Mojo function

crd2idx

crd2idx[: ImmutableOrigin, : ImmutableOrigin, : ImmutableOrigin, //, crd_t: IntTuple[$2], shape_t: IntTuple[$1], stride_t: IntTuple[$0], out_type: DType = uint64](crd: RuntimeTuple[crd_t, element_bitwidth=element_bitwidth, unsigned=unsigned], shape: RuntimeTuple[shape_t, element_bitwidth=element_bitwidth, unsigned=unsigned], stride: RuntimeTuple[stride_t, element_bitwidth=element_bitwidth, unsigned=unsigned]) -> SIMD[out_type, 1]

Converts multi-dimensional coordinates to a linear index.

This function is the inverse of idx2crd, transforming a set of coordinates into a flat index based on the provided shape and stride information. This is essential for mapping multi-dimensional tensor elements to linear memory.

Parameters:

crd_t (IntTuple[$2]): Type of the coordinates.
shape_t (IntTuple[$1]): Type of the shape.
stride_t (IntTuple[$0]): Type of the stride.
out_type (DType): The output data type for the index (default: uint64).

Args:

crd (RuntimeTuple[crd_t, element_bitwidth=element_bitwidth, unsigned=unsigned]): The coordinates to convert.
shape (RuntimeTuple[shape_t, element_bitwidth=element_bitwidth, unsigned=unsigned]): The shape of the multi-dimensional array.
stride (RuntimeTuple[stride_t, element_bitwidth=element_bitwidth, unsigned=unsigned]): The stride values for each dimension.

Returns:

A scalar value representing the linear index corresponding to the given coordinates.