS upheaval about using Unity by default instead of Gnome 2 or Gnome Shell (depending on who you ask). People want to feel like the default is what they want, and they get upset when they don't get their way.

The other huge limiting factor of Wayland is that you now have to maintain all your apps, all your GUI toolkits, all your OpenGL functionality, all of it -- with two different backends, Wayland and X11. This causes a crapton of maintenance burden in the medium-term while we have to support BOTH X11 and Wayland, so that people can switch back and forth if they aren't satisfied with Wayland but they still want to use that distro. It also makes it harder for Mesa maintainers, because I 100% guarantee you that Wayland's use of drivers will expose bugs in drivers, and conversely, Wayland will do things in a broken way that X11 does correctly, or vice versa. So there will be bugs everywhere, with the developers beset upon by two separate fronts. It's going to be ugly.

I'm not against progress, I'm just extremely cynical that people will work together in the correct ways and at the correct times in order to painlessly effect a transition from X11 to Wayland on any distro. No matter which way you try to slice it, it will be as much of an upheaval for our microcosm (the free desktop community) as IPv6 is going to be for the world, or electric cars.

The problem that all three of these things (X11, IPv4 and gas cars) have in common is that there is a tremendous installed base or deployed force with the old/existing technology, and upgrading that force is tremendously expensive in one way or another. In the case of the technologies, it's a labor-intensive process, because equivalent functionalities and legacy programs have to be accounted for: regular users won't be happy with something new until it does what the old thing did, as well as it did it. In the case of the physical stuff (e.g. cars), it's a materials-intensive process, because constructing hundreds of millions of new cars with batteries is going to be enormously expensive. So as you see, even when the transition doesn't involve any physical hardware, as Wayland doesn't, it's still a much more difficult process than it needs to be to get a large group of humans to stop doing one thing and start doing another.

[Dukenukemx]

Finally some Wine benchmarks already.

[Larian]

In our case, it's just a matter of time until there is someone who boards the boat first; i.e., a distro that doesn't give a rat's ass whether they're the only one on the boat or not. Either they don't have any friends (haha) or they just don't care if their friends are there or not.

This may sound cheeky, but since when has Ubuntu cared about what their desktop users wanted? They've made it plain that they're going to do whatever they want, and the community can deal with it. As they have a history of introducing things which were not ready for prime-time as the default modules (PulseAudio, Unity), why is it so hard to believe they would be the most likely distro to adopt Wayland and cram it down everyone's throat?

I don't think your points are bad, I just think you're ignoring a pattern.

[Nobu]

Like I said, I'm sure it won't be default soon. But if someone's going to do it (whether or not it's the right choice at the time), it'll probably be Ubuntu.

[smitty3268]

I don't think Fedora is typically that worried about breaking proprietary drivers

They've had releases before where fglrx didn't work, the only difference is that nvidia wouldn't either.

We'll see what happens.

It's also entirely possible that binary driver support is much closer than you think. It should be fairly easy to provide, and the hardware manufacturers have never said they were against wayland, either. They simply aren't interested in even looking at it until it stabilizes - it would be counterproductive for them to try and support a constantly changing API while it's still under development.