This is the strided equivalent of https://docs.pytorch.org/docs/stable/generated/torch.vander.html and https://numpy.org/doc/stable/reference/generated/numpy.vander.html.

API signature:

function dvander( M, N, order, x, strideX, out, strideOut1, strideOut2 )

where

• M is the number of elements in x and number of rows in out.
• N is the number of columns in out.
• order specifies whether the order of powers increase from left to right or decrease.
• x: input strided array.
• strideX: stride length.
• out: output strided array corresponding to a matrix.
• strideOut1: stride length of first dimension.
• strideOut2: stride length of second dimension.

Depending on whether out is row- or column-major, values should be set in a cache-optimal way.

This will eventually enable blas/ext/vander for creating Vandermonde matrices.