Commissioning of the ATLAS High Level Trigger with Proton Collisions at the LHC

ATLAS is one of the two general-purpose detectors at the Large Hadron Collider (LHC). The ATLAS trigger system uses fast reconstruction algorithms to efficiently reject a large rate of background events and still select potentially interesting ones with good efficiency. After a first processing level (Level 1) using custom electronics, the trigger selection is made by software running on two processor farms, containing a total of around two thousand multi-core machines. This system is known as the High Level Trigger (HLT). To reduce the network data traffic and the processing time to manageable levels, the HLT uses seeded, step-wise reconstruction, aiming at the earliest possible rejection of background events. The recent LHC run at the end of 2009 provided the first proton collisions at the LHC, which soon became the world's highest energy particle collider. The ATLAS trigger was essential during this period to select events with beam activity, and achieve the first ATLAS physics results. During most of the run, event selection was done by the Level 1 trigger only, with the selected data being used to commission the HLT. At the end of the run, the HLT was used to reject background events online, to cope with the increasing LHC luminosity. During the ongoing 2010 LHC run, the HLT will play a crucial role in selecting collision data for the extraction of physics results. The experience gained during the 2009 run will be essentia l for the coming year. After giving an overview of the trigger design and its innovative features, this paper focusses on the valuable experience gained in running the trigger in the the 2009 LHC run. In addition to the monitoring and configuration infrastructures, it will emphasize the trigger operation and the performance of HLT algorithms, including comparisons to the precise offline reconstruction. Finally, it will discuss the HLT plans for the 2010 run.