HLT

HLT, in the context of experimental high-energy physics, stands for High-Level Trigger. It is a software-based data selection layer used in collider experiments to decide which events recorded by the detector should be kept for further analysis. It sits between the hardware-based Level-1 trigger, which rapidly reduces the raw event rate, and the permanent data storage, ensuring that only events with potential physics value are retained. By filtering events in real time, the HLT lowers the data rate from tens of thousands of events per second to a more manageable level, typically a few hundred to a few thousand events per second, depending on luminosity and experiment.

HLT operates on a computing farm, using fast, offline-like reconstruction and selection algorithms. It relies on up-to-date detector calibrations and conditions, reconstructs objects such as tracks, leptons, jets, and missing transverse energy, and applies physics criteria or multivariate analyses to decide whether to accept an event. The decisions are organized into trigger paths or menus; an event may fulfill multiple paths, and prescales may downscale less critical ones. Accepted events are stored in reduced data formats for offline processing, while rejected events are discarded.

The HLT is a central component of the data acquisition architecture in modern collider experiments such as ATLAS, CMS, and LHCb. It enables operation at high instantaneous luminosities by performing sophisticated event selection in real time, leveraging commodity CPU farms and, increasingly, accelerators. The design emphasizes speed, modularity, and configurability, allowing trigger menus to be adjusted between data-taking runs to optimize sensitivity to different physics goals.