Numpy with eigen c++ binding
it seems like numpy is moving toward c++ implementation, so is there any plan to have eigen c++ library also inline with numpy to get better performance? is there any ongoing work happening around this?
On 6/6/23 06:46, darshan patel wrote:
it seems like numpy is moving toward c++ implementation, so is there any plan to have eigen c++ library also inline with numpy to get better performance? is there any ongoing work happening around this?
NumPy does not currently use eigen and I am not aware of any plans to do so. Matti
It looks as though pybind11 can serve as a bridge between NumPy and Eigen: https://pybind11.readthedocs.io/en/stable/advanced/cast/eigen.html On Sun, Jun 11, 2023 at 2:39 AM Matti Picus <matti.picus@gmail.com> wrote:
What would be the point? You would need to add support in Eigen for most of what makes Numpy Numpy. Eigen is really about 2D, Tensors are not even close to be something working properly (not even talking about the dynamic number of dimensions that breaks), and I'm not talking about broadcasting rules. Matthieu Le dim. 11 juin 2023 à 08:38, Lev Maximov <lev.maximov@gmail.com> a écrit :
-- Quantitative researcher, Ph.D. Blog: http://blog.audio-tk.com/ LinkedIn: http://www.linkedin.com/in/matthieubrucher
FWIW xtensor is supposed to be more of a drop in C++ replacement.. --- Rohit ________________________________ From: Matthieu Brucher <matthieu.brucher@gmail.com> Sent: Sunday, 11 June 2023 08:27 To: Discussion of Numerical Python Subject: [Numpy-discussion] Re: Numpy with eigen c++ binding What would be the point? You would need to add support in Eigen for most of what makes Numpy Numpy. Eigen is really about 2D, Tensors are not even close to be something working properly (not even talking about the dynamic number of dimensions that breaks), and I'm not talking about broadcasting rules. Matthieu Le dim. 11 juin 2023 à 08:38, Lev Maximov <lev.maximov@gmail.com> a écrit :
-- Quantitative researcher, Ph.D. Blog: http://blog.audio-tk.com/ LinkedIn: http://www.linkedin.com/in/matthieubrucher
On 6/6/23 06:46, darshan patel wrote:
it seems like numpy is moving toward c++ implementation, so is there any plan to have eigen c++ library also inline with numpy to get better performance? is there any ongoing work happening around this?
NumPy does not currently use eigen and I am not aware of any plans to do so. Matti
It looks as though pybind11 can serve as a bridge between NumPy and Eigen: https://pybind11.readthedocs.io/en/stable/advanced/cast/eigen.html On Sun, Jun 11, 2023 at 2:39 AM Matti Picus <matti.picus@gmail.com> wrote:
What would be the point? You would need to add support in Eigen for most of what makes Numpy Numpy. Eigen is really about 2D, Tensors are not even close to be something working properly (not even talking about the dynamic number of dimensions that breaks), and I'm not talking about broadcasting rules. Matthieu Le dim. 11 juin 2023 à 08:38, Lev Maximov <lev.maximov@gmail.com> a écrit :
-- Quantitative researcher, Ph.D. Blog: http://blog.audio-tk.com/ LinkedIn: http://www.linkedin.com/in/matthieubrucher
FWIW xtensor is supposed to be more of a drop in C++ replacement.. --- Rohit ________________________________ From: Matthieu Brucher <matthieu.brucher@gmail.com> Sent: Sunday, 11 June 2023 08:27 To: Discussion of Numerical Python Subject: [Numpy-discussion] Re: Numpy with eigen c++ binding What would be the point? You would need to add support in Eigen for most of what makes Numpy Numpy. Eigen is really about 2D, Tensors are not even close to be something working properly (not even talking about the dynamic number of dimensions that breaks), and I'm not talking about broadcasting rules. Matthieu Le dim. 11 juin 2023 à 08:38, Lev Maximov <lev.maximov@gmail.com> a écrit :
-- Quantitative researcher, Ph.D. Blog: http://blog.audio-tk.com/ LinkedIn: http://www.linkedin.com/in/matthieubrucher
participants (5)
-
darshan patel -
Lev Maximov -
Matthieu Brucher -
Matti Picus -
Rohit Goswami