Speeding up RooFit with auto-vectorization and batch evaluation.

RooFit is a module of the ROOT data analysis framework, used in all LHC and other experiments. The capabilities of it include expressing statistical models, with binned or unbinned likelihoods, parameter and error estimation, as well as the evaluation of statistical tests. Despite the importance of the module, development of it became effectively inactive since 2011 and was not resumed until 2018. The aim of my project was to rewrite critical parts of the code that were identified as bottlenecks, to boost the performance of RooFit, by gaining advantage of the cache and SIMD (Single Instruction Multiple Data) instructions.