Operational Experience of the ATLAS High Level Trigger with Single-Beam and Cosmic Rays

ATLAS is one of two general-purpose detectors at the LHC. Using fast reconstruction algorithms, the trigger system needs to efficiently reject a large rate of background events while keeping potentially interesting ones with high efficiency. After a first level trigger implemented in custom electronics, the trigger selection is made by software running on two processor farms (the High Level Trigger system), containing a total of around two thousand multi-core machines. To reduce the network data traffic and the processing time to manageable levels, the HLT uses seeded, step-wise event reconstruction, aiming at the earliest possible rejection of background events. The LHC start up and single-beam run periods in 2008 provided a "stress test" of the trigger system. Following this period, ATLAS continued to collect cosmic-ray events for detector alignment and calibration as well as for commissioning the trigger. These running periods allowed us to exercise the trigger system online, including its configuration and monitoring infrastructure, as well as reconstruction and selection algorithms. Several tracking, muon-finding, and calorimetry algorithms were commissioned under different running conditions. Frequent changes of the trigger configuration were required to cope with the parallel commissioning of the ATLAS sub-detectors. The experience gained while running the trigger system online was very valuable to design and implement a n optimal strategy for the collision data taking period of 2009. This paper focuses on the operational experience gained in running the trigger in the fast-changing environment of the detector commissioning with cosmic rays and single beam runs.