Device side linear algebra libraries / Cooperative groups implementation

The aim of this project was to implement GPU device-side libraries for linear algebra operations in the context of the Pixeltrack-Standalone framework, specifically for the code of PixelTriplets - Broken line. The focus was on evaluating the performance and suitability of MAGMA and cuBLASDx libraries for this purpose. The core requirement involved performing small-scale linear algebra operations, particularly the Cholesky decomposition, used for factorizing positive-definite matrices into the product of a lower triangular matrix and its transpose (LLT ), on matrices as small as 6 ×6.