Sunil Pai

Hi, in parallel with this, I'd like to add an implementation for bicgstab (which i believe is also matrix free)! I've started working on something similar in [cupy](https://github.com/cupy/cupy/pull/3569) so I...

Excellent! If GMRES works out too, then we can even do a more general form of BiCGSTAB, which may be useful for EM problems (as suggested by [meep](https://fossies.org/linux/meep/src/bicgstab.cpp)). Not sure...

Hi @shoyer, I have a general question as I prepare this PR. What is the motivation of defining pytree functions like `_add` and `_sub` in the `_cg_solve` method? Is this...

@shoyer I added a PR for `bicgstab` but it's still WIP, was just looking for some comments before I finalize the implementation and tests but I think it's close! I...

Hi @shoyer, re: #3796 I think we should think about either a more robust testing pipeline for all matrix-free methods or just copy `scipy`'s [tests](https://github.com/scipy/scipy/tree/master/scipy/sparse/linalg/isolve/tests) and use that as our...

> A pull request that adds nanometers would be appreciated. Will get started on this, thank you! > That should be supported already. If you feel something is missing or...

Yes, here it is: https://codesandbox.io/p/sandbox/graticule-forked-tkvlnq

I managed to change some args to get closer to the desired behavior using (see modify sandbox above): ```javascript intervals: [10000, 1000, 100, 10, 1, 0.1, 0.01, 0.001, 0.0001], lonLabelFormatter:...

One other issue though @ahocevar. By default in the above sandbox (ie without any further zooming), the I expect to see the Graticule based on the `minZoom` option but I...

Bro, just make a neural net to generate ascii art for you. But yeah these are mostly nice-to-haves. I need a congratulation message from Rishi in my life though.