Heterogeneous techniques for rescaling energy deposits in the CMS Phase-2 endcap calorimeter

Abstract

We present the porting to heterogeneous architectures of the algorithm used for applying linear transformations of raw energy deposits in the CMS High Granularity Calorimeter (HGCAL). This is the first heterogeneous algorithm to be fully integrated with HGCAL’s reconstruction chain. After introducing the latter and giving a brief description of the structural components of HGCAL relevant for this work, the role of the linear transformations in the calibration is reviewed. The many ways in which parallelization is achieved are described, and the successful validation of the heterogeneous algorithm is covered. Detailed performance measurements are presented, including throughput and execution time for both CPU and GPU algorithms, therefore establishing the corresponding speedup. We finally discuss the interplay between this work and the porting of other algorithms in the existing reconstruction chain, as well as integrating algorithms previously ported but not yet integrated.

Clustering in the Heterogeneous Reconstruction Chain of the CMS HGCAL Detector

“…• Data Input: The input is now received from the preceding hit calibration stage [7] in the GPU, avoiding additional host-device transfers present in the standalone version.…”

“…Both validation and performance measurements are executed starting from three uncalibrated CPU collections of hits (one per subdetector). The hits are transferred to the GPU, are calibrated using techniques described in [7], and are directly forwarded to CLUE. Finally, GPU data in SoA format is converted back to the Legacy CPU format (as in fig.…”

“…This work presents the second HGCAL heterogeneous algorithm to be fully integrated within CMSSW after hit calibration [7]. It describes the algorithm's porting to CMSSW and the GPU implementation of additional, non embarassingly-parallel cluster-related steps.…”

“…This paper describes the integration of the standalone GPU CLUE (CLUstering of Energy) algorithm [11] with the CMS software (CMSSW) [10]. This effort follows the development of the CMSSW GPU based calibration of energy deposits [7], which immediately preceeds CLUE in the reconstruction chain. This work thus measures the performance of two HGCAL GPU tasks without resorting to intermediate CPU/GPU data transfers.…”