Oliver Schulz
Oliver Schulz
> A project I'm working on requires me to iteratively update a matrix X'X as I tack on columns to X in batches. You mean `X' * X`? Maybe there's...
> > You mean `X' * X`? Maybe there's also another way to approach this - what shape does `X` have, and in which direction(s) does it grow in your...
> heavy_computation asks for P::AbstractMatrix Ah, yes, that's a problem when using LinearMaps sometimes, I suggested to make `LinearMap` a subtype of `AbstractMatrix` a few days ago: JuliaLinearAlgebra/LinearMaps.jl#180
> It's a complex optimization that involves factorizations of P so I don't think the linear maps approach works here Yes, if there's explicit factorization and so on, LinearMaps won't...
My you *do* have a tricky use case there! :-)
Also ``` julia> using ElasticArrays, CUDA julia> CUDA.allowscalar(false) julia> cu(rand(2,2)) * cu(ElasticArray(rand(2,2))) ERROR: Scalar indexing is disallowed. ```
I think fixing this would require custom methods for (at least) `LinearAlgebra.mul!`, `Broadcast.copy(bc::Broadcasted{
@vlepori looking at this again, maybe the primary way to do it should be `deleteat!(A, ((), i))`, but we should support `deleteat!(A, (), i)` as well. It matches how handle...
Oh, and could you add a test or two?
Ok, let's run Ci on it then. :-)