I’m wondering if there is some sample codes to demonstrate the new features for CUDA 3.0. For example:
1.C++ Class Inheritance and Template Inheritance support for increased programmer productivity
2.Concurrent Kernel Execution
3.ELF compilation (Do we still need to use extern “C” for *.cpp files to access *.cu functions and classes?)
…
Also, does CUDA 3.0 support function pointer in device code? I guess if class inheritance works, then function pointer should do. But is this true without Fermi architecture support?