Microsoft Confirms DirectX 12 Is Alive and Well, Demo Coming At GDC 127
MojoKid writes "Buzz has been building for the last week that Microsoft would soon unveil the next version of DirectX at the upcoming Games Developer Conference (GDC). Microsoft has now confirmed that its discussion forums at the show won't just be to discuss updates to DX11, but that the company is putting a full court press behind DirectX 12. The company responded sharply over a year ago, when an AMD executive claimed that future versions of the API were essentially dead, but it has been over four years since DX11 debuted. To date, Microsoft has only revealed a few details of the next-generation API. Like AMD's Mantle, it will focus on giving developers "close-to-metal" GPU resource access and reducing CPU overhead. Like Mantle, the goal of DirectX 12 is to give programmers more control over performance tuning, with an eye towards better multi-threading and multi-GPU scaling. Unlike Mantle, DirectX 12 will undoubtedly support a full range of GPUs from AMD, Intel, Nvidia and Qualcomm. Qualcomm's presence is interesting. With Windows RT all but moribund, Qualcomm's interest in that market may have seemed incidental. However, the fact that the company is involved with the DX12 standard could mean that the handset and tablet developer is serious about the Windows market in the long term."
Better be for Windows 7 (Score:2, Insightful)
or devs Will not use it
Re:Why not open source it? (Score:4, Insightful)
Open development, more involvement from the community, more trust. Would be good for both Microsoft and its users.
Does it really need to be? Despite the name you will find that most OpenGL implementations aren't open source.
or... (Score:2, Insightful)
>>> However, the fact that the company is involved with the DX12 standard could mean that the handset and tablet developer is serious about the Windows market in the long term." ...Or it could mean that even though they already know Windows phone is almost certainly dead, being seen to be playing nice with Microsoft is worth the relatively small cost of 1 developer who is only actually working on this in any otherwise slack time.
Re:Why not open source it? (Score:5, Insightful)
But the OpenGL standard is very much open. Unlike DirectX.
Which goes a big way to explain why it is king everywhere except Windows PCs.
Fucking idiotic (Score:2, Insightful)
Unbelievable. Another brand new graphics API to front the graphics hardware so that developers have to completely rewrite their software yet again.
People need to get over their damn obsession with "new versions" and remember what the point of a programming API really is. It is to provide a stable and comprehensive interface for doing a task so that developers do not need to hit a moving target or relearn their entire skillset every six months. The reason OpenGL was so successful was that it did not try to capture the current state of graphics hardware. Instead, it captured the essential aspects of a 3D graphics pipeline and created an API for that, and let the drivers handle mapping features to hardware. With this model, OpenGL 1.x was stable for decades. You could run the exact same software on a pure-software implementation on your dinky home PC that you could on the pure-hardware big-iron SGIs. The only difference was in how fast it ran, there was no changing how you interfaced with the hardware, there was no learning an entirely new way of doing things.
Fundamentally a 3D pipeline involves vertex transformation, rasterization, and pixel coloring. That's it. OpenGL 1.x captured this beautifully with its interface for specifying vertices for primitives (triangles), setting up the matrix transformations, and specifying the coloring and interpolation modes. It was simple and effective. Hardware manufacturers came out with dozens of generations of new hardware, and the developers only had to care that "things got faster" and how many more polygons they could afford to push per frame. Then came the programmable hardware, and shaders. OpenGL responded by replacing the matrix transformations with a vertex program, and the pixel coloring with a pixel program. Simple, done. One major version update captured the whole thing with a nice high-level shader language that was intended for many generations of shader improvements.
Unfortunately, the Direct3D dipshits got their hands into the OpenGL ARB and suddenly the shader language is gimped to only represent the current generation of hardware without the simple additions required to make it future-proof. Then, surprise surprise, we need to have many more versions of OpenGL simply to fix the deficiencies that were deliberately added in the first place to sabotage the API. Now we have the same bullshit version bloat in OpenGL that we had avoided for decades until Direct3D came into the picture.
Direct3D embodies everything that is wrong with modern software development. Programming APIs need to express the theoretical capabilities of the task, not the limited expressions people currently use. And trying to make the API "closer to the bare metal"? Ahh, I get it! Driver developers think it's too much work to provide a consistent graphics API, so instead they're just going to make the game developers do all the hard work of driver development and debugging for them! Brilliant!
Fucking lazy assholes. This is why I got out of computer graphics after studying it for years in graduate school. It used to be you could approach it as a clean science, now it's devolved into bug-chasing of half-assed microcode.
Re:Why not open source it? (Score:5, Insightful)
Re:Why not open source it? (Score:5, Insightful)
So your saying that if Microsoft opened up DirectX it would be possible to use on the same devices where OpenGL is used now?
No shit sherlock. Then DirectX would be as *open* as OpenGL.
Being a open standard has everything to do with OpenGL's adoption.
Re:Why not open source it? (Score:1, Insightful)