vload from array with ndims=2

[davidbp]

In the readme one can read that vload allows to load from an array

arr = Vector{Float64}(undef, 100)
xs = vload(Vec{4,Float64}, arr, 1)

But for a "matrix" this does not seem to work.

arr = Matrix{Float64}(undef, 100, 100)
xs = vload(Vec{4,Float64}, arr, 1)

Is there a way to do it?

[davidbp]

Looking at the source code of vload it seems the code

Vec(Intrinsics.maskedload(ptr, mask.data, Val(Aligned), Val(Nontemporal)))

would work even if the input data of vload was any array.

If we consider

X = Int.(reshape(1:32,4,8))

to be

4×8 Matrix{Int64}:
 1  5   9  13  17  21  25  29
 2  6  10  14  18  22  26  30
 3  7  11  15  19  23  27  31
 4  8  12  16  20  24  28  32

Could the package allow us to do the following?

xs = vload(Vec{4,Float64}, X, 1) # to load [1,2,3,4]
xs = vload(Vec{4,Float64}, X, 4) # to load [5,6,7,8]
xs = vload(Vec{4,Float64}, X, 2) # to load [2,3,4,5]

[KristofferC]

Maybe you can use vec as a workaround for now?

[davidbp]

You mean store X_as_vec = vec(X) and then work with X_as_vec ?

Yes, but I would love to be able to work with 2D arrays as mentioned above (actually it seems the backend code of vload would work with any array!).
Specially because in some cases I don't want copy the data of a matrix into a vector before calling a function.
The fact we can't do it seems to be of a limitation of the input types of vload than the code that vload ends up calling.

[KristofferC]

vec does not copy. And as I said, this would just be an immidiately workaround.