The performance of the ATLAS Inner Detector Trigger algorithms in pp collisions at the LHC

The ATLAS Experiment, Inner Detector trigger algorithms have been running online during data taking with proton-proton collisions at the Large Hadron Collider (LHC) since December 2009. Preliminary results on the performance of the algorithms in collisions at a centre-of-mass energies of 900GeV and 7TeV are discussed. The ATLAS trigger performs the online event selection in three stages. The Inner Detector information is used in the second and third triggering stages, referred to as Level-2 trigger (L2) and Event Filter (EF) respectively, or collectively as the the High Level Trigger (HLT). The HLT runs software algorithms on large farms of commercial CPUs and is designed to reject collision events in real time, keeping the most interesting few events in every thousand. The average execution times per event at L2 and the EF are around 40~ms and 4~s respectively and the Inner Detector trigger algorithms can use only a fraction of these times. Within these times, data from interesting regions of the Inner Detector have to be read out through the network, unpacked, clustered and converted to the ATLAS global coordinates. The pattern recognition follows to identify the trajectories of charged particles (tracks), which are then used in combination with information from the other subdetectors to accept or reject events depending on whether they satisfy certain trigger signatures.