This is much simpler than using the C functions. As a bonus, I/O errors
throw exceptions.

fixes #18

This is closer to value_type which is used in STL containers.

Adding simple heat equation benchmark for CUDA.

Used shorter __cuda_callable__ instead of __host__ __device__ and
testing if HAVE_CUDA is defined.

Adding --fast-build optin for the build script to speed-up compilation.

Renaming getParameterType to getType.