Bugfix: the complex-to-complex FFTs were not supported with CUDA.

Description

The code was not working for complex-to-complex FFTs using CUDA. This merge request adds the interface for it, as well as the corresponding kernels for mesh_to_cube and cube_to_mesh operations. Some names are adapted to be more meaningful (accel instead of clfft).

This is now tested by the test on hybrid functionals.

News snippet

Add the support of complex-to-complex FFTs using CUDA.

Checklist

• I have checked that my code follows the Octopus coding standards
• I have added tests for all the new features added in this request.