Improvements in vectorization and parallelization of ROOT Math libraries
- Mentors
- Xavier Valls, Lorenzo Moneta
- Organization
- CERN-HSF
ROOT is a data analysis software used by scientists all over the world; therefore, its efficiency is critical for the scientific community in general and for CERN in particular. This project aims to improve the efficiency of one of the most used ROOT features: the mathematical functions and the fitting methods, which will be accomplished by making the corresponding code vectorized and/or parallelized. The main tasks that will be covered during the coding period are:
- Complete the parallelization and vectorization of all the fitting methods available in ROOT:
Chi2FCN
,LogLikelihoodFCN
andPoissonLikelihoodFCN
. - Adapt the gradient function interfaces for thread-based parallelization and vectorization.
- Vectorize
TFormula
and predefined ROOT functions (gaus
,landau
,expo
,crystalball
, ...). - Vectorize the most used mathematical and statistical functions in
ROOT::TMath
andTMath
.
If time allows it, the special and statistical functions implemented in ROOT will be also vectorized and/or parallelized.