Uh oh!
There was an error while loading. Please reload this page.
numpcypp v1.21.2
Background
NumPy is fast, but its ceiling is locked by Python. numpcpp keeps NumPy's familiar usage patterns while letting C++ break through Python's performance ceiling to accelerate your code further.
This Release
- Pure header-only C++ library — zero external dependencies, C++17 compiler is all you need
- Pixel-level alignment with NumPy APIs:
numpy.*,numpy.linalg.*,numpy.einsum - float64 / float32 support — all 336 precision tests pass
Install
sudo dpkg -i numpycpp-dev-1.21.2-Linux.deb # headers → /usr/include/numpycpp/Usage (CMake)
find_package(numpycppREQUIRED)
target_link_libraries(myappPRIVATEnumpycpp::numpycpp)Usage (C++)
#include"numpy/core.h"
std::vector<double> data = {1.0, 4.0, 9.0};
std::vector<double> result(data.size());
numpy::sqrt(data.data(), result.data(), data.size());
// result → {1.0, 2.0, 3.0}Full Changelog: https://github.com/array2d/numpycpp/commits/v1.21.2