High Performance Computing software libraries (Intel) for data reduction, barycenter, first and second momenta calculation.
The deliverable participation of the LAPP is composed of two different parts, a High Performance Computing library and some programs described below.
The library will be composed of fast elementary functions, those meet the requirements for actual big experiments (CTA, SKA, …), to guarantee the reusability of those functions. For example, we can perform a vector reduction 10 times faster than numpi in python. Or barycenter calculation more than 3 times faster than the same vectorized by GCC, in C of C++.
For the moment, we have the basis of the library and a developer documentation. This library will be in development, we will also provide the documentation as soon as possible.