Packaging HEP Heterogeneous Mini-apps for Portable Benchmarking and Facility Evaluation on Modern HPCs

2025-05-13

AI Summary

High-energy physics (HEP) experiments face significant challenges in deploying on GPU-accelerated HPC platforms, including rapid hardware evolution, complex software portability, and non-reproducible benchmarking. To address these, we propose the first fully automated containerization and configuration framework tailored to HEP heterogeneous micro-applications (e.g., Patatrack, FastCaloSim). Our framework integrates Docker/Singularity containerization, Spack-based dependency management, CI/CD-driven cross-platform configuration generation, and a GPU micro-benchmark abstraction layer. It enables out-of-the-box deployment and fully reproducible, zero-manual-intervention performance evaluation, supporting long-term performance tracking and regression analysis across multi-vendor GPU clusters (NVIDIA and AMD). End-to-end validation across international HPC facilities demonstrates a 90% reduction in deployment time, substantially improving infrastructure assessment efficiency and scientific rigor in hardware selection decisions.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web applications in cross-disciplinary domains and verticals such as mixed reality, smart cities, and digital health
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

High Energy Physics (HEP) experiments are making increasing use of GPUs and GPU dominated High Performance Computer facilities. Both the software and hardware of these systems are rapidly evolving, creating challenges for experiments to make informed decisions as to where they wish to devote resources. In its first phase, the High Energy Physics Center for Computational Excellence (HEP-CCE) produced portable versions of a number of heterogeneous HEP mini-apps, such as ptor, FastCaloSim, Patatrack and the WireCell Toolkit, that exercise a broad range of GPU characteristics, enabling cross platform and facility benchmarking and evaluation. However, these mini-apps still require a significant amount of manual intervention to deploy on a new facility. We present our work in developing turn-key deployments of these mini-apps, where by means of containerization and automated configuration and build techniques such as Spack, we are able to quickly test new hardware, software, environments and entire facilities with minimal user intervention, and then track performance metrics over time.

Problem

• Packaging HEP mini-apps for portable benchmarking on HPCs
• Reducing manual intervention in deploying heterogeneous HEP applications
• Enabling automated performance tracking across diverse GPU environments

Innovation

• Portable heterogeneous HEP mini-apps for benchmarking
• Containerization for automated deployment and testing
• Spack for configuration and build automation