Bridgman when can we expect to see products with this tech??
I understand there are benefits to GPU computing, when there are many paralelized calculations to be performed, however, this is not always the case, or there would be no reason for AMD including the UVD block in their GPUs and APUs in the first place.
What I grasp from the HSA inniciative is more like:
CPU: very good for serial general calculations
GPU: very good for paralel calculations, still getting better at being general
There are, however, other workloads, such as video transcoding for example that are not specially suitable for neither, hence intel, amd, nvidia and several arm soc vendors include a video transcoding block in their chips which is several times more efficient than either cpu, gpu of both at this specific workload.
What I would like to know is if the HSA foundation has plans to do to these blocks the same it is doing to the GPU: helping programmers use the speciallysed logic transparently.
Please correct me if I'm wrong in any of the above assumptions.
could you estimate how much time you would need to invest to get gpu based aka shaderbased accelleration of x264 720p and 1080p accelleration.
for a guy who is able to programm in C but have no experince in kernel or x or driver development?
I would maybe try it if it would have any change of completion without working on it for 6 months full-time. And bridgeman did say its all there to get it running easy.
he never explizitly said easy but if that task would cost 100.000$ manpower-costs it would be a unneeded statement because then it would be clear to anybody that it will not happen ever, if amd does it not by them self.
Christian Konig already wrote code for shader based h264. I am not sure if he released it and where.
Originally Posted by blackiwid
If you are interested better ask in the mailing list.
For the Intel demo, it's using GStreamer + libva acceleration, and rendering that directly into a GPU texture which Wayland uses to render onto the screen.
Originally Posted by e.a.i.m.a.
According to the devs, I think that was the best their hardware could do - (Sandybridge, i think, or maybe Ivy), but future hardware would allow them to use a hardware overlay to display the video instead which would save a lot of power usage compared to using the GPU texture. The problem with current hardware was that the VA decoded format was incompatible with the current hardware overlays.
The UVD unit also use the shaders without the usual drawbacks and overheads now you can imagine how good this work with the opensource drivers on a hd2900 with a broken uvd unit and or on a hd3870 with closed source only uvd unit and so one and so one.
Originally Posted by bridgman
Your marketing speech is a ridiculous attempt to sell the new DRM TrustZone system then we get the (OLD-UVD unit+TrustZone) = HSA!!!
And Intel do have "Intel Trusted Execution Technology" already because of this they can deliver the video acceleration right now with opensource drivers.
But hey the fools will get the truth when the HSA anti-consumer hardware is ready to sell then we all know how good TrustZone will work against the customers. . .
Any positive imagination is just nativity-->"Fixed that for you"
09-26-2012, 09:35 PM
Is there such a thing as a "triple facepalm" ?
Originally Posted by necro-lover