Any way to implement real time telecommunication technology such as LTE using CUDA?

[happy_jason]

This is requirement that to implement complex algorithm such as LTE coding/decoding with the power of GPU. But it has very strict realtime requirement. If CUDA is supported for some embedded OS such as RT Linux or Vxworks, it would be much easier for this purpose. Does anyone know there is alternative ways? Thanks.

[cbuchner1]

This is requirement that to implement complex algorithm such as LTE coding/decoding with the power of GPU. But it has very strict realtime requirement. If CUDA is supported for some embedded OS such as RT Linux or Vxworks, it would be much easier for this purpose. Does anyone know there is alternative ways? Thanks.

We’re doing this for simulation purposes (physical channel modeling, block error model) but not for performing any actual channel coding/decoding at bit level.

EDIT: The closest that you will ever get to realtime with CUDA is when you enable the Realtime Preemption patches in your Linux kernel.

Christian