End-to-end deep learning inference with CMSSW via ONNX using docker

• DOI: 10.48550/arxiv.2309.14254
• Publication Date: 2023-01-01
• Authors: Purva Chaudhari; Shravan Chaudhari; R. Chudasama; Sergei Gleyzer
• Keywords: High Energy Physics - Experiment (hep-ex); Physics; QC1-999; Physics - Data Analysis, Statistics and Probability; 0103 physical sciences; FOS: Physical sciences; 01 natural sciences; Data Analysis, Statistics and Probability (physics.data-an); High Energy Physics - Experiment

Deep learning techniques have been proven to provide excellent performance for a variety of high-energy physics applications, such as particle identification, event reconstruction and trigger operations. Recently, we developed an end-to-end deep approach identify various particles using low-level detector information from collisions. These models will be incorporated in the CMS software framework (CMSSW) enable their use or operation real-time. Incorporating these computational tools experimental presents new challenges. This paper reports implementation inference with framework. The has implemented on GPU faster computation ONNX. We benchmarked ONNX CPU NERSCs Perlmutter cluster by building docker image