Speaker
Description
The Large Hadron Collider (LHC) collides bunches of protons spaced 25 ns apart at a total center of momentum energy of 13.6 TeV, producing an event rate of 40 MHz. This generates about a petabyte worth of information every second, but this is far too much data to feasibly save for offline analysis. To increase the chances of saving interesting physics events, the ATLAS detector implements a two-tiered trigger system designed to reduce the rate of accepted events that are stored for later analysis. The first tier of this trigger system is the Level 1 hardware trigger. It reduces the initial 40 MHz rate down to 100 kHz at a latency of 2.5 μs. For Run 3 of the LHC, upgrades were made to the Level 1 trigger, including new feature extractors (FEXs) designed to increase sensitivity to key physics channels. In particular, the global feature extractor (gFEX) takes advantage of its single board architecture to implement algorithms which cover the entire range of the calorimeters. This makes it ideal for identifying large radius jets, indicative of Lorentz boosted objects, as well as global values of interest such as missing transverse energy and total transverse energy, all of which are key signatures in many beyond Standard Model physics scenarios. This talk will cover the algorithms used by gFEX, the work done to validate these algorithms, and a discussion of gFEX performance so far in ATLAS data-taking during Run 3 of the LHC.
