Description
The new trigger system will be able to receive tracks reconstructed from silicon tracker hits by a level-1 track trigger and will evaluate algorithms that were previously reserved to the high-level trigger, such as vertex finding and particle flow. The final element in the level-1 trigger, the Global Trigger, will receive a large number of physics object collections reconstructed with different flavours of algorithms and will perform the final event selection through parallel evaluation of O(10^3) trigger algorithms. Besides cut-based algorithms, which will include topological conditions such as invariant or transverse mass, the Global Trigger will also be able to evaluate machine-learning based algorithms . We present studies leading to the design of the Global Trigger system, and discuss challenges in the FPGA firmware implementation on Xilinx Virtex Ultrascale Plus FPGAs, such as optimising resource usage by working at higher clock frequencies while keeping latency as low as possibl
