I would like to use Intel IPP (Integrated Performance Primitive) library functions sparingly in my C++ code. I could observe that, though these functions are C-style functions, there is a logical pattern in the function names that depends on the data type.

How can we write a C++ wrapper based on templates?

A simple example will be the FIR filter where we need to use ippsFIRSR_32f for float and ippsFIRSR_64f for double data types.

Is there any way to write a very generic C++ template wrapper for all IPP calls?