GooFit: A library for massively parallelising maximum-likelihood fits

Abstract

Fitting complicated models to large datasets is a bottleneck of many analyses. We present GooFit, a library and tool for constructing arbitrarily-complex probability density functions (PDFs) to be evaluated on nVidia GPUs or on multicore CPUs using OpenMP. The massive parallelisation of dividing up event calculations between hundreds of processors can achieve speedups of factors 200-300 in real-world problems.

Presented at the CHEP 2013 conference

References in corpus (1)

• The RooFit toolkit for data modeling

Cited by in corpus (8)

• Studies of the resonance structure in $D^0\to K_S^0{K}^{\pm }{\pi }^{\mp }$ decays
• \texttt{GooStats}: A GPU-based framework for multi-variate analysis in particle physics
• GNA: new framework for statistical data analysis
• Amplitude analysis of four-body decays using a massively-parallel fitting framework
• Massively Parallel Computing at the Large Hadron Collider up to the HL-LHC
• Power-aware applications for scientific cluster and distributed computing
• MoreFit: A More Optimised, Rapid and Efficient Fit
• A GPU based multidimensional amplitude analysis to search for tetraquark candidates