Open-Source Doom 3 Running On Wayland

[9a3eedi]

Maybe it's just me, but the framerate in that video seems pretty low for a game that's supposed to be 8 years old now. (it's not THAT low, but being a 8 year old game I would expect at least a constant 60 fps or higher)

I wonder what sort of hardware the guy is running, and what's causing the low framerate...

And it would be interesting to see the difference in framerate/performance between X11 and Wayland, that way we can know for sure if Wayland really is all what the developers seem to say it is.. a "simpler" and "faster" display protocol than X.

[Hamish Wilson]

This is Doom 3 running on the free software drivers (as those are the only ones compatabile with Wayland). Since they still do not have the same performance as the proprieraey blobs of course the framerate is not going to be the highest. I do not understand the confusion.

[blackiwid]

ok libreoffice blender and co are imporatant but frist there was only talked about firefox, I could not care less about firefox ^^

Today standard is webkit its faster so go for it ^^. just use chromium... if you want a mainstream browser with many plugins... if you dont care about plugins epiphany could be an alternative soon when they ported it to the newest webit version... Who uses freely firefox today? Some people who are ultra-conservative maybe they also even hate gnome-shell ^^

[asdx]

Just remember, that "frozen" does not mean "complete". Wayland 1.0 will lack in the protocol for anyone trying to support a full DE (or games, like you saw with that Doom 3 video). 1.0 only means, that the protocol existing so far is not going to be broken.

Also, 1.0 does not mean it can become a default windowing system in desktop distros. It means, that toolkit, DE, and application developers can now start serious work on supporting Wayland without playing continuous catch-up with protocol changes.

It is all exciting, yes, but distro migration to Wayland will be a long path. And when it's done right, you probably can't even see the difference, unless you know what to look for.

I see. Thanks for explaining.

[9a3eedi]

This is Doom 3 running on the free software drivers (as those are the only ones compatabile with Wayland). Since they still do not have the same performance as the proprieraey blobs of course the framerate is not going to be the highest. I do not understand the confusion.

I see.. I realize that opensource drivers are lower in performance compared to the blobs, but I would think they would at least be able to handle Doom 3 since it's an old game now.

But then again, it could be a weak graphics card too..

[dhewg]

That video wasn't about performance, its about the port running on wayland.
It was a debug build, weston was running on x11, and the capture process itself adds a significant overhead. So the on screen fps display you see in the video doesn't reflects the performance of the port itself or the open source drivers.

[Hamish Wilson]

What frames were you getting without the capture then? Because those are about the frames I do get when running it on my machine, although you may have a better rig.

Also, I am curious, which utility did you use to make the capture?

EDIT: Never mind, I read the article again. "The video was done using Wayland's nice video capturing support."

[crazycheese]

Maybe it's just me, but the framerate in that video seems pretty low for a game that's supposed to be 8 years old now. (it's not THAT low, but being a 8 year old game I would expect at least a constant 60 fps or higher)

I wonder what sort of hardware the guy is running, and what's causing the low framerate...

And it would be interesting to see the difference in framerate/performance between X11 and Wayland, that way we can know for sure if Wayland really is all what the developers seem to say it is.. a "simpler" and "faster" display protocol than X.

What hardware were you using for this?

Radeon 6850 and i5-3570K at stock clocks.

(charlimit)

[9a3eedi]

That video wasn't about performance, its about the port running on wayland.
It was a debug build, weston was running on x11, and the capture process itself adds a significant overhead. So the on screen fps display you see in the video doesn't reflects the performance of the port itself or the open source drivers.

I realize it wasn't about performance, but I was just wondering why the framerate was lower than I expected

Your explaination makes more sense now, thanks.. since it's a debug build and with all that overhead, it's no wonder the framerate is somewhat low.

[dhewg]

What frames were you getting without the capture then?

With X11, Radeon 6850 and i5-3570K, SwapbuffersWait off, vblank_mode=0, ultra quality settings @640x480 windowed, 64bit release build (vanilla is 32bit only) I ran 'timeDemo demo1.demo' 4 times, dropped the first result, and I get:

vanilla			122	123	120
release sdl1		119	121	119
release sdl2		119	120	117

I currently cannot run weston on drm because of https://bugs.freedesktop.org/show_bug.cgi?id=52267. So only weston on X11 for me atm, which doesn't make much sense to benchmark (For the record, its ~95fps).

Once that bug is fixed I can post some numbers...