Started putting together a new toy gfx framework based on python with pybind11 for directx12 etc.
February 15, 2025 at 6:51 PM
Started putting together a new toy gfx framework based on python with pybind11 for directx12 etc.
Cython was where we were headed due to Numpy, but we're pivoting to pybind11 due to its own support for Numpy, using C++ directly rather custom lang & tool, and being where a lot of mindshare seems to be now. See here for numpy support in pybind11: pybind11.readthedocs.io/en/stable/ad...
NumPy - pybind11 documentationContentsMenuExpandLight modeDark modeAuto light/dark mode
pybind11.readthedocs.io
January 9, 2025 at 10:55 PM
Cython was where we were headed due to Numpy, but we're pivoting to pybind11 due to its own support for Numpy, using C++ directly rather custom lang & tool, and being where a lot of mindshare seems to be now. See here for numpy support in pybind11: pybind11.readthedocs.io/en/stable/ad...
Given the focus on performance, I would suggest to switch from pybind11 to nanobind. Should just be a tiny change 😇
August 11, 2025 at 3:05 PM
Given the focus on performance, I would suggest to switch from pybind11 to nanobind. Should just be a tiny change 😇
LLVM Weekly - #573, December 23rd 2024. EuroLLVM and AsiaLLVM save the date, profile guided static data partitioning, 19.1.6 released, type sanitizer committed, ClangIR upstreaming continues, pybind11 => nanobind for MLIR Python bindings, and more llvmweekly.org/issue/573
LLVM Weekly - #573, December 23rd 2024
llvmweekly.org
December 23, 2024 at 6:16 PM
LLVM Weekly - #573, December 23rd 2024. EuroLLVM and AsiaLLVM save the date, profile guided static data partitioning, 19.1.6 released, type sanitizer committed, ClangIR upstreaming continues, pybind11 => nanobind for MLIR Python bindings, and more llvmweekly.org/issue/573
今日いよいよOpenMP&C++&Pybind11のバグの温床から脱却して仮想環境丸ごとぶち込むPythonのmultiprocessing単にしたろうかな。
October 25, 2024 at 5:39 AM
今日いよいよOpenMP&C++&Pybind11のバグの温床から脱却して仮想環境丸ごとぶち込むPythonのmultiprocessing単にしたろうかな。
The pybind11-stubgen project seeks support in #maintaining it! We #pybind11 maintainers would move it to the pybind mainline org, but we need a community to help support the .pyi stub generator <3
Chime in here:
github.com/sizmailov/py...
#opensource #sustainability #CallForMaintainers
Chime in here:
github.com/sizmailov/py...
#opensource #sustainability #CallForMaintainers
Call for maintainers · Issue #244 · sizmailov/pybind11-stubgen
I've started pybind11-stubgen when there were no good alternatives to make good enough stubs for pybind modules. As of today I don't have a C++ project that needs python bindings, so I'm a bit out ...
github.com
January 16, 2026 at 9:48 PM
The pybind11-stubgen project seeks support in #maintaining it! We #pybind11 maintainers would move it to the pybind mainline org, but we need a community to help support the .pyi stub generator <3
Chime in here:
github.com/sizmailov/py...
#opensource #sustainability #CallForMaintainers
Chime in here:
github.com/sizmailov/py...
#opensource #sustainability #CallForMaintainers
also finished my python class for the semester. I always end with an example of extension modules, comparing C++ (pybind11), Fortran (f2py), Cython, and Numba to python + numpy:
sbu-python-class.github.io/python-scien...
all of these tools are really great.
#astro
sbu-python-class.github.io/python-scien...
all of these tools are really great.
#astro
Example Extension — PHY 546: Python for Scientific Computing
sbu-python-class.github.io
May 9, 2025 at 10:59 AM
also finished my python class for the semester. I always end with an example of extension modules, comparing C++ (pybind11), Fortran (f2py), Cython, and Numba to python + numpy:
sbu-python-class.github.io/python-scien...
all of these tools are really great.
#astro
sbu-python-class.github.io/python-scien...
all of these tools are really great.
#astro
C++ W! With my dataset, C++ library called from Python with pybind11 uses ~20GB of memory and takes ~80 minutes while the Cython version took over 60GB of memory and took 3+ hours.
May 29, 2025 at 11:56 PM
C++ W! With my dataset, C++ library called from Python with pybind11 uses ~20GB of memory and takes ~80 minutes while the Cython version took over 60GB of memory and took 3+ hours.
too much multitasking today, but got most of my last python lecture of the semester in order: extension modules
here's a Mandelbrot generator with python + numpy, Numba, Fortran + f2py, and C++ + pybind11:
sbu-python-class.github.io/python-scien...
still might need some optimizations
#astro
here's a Mandelbrot generator with python + numpy, Numba, Fortran + f2py, and C++ + pybind11:
sbu-python-class.github.io/python-scien...
still might need some optimizations
#astro
Example Extension — PHY 546: Python for Scientific Computing
sbu-python-class.github.io
April 23, 2024 at 11:12 PM
too much multitasking today, but got most of my last python lecture of the semester in order: extension modules
here's a Mandelbrot generator with python + numpy, Numba, Fortran + f2py, and C++ + pybind11:
sbu-python-class.github.io/python-scien...
still might need some optimizations
#astro
here's a Mandelbrot generator with python + numpy, Numba, Fortran + f2py, and C++ + pybind11:
sbu-python-class.github.io/python-scien...
still might need some optimizations
#astro
It turns out that all builds are broken. Not as crash, but as returning rubbish.
The Codex is not guilty though - that is because of old pybind11<->numpy 2.0 interactions.
Sadly, Codex couldn't figure out on itself, so I had time to investigate and fix manually.
PyPI updated.
The Codex is not guilty though - that is because of old pybind11<->numpy 2.0 interactions.
Sadly, Codex couldn't figure out on itself, so I had time to investigate and fix manually.
PyPI updated.
January 8, 2026 at 2:51 PM
It turns out that all builds are broken. Not as crash, but as returning rubbish.
The Codex is not guilty though - that is because of old pybind11<->numpy 2.0 interactions.
Sadly, Codex couldn't figure out on itself, so I had time to investigate and fix manually.
PyPI updated.
The Codex is not guilty though - that is because of old pybind11<->numpy 2.0 interactions.
Sadly, Codex couldn't figure out on itself, so I had time to investigate and fix manually.
PyPI updated.
I wonder how much effort it would be to wiggle sol2 into C++23
I really don't want to use the lua C API directly, especially as I'm also already using pybind11
Plus the ability to swap out for luajit is nice.
Weh
I really don't want to use the lua C API directly, especially as I'm also already using pybind11
Plus the ability to swap out for luajit is nice.
Weh
February 17, 2025 at 2:43 AM
I wonder how much effort it would be to wiggle sol2 into C++23
I really don't want to use the lua C API directly, especially as I'm also already using pybind11
Plus the ability to swap out for luajit is nice.
Weh
I really don't want to use the lua C API directly, especially as I'm also already using pybind11
Plus the ability to swap out for luajit is nice.
Weh
The Python bindings were redesigned from scratch using a new library (pybind11→nanobind), fixing long-standing performance issues. Code generation of various operations was improved as well. We see 10-20× faster tracing in Dr.Jit and 1.2-2× wall-clock (inverse) rendering time.
November 26, 2024 at 3:09 PM
The Python bindings were redesigned from scratch using a new library (pybind11→nanobind), fixing long-standing performance issues. Code generation of various operations was improved as well. We see 10-20× faster tracing in Dr.Jit and 1.2-2× wall-clock (inverse) rendering time.
Uno v2.2.0 is out with C, Julia and Python interfaces 🤩
- pure #Julia interfaces to NLPModels.jl and MathOptInterface.jl
- #Python bindings using #pybind11
Developed at @zuseinstitute.bsky.social
#optimization #mathematicaloptimization #numericaloptimization #constrainedoptimization #oss #foss #uno
- pure #Julia interfaces to NLPModels.jl and MathOptInterface.jl
- #Python bindings using #pybind11
Developed at @zuseinstitute.bsky.social
#optimization #mathematicaloptimization #numericaloptimization #constrainedoptimization #oss #foss #uno
GitHub - cvanaret/Uno: A next-gen SQP & barrier solver for nonlinearly constrained optimization
A next-gen SQP & barrier solver for nonlinearly constrained optimization - cvanaret/Uno
github.com
October 1, 2025 at 1:52 PM
Uno v2.2.0 is out with C, Julia and Python interfaces 🤩
- pure #Julia interfaces to NLPModels.jl and MathOptInterface.jl
- #Python bindings using #pybind11
Developed at @zuseinstitute.bsky.social
#optimization #mathematicaloptimization #numericaloptimization #constrainedoptimization #oss #foss #uno
- pure #Julia interfaces to NLPModels.jl and MathOptInterface.jl
- #Python bindings using #pybind11
Developed at @zuseinstitute.bsky.social
#optimization #mathematicaloptimization #numericaloptimization #constrainedoptimization #oss #foss #uno
Nice! pybind11 is great.
Could ADBC be added next? It has huge performance advantages over ODBC.
arrow.apache.org/adbc/current...
Could ADBC be added next? It has huge performance advantages over ODBC.
arrow.apache.org/adbc/current...
Driver Implementation Status
Implementation Status: Experimental drivers are not feature-complete and the implementation is still progressing. Beta drivers are (mostly) feature-complete but have only been available for a short...
arrow.apache.org
August 13, 2025 at 1:32 PM
Nice! pybind11 is great.
Could ADBC be added next? It has huge performance advantages over ODBC.
arrow.apache.org/adbc/current...
Could ADBC be added next? It has huge performance advantages over ODBC.
arrow.apache.org/adbc/current...
Cython is great. For MetPy we’re going to be using straight C++ with pybind11 (and scikit-build-core)
July 27, 2025 at 5:25 PM
Cython is great. For MetPy we’re going to be using straight C++ with pybind11 (and scikit-build-core)
a lot of this code was tightly connected in a way that required it to all be built together:
`pybind11` bindings generated from C++ code, `solib` dependencies on OpenMPI, and a bunch of FFI from Haskell to C++ (and vice versa).
`pybind11` bindings generated from C++ code, `solib` dependencies on OpenMPI, and a bunch of FFI from Haskell to C++ (and vice versa).
November 13, 2024 at 6:20 PM
a lot of this code was tightly connected in a way that required it to all be built together:
`pybind11` bindings generated from C++ code, `solib` dependencies on OpenMPI, and a bunch of FFI from Haskell to C++ (and vice versa).
`pybind11` bindings generated from C++ code, `solib` dependencies on OpenMPI, and a bunch of FFI from Haskell to C++ (and vice versa).
pybind11 3.0 is out (in time for #scipy2025)! Support for Python 3.14, 3.14t, GraalPy, PyPy 3.11, and iOS have been added. smart-holder merged, native enums, multiphase init, subinterpreters (with per-interp GIL), better typing, C++23, and much more! See iscinumpy.dev/post/pybind1...
pybind11 3.0.0
Pybind11 3.0 has been released! I would like to highlight some of the key
changes below; be sure to check out the changelog and upgrade guide for
more information! This release includes an ABI bump, t...
iscinumpy.dev
July 10, 2025 at 10:06 PM
pybind11 3.0 is out (in time for #scipy2025)! Support for Python 3.14, 3.14t, GraalPy, PyPy 3.11, and iOS have been added. smart-holder merged, native enums, multiphase init, subinterpreters (with per-interp GIL), better typing, C++23, and much more! See iscinumpy.dev/post/pybind1...
scikit-build is definitely what I would recommend if you can make it work; not sure what it's status is with regard to SWIG (which has definitely fallen out of vogue). We'll be adding compiled extensions with that to MetPy in the short term, but those will be using C++ & Pybind11.
January 9, 2025 at 9:59 PM
scikit-build is definitely what I would recommend if you can make it work; not sure what it's status is with regard to SWIG (which has definitely fallen out of vogue). We'll be adding compiled extensions with that to MetPy in the short term, but those will be using C++ & Pybind11.
I had the misfortune of having to work with a codebase using Boost.Python a couple of years ago. I’d go with pybind11 today
April 19, 2025 at 2:55 AM
I had the misfortune of having to work with a codebase using Boost.Python a couple of years ago. I’d go with pybind11 today
the snake community have ctypes, cffi, cython, pybind11 and more (CPython is just the name of the official reference python implementation in C like GNU R for R).
If we really want to covert all of these, we need more packages like Rffi and nCompiler for example, not just Rcpp et Co
If we really want to covert all of these, we need more packages like Rffi and nCompiler for example, not just Rcpp et Co
August 14, 2025 at 10:19 PM
the snake community have ctypes, cffi, cython, pybind11 and more (CPython is just the name of the official reference python implementation in C like GNU R for R).
If we really want to covert all of these, we need more packages like Rffi and nCompiler for example, not just Rcpp et Co
If we really want to covert all of these, we need more packages like Rffi and nCompiler for example, not just Rcpp et Co
I think I figured out which language I should spend time on learning to pick up as my 2nd language as a Python developer: C++ is the winner because you can wrap C++ into Python using Pybind11 which speeds up processing time as one of the biggest cons of Python is it runs slow and C++ runs fast.
April 23, 2025 at 1:53 PM
I think I figured out which language I should spend time on learning to pick up as my 2nd language as a Python developer: C++ is the winner because you can wrap C++ into Python using Pybind11 which speeds up processing time as one of the biggest cons of Python is it runs slow and C++ runs fast.
ほんとにopenMPとPybind11の組み合わせが諸悪の根源すぎてこの癌を摘出したいのに、この作業が時間掛かる上に別にプロジェクトに進捗を出すわけじゃないから一生後回しなのがキツい。
October 14, 2024 at 3:36 PM
ほんとにopenMPとPybind11の組み合わせが諸悪の根源すぎてこの癌を摘出したいのに、この作業が時間掛かる上に別にプロジェクトに進捗を出すわけじゃないから一生後回しなのがキツい。
djzeeez finally figured out how use boost-like outcome support with pybind11, I had to go outside of the boost library sources to the origin tho github.com/ned14/outcome/
GitHub - ned14/outcome: Provides very lightweight outcome<T> and result<T> (non-Boost edition)
Provides very lightweight outcome<T> and result<T> (non-Boost edition) - ned14/outcome
github.com
February 23, 2025 at 3:14 AM
djzeeez finally figured out how use boost-like outcome support with pybind11, I had to go outside of the boost library sources to the origin tho github.com/ned14/outcome/