OpenCL dynamic parallelism / GPU-created threads?

Question

CUDA 5 has just been released and with it the ability to generate GPU threads from another GPU stream (main?), Minimizing the callouts between the CPU and GPU that we have seen so far.

What plans exist to support threads created on the GPU in the OpenCL arena? Since I cannot afford to choose a closed standard (my user base is “every player”), I need to know when OpenCL is ready for prime time in this regard.

Answer 1

The OpenCL standard is usually the CUDA backtrack (with the exception of the device split function), and I assume that this function will be added to OpenCL in a year.

EDIT August 8, 2013. This feature was introduced in OpenCL 2.0 .