Speaker
Description
The CMS collaboration has chosen a novel high granularity calorimeter (HGCAL) for the endcap regions as part of its planned upgrade for the high-luminosity phase of the LHC. The HGCAL will have fine segmentation in both the transverse and longitudinal directions, which leads to about six million readout channels in total. It will be part of the Level 1 (L1) trigger of the CMS experiment, and its high granularity coupled with the 40MHz sampling at the L1 trigger pose a significant challenge in terms of data manipulation and processing. Trigger cells and module energy sums are first constructed on-detector in the HGCAL front-end ASICs and used off-detector to build higher-level objects called trigger primitives in a two-stage FPGA-based system composed of about 200 ATCA boards. Several types of trigger primitives using the HGCAL data have been developed in order to provide complementary information to the central L1 trigger system of CMS. These trigger primitives are three-dimensional clusters that can efficiently be used by particle-flow algorithms, as well as projective tower energy sums, which provide a complete two-dimensional map of the detector. Reconstructing such objects is a complex task given the huge data throughput. It must in particular satisfy both the hardware and latency constraints of the system and provide optimal reconstruction performance for triggering under the harsh HL-LHC conditions. The algorithms reconstructing these trigger primitives are currently being optimized and implemented in hardware, and these implementations are undergoing an extensive validation using dedicated test systems and data collected during test beam. The status of the HGCAL trigger primitive generation system will be presented, together with the latest results on the implementation, hardware validation and optimization of its algorithms.