Encoders don't like gpus.
 

I have noticed that a lot of encoders are not fully coded to handle distributing work to the gpu, even when the gpu can handle it. I am think it's an issue of lack of Knowledge and motivated creative coders. Like x264 for example would be one of the fastest encoders on the planet if it could get a team of programmers that code code for such jobs. I have talked to some x264 people, and they say those project always start and get no where. So what is the issue? Is it the complex coding required or lack of motivation from talented programs?

If I am over stepping my boundaries on this I am sorry. I am saying the hardware can handle what is thrown at it, but why can't we code for it?

General computing on the GPU is a relatively recent development. It depends on having the required drivers, which depending on graphics card may be proprietary, so it may only benefit a minority of users.

The situation probably isnt helped by nVidia having totally closed drivers on linux (yes, theres nouveau, but can that do what the closed drivers do? Not really, even with gallium 3D performance is slower with nouveau). ATI/AMD is a tiny bit better in that regard, but still, its not an idea situation. As for intel....ha, I wont give any intel video chip the 'GPU' tag.