High Performance Computing Group
High Energy Physics Center for Computational Excellence
The High Energy Physics Center for Computational Excellence (HEP-CCE) is a multi-lab activity aimed at bringing new computational capabilities to bear in support of High Energy Physics (HEP) science goals. These include porting and optimizing HEP codes and frameworks for next-generation high-performance computing (HPC) systems, data-intensive computing tasks on HPC platforms, and addressing computation-related training issues for the HEP workforce. Many of these activities are conducted in partnership with DOE Advanced Scientific Computing Research (ASCR) scientists.
HPC systems at DOE supercomputing facilities can be a powerful resource for HEP experiments, provided the software can be suitably refactored, while satisfying two key requirements: efficiency and portability. The aim of this HPC-CCE project is to address this question via pilot sub-projects and collaborative research studies with the DOE-ASCR HPC community in: 1) portable parallelization strategies, 2) fine-grained input/output (I/O) and related storage issues, 3) optimizing event generators, and 4) running complex workflows on HPC systems.
CSI scientists are engaged in every aspect of the project with a particular focus on portable applications and workflows across different HPC systems. Portable programming models, such as Kokkos, SYCL and OpenMP, have been investigated for ATLAS and Deep Underground Neutrino Experiment testbeds with demonstrated speedups on graphics processing unit (GPU) accelerators. These studies will inform the HEP experiments of the potential of different portability solutions on GPU accelerators.
Publications
Atif, M., et al., Evaluating Portable Parallelization Strategies for Heterogeneous Architectures in High Energy Physics, (2023). http://arxiv.org/abs/2306.15869, arXiv:2306.15869.
Bhattacharya, M., et al., Portability: A Necessary Approach for Future Scientific Software, in: Snowmass 2021, 2022. http://arxiv.org/abs/2203.09945, arXiv:2203.09945 [physics.comp-ph]
Lin, Meifeng, et al. “Portable Programming Model Exploration for LArTPC Simulation in a Heterogeneous Computing Environment: OpenMP vs. SYCL.” arXiv preprint arXiv:2304.01841 (2023).
Dong, Zhihua, et al. “Evaluation of Portable Programming Models to Accelerate LArTPC Detector Simulations.” Journal of Physics: Conference Series. Vol. 2438. No. 1. IOP Publishing, 2023.
Yu, Haiwang, et al. “Evaluation of Portable Acceleration Solutions for LArTPC Simulation Using Wire-Cell Toolkit.” EPJ Web of Conferences. Vol. 251. EDP Sciences, 2021.
Pascuzzi, Vincent and Mehdi Goli, Mehdi. “Benchmarking a Proof-of-Concept Performance Portable SYCL-based Fast Fourier Transformation Library”, IWOCL’22: International Workshop on OpenCL, May 2022, Article No.: 20, Pages 1-9, DOI: 10.1145/3529538.3529996
Pascuzzi, Vincent and Mehdi Goli, Mehdi. “Achieving near native runtime performance and cross-platform performance portability for random number generation through SYCL interoperability”, arXiv:2109.01329.
Dong, Zhihua, et al. “Porting HEP Parameterized Calorimeter Simulation Code to GPUs.” Frontiers in Big Data. arXiv:2103.14737v2.
Yu, Haiwang, et al. “Evaluation of Portable Acceleration Solutions for LArTPC Simulation Using Wire-Cell Toolkit.” vCHEP 2021. arXiv:2104.08265v1.
