Valve's L4D2 Is Faster On Linux Than Windows

Using a NVIDIA GeForce GTX 680 graphics card with an Intel Core i7 3930K processor, Windows 7 SP1 was running Left 4 Dead 2 with the Direct3D renderer at 270 FPS while under Linux with OpenGL they are now at 315 FPS! Using the OpenGL renderer on Windows isn't also quite as good with its average frame-rate at around 303 FPS.

As though it matters. And wouldn't it be interesting to see the Windows 8 numbers? And why Valve is messing around with a dated as hell engine and porting an old game to Linux? Wouldn't their time be better spend on HL2 EP3?
 
Wow... And I thought the whole work just began with Steam Linux...

Maybe Christmas 2012 Steam sale will be also be available on Steam Linux. :)

Nice to see great results.
 
Good thing too, that boost from 270 FPS to 315 FPS will make all the difference.
 
It would run on it, if it was made for it.

The importance here is on directx vs opengl. Did you not understand?

What is there to understand? Valve is porting a game, probably optimizing things along the way and that runs better compared to the years old DX version on a version of Windows soon to be outdated. Plus the Source engine is terribly dated.

I'm not slamming Linux here, its just that this means so little in the grand scheme of things that it is interesting that Valve is dedicating resources to this when a LOT more people would care about HL2 EP3. I think it would certainly be more profitable for Valve.
 
Well you know how Valve is... they don't work on anything that the people there don't want to work on. So if none of the devs have interest in making HL2 EP3... well there ya go. :)

But what this does show (which we already knew) is that OpenGL can likely hang with Direct3D in terms of performance. So maybe it doesn't have to be completely shunned by devs.
 
I think it would certainly be more profitable for Valve.

A very short-sighted attitude, which I'm glad Valve doesn't seem to share. First of all, it's not like it's hundreds of people working on this, and secondly, investing in QUALITY is never wrong.

Also, there are too many of these threads.
 
A very short-sighted attitude, which I'm glad Valve doesn't seem to share. First of all, it's not like it's hundreds of people working on this, and secondly, investing in QUALITY is never wrong.

Also, there are too many of these threads.

So "investing" in a platform virtually nobody uses is a better use of time?

DNF took forever to come out. I guess that was a quality game, huh?
 
But what this does show (which we already knew) is that OpenGL can likely hang with Direct3D in terms of performance. So maybe it doesn't have to be completely shunned by devs.

Correct me if I'm wrong, but why is this even relevant? If Valve were comparing a modern version of D3D to OGL it would be notable. But even since L4D2 came out there were advances made that make D3D even more efficient than it already was.

Its like saying running a benchmark on XP produces lower results than W7, or iOS2 is slower than iOS5. That may be the case, but you're comparing a modern technology to an old one.
 
If I'm not mistaken didn't Gabe (valve dude) say that windows 8 was complete shit? Knowing that them preparing a non windows market place seems like common sense.


Seems Valve is not focused on quick easy money and actually prepares for different possible outcomes....... Benefits of being a private company.


Though even if windows 8 is a complete flop nobody is moving to linux. Windows could probably even fuck up windows 9 and still nobody is moving to linux.......Unless they stop selling Windows 7.....
 
If I'm not mistaken didn't Gabe (valve dude) say that windows 8 was complete shit?

That quote was taken out of context. If I'm understanding correctly he's speaking about the fact that manufacturers will have to start making tablets which are more expensive to produce than the laptops and desktops they are currently selling, and then sell them at around the same price. This wouldn't be a problem if it weren't for the fact that the PC manufacturers already sell product with thin margins.

Of course, the media will sensationalize any quote they can. It gets page views, commercials watched, and papers sold.

Gabe Newell said:
“I think Windows 8 is a catastrophe for everyone in the PC space. I think we’ll lose some of the top-tier PC/OEMs, who will exit the market. I think margins will be destroyed for a bunch of people. If that’s true, then it will be good to have alternatives to hedge against that eventuality.

That says nothing of "Windows 8 is complete shit." That's saying MS's hardware partners are finally going to have to compete on more than prices, and many of them will not be able to.
 
What is there to understand? Valve is porting a game, probably optimizing things along the way and that runs better compared to the years old DX version on a version of Windows soon to be outdated. Plus the Source engine is terribly dated.

Valve is not porting just a single game. Valve is porting their entire engine - that includes a whole bunch of games.

I'm not slamming Linux here, its just that this means so little in the grand scheme of things that it is interesting that Valve is dedicating resources to this when a LOT more people would care about HL2 EP3. I think it would certainly be more profitable for Valve.

Umm... No. Making the engine cross-platform will mean that they will be able to easily distribute their games on:
PC - Mac - Linux
Xbox - PS3 - Wii?

Don't you see $$$ here? Adding another platform will never be a problem for them, as they already have so much cross-compatibility going on.

It means A LOT in the grand scheme of things.
 
Mad that Valve is porting an 8 year old engine and old games to Linux? If you say so phide.;)

So you say it's old.

Unreal Engine is old. In fact, it's probably 13-14 years old at this point. And yet most console and PC games nowadays use it. And, well, the newest revision looks pretty sick, no doubt.

It's a similar story with Source engine. The source engine of today isn't the source engine of 2004. They use the same name - yes - but I think that's where the similarities end.

I think a decent example of that is Half Life 2 Lost Coast. It improved on the Source engine by adding / improving on (according to Wikipedia) HDR in the original game. Just an example.
 
I think a decent example of that is Half Life 2 Lost Coast. It improved on the Source engine by adding / improving on (according to Wikipedia) HDR in the original game. Just an example.
Lost Coast is a good example, actually, because the changes made to the renderer run quite a bit deeper than what they added (the things we can see by running through it). Adding one feature required a major reworking of the renderer. It's been reworked again since them, perhaps a couple times. It's not unlikely at all that more than 75% of the renderer, as it existed in 2004, has been completely discarded and replaced with different code.

The API hasn't changed, but that doesn't mean nothing else has.
 
Steam is 9 years old, if age of software is a serious consideration. :confused:

Hard to see how it's not a consideration, especially with software. The Source Engine is pulling around 300 FPS on a single top end card in L4D2, that's not exactly a sign of top-line game engine.

I don't have anything against Valve doing this, but if a company were coming out with a port of an engine to Windows that was past it's prime, how many Windows users would be going wild about it.

But sure, this is great for Linux, a lot more companies will need to do the same thing and with their newest efforts, not something from years ago to push gaming on Linux beyond the Linux crowd. That's all I'm saying and it's perfectly reasonable. How many average users are going to say "WOW! Linux now supports games I played years ago! I'm switching to Linux!"
 
But sure, this is great for Linux, a lot more companies will need to do the same thing and with their newest efforts, not something from years ago to push gaming on Linux beyond the Linux crowd. That's all I'm saying and it's perfectly reasonable. How many average users are going to say "WOW! Linux now supports games I played years ago! I'm switching to Linux!"

Of course not.

But the point is to get the ball rolling.

It's all kind of moot anyway since the real effort here is probably about laying the groundwork for a console or something like that.
 
Steam on Linux would make installation of games a lot easier...
 
So many people seem to be completely missing the point of this.

Why is Valve porting such an old game no one cares about?

It could be L4D2, TF2, HL2, Portal 2, doesn't matter what it is. The entire idea is just a stepping stone in porting the Source engine to Linux. They have the experience going to OpenGL thanks to the Mac port, but Linux is another beast. This is a baby step, a small part of the bigger picture. Its L4D2 because its a game people can play, one thats not constantly changing (like TF2 and its constant updates) and one thats multiplayer so people can play with their friends.

Source is so old, who cares.

The Source engine we have today is nearly nothing like the Source engine that debutted with HL2. Much like the Steam we have today is nearly nothing like the Steam that first debuted and everyone dispised. It is still slightly dated, but its constant upgrading have kept it relevent and played by millions all over the planet.

How dare they do this, they should be working on HL3!

Valve is more then the eleven people currently working on this project. They are, in fact, a good deal bigger then that. Valve is doing a lot of things at once, this is just one of them. It bogggles me how people think a company like Valve has to focus on just one thing, because if you've read ANYTHING about how Valve works, thats simply not how things go. Even if you did have everyone focusing on HL3, you'd have a lot of redundant, useless work....then those people would need to find something else to do. Like work on the Linux port. As I said in the beginning, the L4D2 is a stepping stone, and in the end WILL benefit HL3 in its own Linux port.

Who cares about Steam on Linux?

A lot of people, namely the vast majority of current Linux users, and then people like me, who are only on Windows because of Steam. Everything else I use works either natively or is easily emulated on Linux. Ubuntu looks really impressive these days, and more and more (to me) is looking like a viable alternative to Windows 8 (which is personally don't like the look of, and the rumors of some of the close-endedness they may be planning).

In my opinion, this post isn't a "LOL OPENGL IS BETTER SEE LIKE 10 MORE FRAMES!!1", Valve isn't that petty from what I've observed. This is a "Its possible, its working, and better then expected."

However, people will read into what they want. Can't really help that. I can say though, if TF2 and CSGO get native Linux ports, you can bet your ass I'll be looking ot Ubuntu as a permanent OS. With Gabe Newell, Mike Morhaime, John Carmack (kind of), Notch and others high and respected in the gaming industry look at Windows 8 with a grimmace, I can hope that maybe a Linux shift could be in the making.
 
Many people give shits about this. I, for one, cannot wait until I can fire up an install of Ubuntu, install steam and play most if not all my games. I really hate windows and the direction windows is going.
 
So many people seem to be completely missing the point of this.



It could be L4D2, TF2, HL2, Portal 2, doesn't matter what it is. The entire idea is just a stepping stone in porting the Source engine to Linux. They have the experience going to OpenGL thanks to the Mac port, but Linux is another beast. This is a baby step, a small part of the bigger picture. Its L4D2 because its a game people can play, one thats not constantly changing (like TF2 and its constant updates) and one thats multiplayer so people can play with their friends.
Taking slow baby steps means shit when the rest of the world is running away.




Valve is more then the eleven people currently working on this project. They are, in fact, a good deal bigger then that. Valve is doing a lot of things at once, this is just one of them. It bogggles me how people think a company like Valve has to focus on just one thing, because if you've read ANYTHING about how Valve works, thats simply not how things go. Even if you did have everyone focusing on HL3, you'd have a lot of redundant, useless work....then those people would need to find something else to do. Like work on the Linux port. As I said in the beginning, the L4D2 is a stepping stone, and in the end WILL benefit HL3 in its own Linux port.
HL3 doesn't even exist, and already you are predicting HL3 on Linux. Thats called HOPE, not reality.

A lot of people
A very small percentage of total PC users.

, namely the vast majority of current Linux users, and then people like me, who are only on Windows because of Steam. Everything else I use works either natively or is easily emulated on Linux. Ubuntu looks really impressive these days, and more and more (to me) is looking like a viable alternative to Windows 8 (which is personally don't like the look of, and the rumors of some of the close-endedness they may be planning).
Windows 8 likely being a turd is not relevant to the discussion, especially since its all speculation at this point.

In my opinion, this post isn't a "LOL OPENGL IS BETTER SEE LIKE 10 MORE FRAMES!!1", Valve isn't that petty from what I've observed. This is a "Its possible, its working, and better then expected."

However, people will read into what they want. Can't really help that. I can say though, if TF2 and CSGO get native Linux ports, you can bet your ass I'll be looking ot Ubuntu as a permanent OS. With Gabe Newell, Mike Morhaime, John Carmack (kind of), Notch and others high and respected in the gaming industry look at Windows 8 with a grimmace, I can hope that maybe a Linux shift could be in the making.
Yes, I cant wait for really old games to be ported to an OS I don't use. YAY Hats on linux!

Listen, If valve wants to port Steam and some of their games to Linux, great. What about the 99% of the other titles that won't be ported? You're still rebooting into Windows, hence not a big win.
 
CHAoS_NiNJA, I agree with you. Some really decent points were made. Kudos.

Even though most people don't even use Linux, I'd bet that almost every single one of them has heard of Linux one way or another, so it's not an "underground" platform. Of course I'm talking about people that care about technology. Also, a very small percentage of PC users have access to all the marvelous and latest hardware that are daily discussed here, but that doesn't mean they aren't relevant.

I wouldn't hope for a Linux shift in the near future, but grabbing attention from hardware and software vendors is always a good thing, and I think this is a very good way of doing so.

We should expect comments like "hur dur... no one cares", but by spending time posting in this thread, I suspect they've proved themselves wrong.
 
IMO, the porting of Source games is primarily to support the rollout of Steam. I expect it should resembling the OSX rollout of Steam very closely.
Just as you wouldn't setup a website without content prepared or a forum without some plan for generating discussions, so is that you don't open an E-commerce store without any products to sell.

It demonstrates to indies and and companies that it's a viable platform that is possible to support. I know that I have chosen titles on Steam specifically because they support both Windows and OSX. So it does influence where dollars are going. I know I was absolutely tickled the first time I fired up Portal 2 on my MBA. Finally, an end of the "OSX has no games" complaint. Expansion of that into linux and seeding it with your own games is not a dummy idea.
 
Looks more like a DirectX 9 limitation creating a bottleneck than anything else. Case in point, when they ran it in OpenGl on Windows as well, the performance gap narrowed to almost nothing.

Valve said in their blog post that they're looking to bring similar optimization to the Windows/DirectX version of the graphics engine. Wouldn't be surprised to see it outperform Linux/OpenGL again after they get done with it. A move to a DirectX 11 renderer would remove a lot of the overhead that they're dealing with on DirectX 9.
 
Back
Top