Ilhan Polat
Ilhan Polat
Did not really sit down and lay out the differences but if this is working, it is a major improvement. Congrats @steppi ! Much appreciated. > One thing that is...
re-run the test, now passing
Nevermind that failure, it is not related but #18654 This looks good to me. But that error change might break some code somewhere. So I'm not sure about it
By the way, if we remove the 2D checks from them, `cosm`, `sinm`, `tanm`, `coshm`, `sinhm`, and `tanhm` are ready to accept nDarray inputs since they are all arithmetics over...
Thanks @ollybritton this is a very needed and nice addition. Here are two immediate comments from my side; 1- Why `cub` and not `cubature`? It does not seem like a...
If you have some identified parts of the code that are deemed to be blockingly slow, I can help with Cython/C translations.
> and because it would be nice to eventually support the array API. We discussed this a few times before. That's not really a blocker if it gets in front...
Way back when we wrote a very terse version of what needs to be done in the most hidden place which at the top of `scipy.linalg.lapack` doc page. https://docs.scipy.org/doc/scipy/reference/linalg.lapack.html If...
That's my laziness when I was converting fortran stack array allocations. I'm not sure whether conversion to heap was a good move but pushing 1000 items to stack is also...
In general, it does not feell right to alloc for a special function/ufunc if it is running over a long array. Especially functions like `kmn` doing some expansion coefficients, every...