Thore Kockerols

Hi, I am looking for an arm64 container generator. Would be great to have this option as well, instead of default x86_64. Many thanks, Thore

Hi, I tried running GenericSchur on GPU (via Google Colab) and unfortunately get stuck very early on. He is resorting to LAPACK Schur, which is not implemented for GPUs. Having...

implement dmperm

8

Hi, it would be nice to have [dmperm](https://github.com/DrTimothyAldenDavis/SuiteSparse/blob/master/CXSparse/Source/cs_dmperm.c) implemented as well. I coudn't find a julia implementation and am using a python workaround for now.

Use `block_gmres` for batched linear solves

4

The latest Krylov.jl version has a block-gmres version which I think can simplify/speed-up some parts of the code.

sparsity gets lost: sparse array input -> dense array output

1

it would be good to preserve sparsity. the culprit in the ForwardDiff extension is: ```julia y_and_dy = map(eachindex(IndexCartesian(), y)) do i Dual{T}(y[i], Partials(ntuple(k -> dy[k][i], Val(N)))) end ``` which always...

reexport AD.ForwardDiffBackend()

4

in order to keep dependencies lean in downstream packages I would recommend to reexport ForwardDiffBackend() ```Julia import Reexport Reexport.@reexport import AbstractDifferentiation: ForwardDiffBackend ```

nonlinear solving: x^a * b = c

2

Is there any capability to solve this kind of problem: ```Julia using SymPy @syms x a b c solve(x^a * b - c, x) # 1-element Vector{Sym}: # (c/b)^(1/a) ```...

`Not` exported but not defined

2

I am reexporting functions from `AxisKeys.jl` and `Aqua.jl` noticed that `Not` is exported but not defined. It would be nice to resolve this one way or another. In the meantime...

change defaults: @addlogprob! not included in PriorContext()

5

Would it be possible to change the default behavior for `@addlogprob!` and not have it included in the `PriorContext()` I understand the current solution is to do this: ```julia if...

eigen for non symmetric matrices

3

I cannot verify your results using numderiv for non-symmetric matrices (I can for symmetric matrices). Could you implement an error that this is not implemented yet?