npm

xtensor

C++ library for multi-dimensional arrays with broadcasting and lazy evaluation.

brewmacoslinux

Try with needOr install directly

Source

About

Multi-dimensional arrays with broadcasting and lazy computing

Examples

create a multi-dimensional array in C++

$ cat > example.cpp << 'EOF'
#include <xtensor/xarray.hpp>
int main() {
  xt::xarray<double> arr {{1, 2, 3}, {4, 5, 6}};
  return 0;
}
EOF

perform element-wise operations on arrays

$ cat > math.cpp << 'EOF'
#include <xtensor/xarray.hpp>
int main() {
  xt::xarray<int> a {1, 2, 3};
  xt::xarray<int> b {4, 5, 6};
  auto result = a + b;
  return 0;
}
EOF

use broadcasting to combine differently shaped arrays

$ cat > broadcast.cpp << 'EOF'
#include <xtensor/xarray.hpp>
int main() {
  xt::xarray<double> mat {{1, 2}, {3, 4}};
  xt::xarray<double> vec {10, 20};
  auto result = mat + vec;
  return 0;
}
EOF

compile C++ code using xtensor headers$ clang++ -I$(brew --prefix xtensor)/include example.cpp -o example

use lazy evaluation to optimize array computations

$ cat > lazy.cpp << 'EOF'
#include <xtensor/xarray.hpp>
int main() {
  xt::xarray<int> a {1, 2, 3};
  xt::xarray<int> b {4, 5, 6};
  auto lazy_result = a + b * 2;
  return 0;
}
EOF