Abstract
One of the challenges of high granularity calorimeters, such as that to be built to cover the endcap region in the CMS Phase-2 Upgrade for HL-LHC, is that the large number of channels causes a surge in the computing load when clustering numerous digitized energy deposits (hits) in the reconstruction stage. In this article, we propose a fast and fully parallelizable density-based clustering algorithm, optimized for high-occupancy scenarios, where the number of clusters is much larger than the average number of hits in a cluster. The algorithm uses a grid spatial index for fast querying of neighbors and its timing scales linearly with the number of hits within the range considered. We also show a comparison of the performance on CPU and GPU implementations, demonstrating the power of algorithmic parallelization in the coming era of heterogeneous computing in high-energy physics.
Highlights
Calorimeters with high lateral and longitudinal readout granularity, capable of providing a fine grained image of electromagnetic and hadronic showers, have been suggested for future high-energy physics experiments (CALICE Collaboration, 2012)
We describe CLUstering of Energy (CLUE), a novel and parallel density-based clustering
CLUE requires the following four parameters: dc is the cut-off distance in the calculation of local density; ρc is the minimum density to promote a point as a seed or the maximum density to demote a point as an outlier; δc and δo are the minimum separation requirements for seeds and outliers, respectively
Summary
Calorimeters with high lateral and longitudinal readout granularity, capable of providing a fine grained image of electromagnetic and hadronic showers, have been suggested for future high-energy physics experiments (CALICE Collaboration, 2012). The deposited energy is collected by the sensors on the layers that the shower traverses. The purpose of the clustering algorithm when applied to shower reconstruction is to group together individual energy deposits (hits) originating from a particle shower. Event reconstruction is tightly constrained by a millisecond-level execution time. This constraint requires the clustering algorithm to be highly efficient while maintaining a low computational complexity. It is highly preferable to have a fully parallelizable clustering algorithm to take advantage of the trend of heterogeneous computing with hardware accelerators, such as graphics processing units (GPUs), achieving a higher event throughput and a better energy efficiency
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
