Excerpts from the LHCb cookbook — RECEPTS for testing lepton universality and reconstructing primary vertices

Lepton universality is an accidental symmetry of the Standard Model of particles physics, which expects that the charged leptons have the same characteristics with the exception of their masses. Experimental tests of lepton universality have shown hints of deviations from the Standard Model. This thesis presents a test of lepton universality using semileptonic $b$-hadron decays to final states involving a muon or an electron using data recorded with the \mbox{LHCb} detector in the years 2016, 2017 and 2018. The test is performed by measuring the branching fraction of the $\overline{B}^0 \rightarrow D^{*+} e^- \overline{\nu}_e$ decay relative to the branching fraction of the $\overline{B}^0 \rightarrow D^{*+} \mu^- \overline{\nu}_\mu$ decay, with the result \begin{equation} R(D^{*+})_{light} = \frac{\mathcal{B}(\overline{B}^0 \rightarrow D^{*+} e^- \overline{\nu}_e)}{\mathcal{B}(\overline{B}^0 \rightarrow D^{*+} \mu^- \overline{\nu}_\mu)} \\ =X \times (2.937\pm 0.014(stat.) \pm 0.063 (syst.)), \end{equation} where $X$ is a blinding factor. This is the first time this ratio is measured at a hadron-hadron collider. It is shown that the systematic uncertainties related to the different behaviour of electrons and muons in the detector can be controlled to a level that a similar sensitivity to measurements at electron-positron colliders can be achieved. The LHCb Upgrade for Run 3 of the Large Hadron Collider will rely on a completely software-based trigger, whose first stage needs to process events at a rate of up to 30MHz. An implementation of the complete first trigger stage on GPUs was developed by the LHCb collaboration. During this thesis, the algorithm to reconstruct the proton-proton interaction points, necessary to identify displaced signatures, is implemented on GPUs. Changes to the original CPU implementation to achieve a fast and efficient algorithm are presented. The GPU algorithm is shown to fulfil the physics performance and throughput requirements of the Run 3 trigger.