To read all comments associated with this story, please click here.
Linux, Android, iOS, PS3, Wii, and MacOS don't have Dirext3D/DirectX. All use OpenGL in some variant.
As more people buy non Windows products, the demand for OpenGL games rises.
If I were a game developer in today's world and wanted to maximize my potential customer pool I would definitely be looking at OpenGL, but I would still have a D3D port for Xbox and Windows.
Middleware my good man; it’s not 1996 any more. Many AAA games are done using middleware platforms because they don’t want to invest the time/talent doing raw ports to the different platforms; the game has to be released across platforms at the same time. (crappy, laggy ports ho!)
Already the OpenGL/DX divide is for the most part non-existant other than for smaller titles and those developers who want/need custom/speed.
What Valve may help drive is a _focus_ on OpenGL as the primary source of effort instead of DirectX. Traditionally OpenGL games on Windows have always been second fiddle to DirectX and that trend may eventually swing the other direction due to the weight of iOS/Android/PS3 vs. PC/XBox/WinPhone; even with middleware (optimisation is everything for mobile systems)
Honestly, if it wasn’t for XBox, Microsoft would have already lost that grip.
This makes sense for a new developer who say, "I want to make games." But for the vast majority of big games out there that would be ported to Linux, these are Windows games. Most of these games use DirectX.
I agree that as more and more developers come in and decide that they want to develop games for many platforms, then OpenGL makes sense. But for the current batch of big Windows games, these target Windows and XBox (with perhaps a crappy port to other consoles) and are written in DirectX.
Even for the current Windows dev who writes Windows and XBox games (using same development environment - XNA), if they decide to write (or rewrite) their current games for OpenGL, what do they do for XBox, which is most likely their premier platform.
I think as mobile games, tablet games, and the non-microsft platform in general grows, we may reach this tipping point. I don't think we are there yet. It is possible that Windows 8 will help, but a lot depends of how successful Microsoft is with Windows Surface on tablets, convertibles, and whether or not they can resurrect the Windows Phone market.
"A secondary question is how hard it is to code OpenGL for Windows Devs."
They're actually very similar abstractions because they're designed to expose the same hardware functionality.
They're so similar I'd be inclined to think a shim layer would be fairly effective, and search reveals that one already exists, even:
http://sourceforge.net/projects/dxglwrap/
http://www.wcn.it/reactos.directx8.wrapper.opengl.jpg
It's author hasn't contributed anything since '09, but it might prove the feasibility of that approach never the less.
"The question is (and maybe I missed it) is - is OpenGL/Windows faster than DirectX/Windows"
"If DirectX is still faster than OpenGL on Windows, there won't be a mass exodus."
According to the article:
Windows OpenGL: 303FPS
Windows DirectX: 270FPS
Linux OpenGL: 315FPS.
We shouldn't extrapolate this directly to imply that linux would win a more complex benchmark. This benchmark might happen to be so fast that it exposes implementation bottlenecks on windows that wouldn't ordinarily be reached with a more challenging problem. In other words, maybe the windows version peters out beyond 200FPS due to no interest in optimising that case.
Either way I think it obviously demonstrates that linux can handle games, it's just a matter of convincing commercial publishers to show an interest. Microsoft may be providing the biggest incentive of all with Metro.
DirectX 9.0c is well known to have a overhead compared to OpenGL, this situation is reversed in Direct X 10 and 11
http://www.osnews.com/permalink?529385





Member since:
2005-07-06
I am glad to hear OpenGL/Linux is faster that Windows/DirectX. However, a lot of Devs will only code for Windows. The question is (and maybe I missed it) is - is OpenGL/Windows faster than DirectX/Windows? A secondary question is how hard it is to code OpenGL for Windows Devs.
If DirectX is still faster than OpenGL on Windows, there won't be a mass exodus. In addition, if OpenGL isn't somewhat easy for Windows Devs using Visual Studio and Microsoft libraries, they will continue using what they have.
The bottom line, is once someone can convince Windows devs to use OpenGL, a good part of the battle is over.