Yixuan Qiu

Ah, personally I'd like to further polish it a little bit and add some new features before the actual release. After all it is marked as "1.0.0" and I wish...

@jschueller I think it is close. I'm going to merge [this branch](https://github.com/NLESC-JCER/spectra/tree/6_jd_base_solver) and do some cleanups. There will be another big API change though: I want to remove the `Scalar`...

Hi All, I have drafted a [migration guide](https://github.com/yixuan/spectra/blob/1.y.z/MIGRATION.md) to document the user-level changes from 0.9.0 to 1.0.0. Hope this makes the transition a little easier. I'm going to marge the...

It seems that `DEIGEN_DEFAULT_TO_ROW_MAJOR` is only for testing and debugging (https://eigen.tuxfamily.org/dox/TopicPreprocessorDirectives.html), and the Eigen documentation does not recommend using it in production code. If the input matrix is row-majored, then...

Yeah, internally Spectra relies on the column-majored layout for efficient operations (e.g. copying columns), and I was even unaware of the existence of the `EIGEN_DEFAULT_TO_ROW_MAJOR` macro, so at least in...

Thanks. I'll increase the iteration number in the next release.

I guess it would be straightforward to implement the randomized SVD, as it basically subsamples a large matrix and performs SVD on a smaller one. It can even rely on...

Thanks Aaron, it looks great to me. I'm not an expert on CMake, so I may ask @guacke @JensWehner @jschueller @shivupa to give some comments.

Sorry for the long delay -- I just didn't have time to focus on this. From the conversations above, it seems that overall this change is useful in certain scenarios,...

Currently there is no such one in the library, but it shouldn't be too hard to write a wrapper function for that. Pull requests are always welcome. :smiley: