Austen

This implementation isn't ideal but should be better in many sparse by dense products. In certain cases like when the right hand side is a dense vector this is actually...

Hello @vbarrielle @mulimoen --- Thank you both for all the work you have put into this crate, it's been a pleasure to work with for my undergraduate research and I...

Ok I'll take a look at that! Looks like a good issue to start studying the internals. Would it be worth it to bring in `criterion` as a bench dependency...

Hi @StefanUlbrich, let me know if you have any questions/suggestions as I am just starting to get involved with this crate

@tomtung Is there any reason other than the potential efficiency improvement of using `mul_add`? It seems from your benchmarks that the improvement isn't present on your machine and I'm getting...

So I ran my own bench with `criterion` on a single csr by dense matrix product with 6 variations. The first 3 are heap access, `clone()`, and `Copy`; all with...

I added benches that are exactly the same as yours and am getting the same result as with my FEM matrix. Hardly any difference between heap access, clones, and copies...

> Thanks, this looks great! > > I only have a few minor comments related to documentation and naming. For example, maybe we should rename `CsrInnerIter` to `CsrRowIter` and `CscInnerIter`...

I like the long term plan if you think it's worth it. Personally I use this pattern a lot and I've even seen it throughout the crate so it is...

What you want is the Schur decomposition yes? https://docs.rs/nalgebra/latest/nalgebra/base/struct.Matrix.html#method.schur If this doesn't work and you need the classical eigen decomp then the Lapack version of the crate has: https://docs.rs/nalgebra-lapack/latest/nalgebra_lapack/struct.Eigen.html