My prof thinks GPU will play a huge role in the future of computing, especially in image/video coding. Real-time encoding of video has always been an issue due to the complexity of the encoder, especially with H.264 being a much more sophisticated encoder than its predecessor. But with the advent of GPU processing, these encoder complexity will become lesser of an issue, assuming that one can achieve parallel programming at the encoder side.
I think the future of video coding must make use of parallel coding paradigm. Maybe the design of such codecs may not give you optimal performance, but it should be aware of what parallel computing can offer, and thus strike a balance between performance and complexity. After all, this is always the main tradeoff in computing, isn't it?
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment