C++ Libraries

Boost

Period.

I have a specifically intense crush on the Boost Graph Library and Spirit.

Trivia: there are more people that have been in space than code contributors to Boost.

Signal Processing

OpenCV

Computer vision library.

FFTW (C library)

The “Fastest Fourier Transformation of the West”, a must-have.

ffmpeg (C library)

Video and audio encoder/decoder library. Second to none.

GPU Rendering

High Performance Computing

DASH

A Partitioned Global Address Space library, mimicking STL for high performance computing clusters.
Containers: array, unordered_map, list, narray (n-dim. array)
Algorithms: essential STL equivalents, some use cases under development, great scalability for distributed sorting

HPX

Similar to DASH, task-parallel programming model.
Containers: no STL-compliant containers (no global random access or not distributed)
Algorithms: all STL algorithms, task-parallel implementations with robust scalability

Blaze

The high performance math library. A real gem!

Machine Learning

Tensorflow

Of course.

Python Bindings

Writing Python modules in C++ is a breeze thanks to pybind11.
Python owes its success in HPC and data science to modules implemented in C++ for performance, like scipy and numpy.