Kernel

The kernel struct contains the type signature together with several (possibly optimized) kernel functions. Each of these functions may be NULL.

typedef void (* gm_c_kernel_t)(xnd_ndarray_t stack[]);
typedef void (* gm_fortran_kernel_t)(xnd_ndarray_t stack[]);
typedef void (* gm_strided_kernel_t)(xnd_ndarray_t stack[]);
typedef void (* gm_xnd_kernel_t)(xnd_t stack[]);

typedef struct {
    ndt_t *sig;

    gm_c_kernel_t C;
    gm_fortran_kernel_t Fortran;
    gm_strided_kernel_t Strided;
    gm_xnd_kernel_t Xnd;
} gm_kernel_t;

The idea is to have highly optimized kernels for contiguous C and Fortran arrays, a generic strided kernel for non-contiguous arrays and Xnd kernels for situations where variable arrays or optional values are needed.

For Numpy arrays the Xnd struct member may be NULL.

Kernel application

The algorithm for gufunc application can be seen in the Python module.

1. Get the function name and the list of xnd function arguments.

2. Get the types of the function arguments.

3. Select the kernel:

   1. Lookup the gufunc in the function table.

   2. Iterate over the type signatures.

      1. If no match is found, return an error.

      2. If a match is found, compute the return type(s) and the number of outer dimensions to be skipped.

         This stage should probably also do broadcasting, which is currently not implemented.

4. Allocate new xnd container(s) for the return values.

5. Input and output containers are pushed on a single stack. The types, which are available at any stage of the array traversal, keep track of the number of in/out args.

6. Call gm_map(), which orchestrates kernel application.

7. The actual kernel {C, Fortran, Strided, Xnd} is selected right before application (in this order of preference).

   If no kernel is found, an error is returned.

More specialized kernel signatures

What to add to {C, Fortran, Strided, Xnd}?

MKL would be an obvious choice. Another idea is to support kernels with closure-like state and constr/destructor functions for the state.