Deepanshu Thakur Email & Phone Number

Jaipur, Rj, In

India

I have worked on the Python bindings for the Hydra C++ library under the CERN-HSF organization. Hydra is a header-only C++ library for data analysis in massively parallel platforms which supports algorithms like "Generation of phase-space Monte Carlo Samples with any number of particles in the final states". Hydra is implemented on top of the Thrust library. One of the main focuses of the thrust library is speed and precision, for example, the simple sort function of thrust library sorts 846M keys per second on GeForce GTX 480.My role in this project was to write the Python bindings for this library, so library can be used with Python without any loss of precision and speed. I have used pybind11 as the main framework to bind the C++ code with Python with custom-written Parsers. I have also used Numba to compile the functions and methods created inside the Python interpreter and passed their address to C++ code, due to the GIL limitations of Python.Technologies used: - SWIG - Boost Python - Pybind11 - Numba - Cython - C/C++ libraries https://github.com/MultithreadCorner/HydraCheck the official Hydra.Python repository. https://github.com/MultithreadCorner/Hydra.PythonA medium blog post for CRTP pattern implementation with Pybind11https://medium.com/@deepanshu2017/a-curiously-recurring-python-d3a441a58174