Assassinator Wrote:

feinicks Wrote:There are hardly any good GPU Based encoders due to the fact that it is a new field to explore. Akin to the scenario: Have water, will fish. But the fish will only come after sometime.

The current encoders don't work well with too many threads, and are also quite heavily dependent on current CPU features such as the SSEx and other instruction sets, which the GPU doesn't have, to speed up the process.

Not denying that current gen encoders, or for that matter any parallel process application, are too CPU dependent. However, my argument is based on the fact that, in the future, application design may change from CPU to GPU to harness the more free and efficient power of the latter. CUDA is a an example of GPU-based instruction set. And already, the importance of CUDA has been recognised. Give it time, and you may have GPUs comparing with CPUs.

Assassinator Wrote:

feinicks Wrote:GPU are very powerful, indeed. There is that ongoing debate about whether in the future, the GPU will replace the CPU as the complete processing unit or not. Wee have a thread here about the same.

The GPUs strength comes in parallel processing. Like that previous picture V1 posted, processing 240 threads simultaneously. (Likewise, the CPU is good at serial processing)

But some stuff simply cannot be processed parallel, and so I don't think the GPU will ever completely replace the CPU. At least not any time soon.

As I quoted above, I am more interested in the future of the GPUs. CPU are reaching a set limit on how smaller and powerful they can get, without compromising on hardware estate. There is a finite value as how small and how many cores you can fit onto a board, without increasing its size (and hence cost) and, more importantly, heat generation. Ever since the advent of discreet graphics cards, GPUs have been spared this dilemma. This is quite evident in the ever increasing sizes of the cards, to allow more efficient channeling of information and dissipation of heat.

In the future, wee might even have developed a hybrid CPU-GPU system. I'm sure that this will be the goal due to the simple fact that though some basic level instruction have to be serial processed, parallel processes have more efficiency and in general greater flexibility and power.

GPUs are far too young as of now and too uni-directional to seriously threaten, but forget not that if properly constructed, a super computer comprising of only GPU core is as potent as its CPU counterpart.