Readers, Multi-GPU, Profiling...

For extending readers in CNTK, please follow :cntkwiki:these instructions <BrainScript-and-Python---Understanding-and-Extending-Readers>

For multi-GPU and multi-machine usage, please follow :cntkwiki:these <Multiple-GPUs-and-machines>.

For profiling, please refer to :cntkwiki:these <BrainScript-and-Python-Performance-Profiler>.