基于统一计算架构加速的γ能谱分析系统设计  

The design of a gamma spectrum analysis system accelerated by CUDA

作　　者：谢浩浩 王仁波 陈锐[3] XIE Haohao;WANG Renbo;CHEN Rui(School of Information Engineering,East China University of Technology,Nanchang 330013,China;Institute of Nuclear Application Technology,East China University of Technology,Nanchang 330013,China;Jiangxi Engineering Research Center of Process and Equipment for New Energy,East China University of Technology,Nanchang 330013,China)

机构地区：[1]东华理工大学信息工程学院,江西南昌330013 [2]东华理工大学核应用技术研究所,江西南昌330013 [3]东华理工大学江西省新能源工艺及装备工程技术研究中心,江西南昌330013

出　　处：《世界核地质科学》2024年第6期1203-1213,共11页World Nuclear Geoscience

基　　金：国家自然科学基金项目(编号:12365026);江西省重点研发计划重点项目(编号:20232BBE50013);江西省自然科学基金(编号:20242BAB25046)资助。

摘　　要：在现代核科学与技术应用中,γ能谱分析是一种重要的技术手段,用于放射性核素的定性和定量分析。然而,传统的γ能谱分析系统在处理高通量数据时面临诸多挑战,尤其是在计算量大、处理速度慢的情况下,常常难以满足实际需求。因此,如何提高γ能谱分析的效率和准确性,成为亟待解决的问题。旨在通过设计一种基于CUDA(Compute Unified Device Architecture)加速的γ能谱分析系统,来提高γ能谱分析的处理速度与效率,从而满足高通量数据处理的需求。为此,本研究采用一种新颖的系统架构CUDA并行计算方法,利用CUDA的并行计算能力来实现γ能谱的快速处理。该系统实现谱光滑、峰值检测和核素识别等关键步骤,能够有效地处理多张能谱图。通过将这些计算任务分配到多个处理核心上,系统可以在短时间内完成传统方法所需的复杂计算从而实现计算任务的加速。此外,为了验证系统的有效性,在实际应用场景中进行多次测试,收集大量数据,以确保系统在不同条件下的稳定性和可靠性。测试结果显示:基于CUDA的γ谱分析系统在处理效率方面相比于传统的CPU计算方式,其处理效率提高数倍,系统在准确性和稳定性方面也得到保证。本研究设计的基于CUDA加速的γ谱分析系统,通过并行计算有效提高数据处理效率,满足高通量数据处理的需求,同时保持较高准确性和稳定性。In modern nuclear science and technology applications,γ-ray spectrum analysis is an important technique for the qualitative and quantitative analysis of radioactive isotopes.However,traditionalγ-ray spectrum analysis systems faces numerous challenges while handling high-throughput data,particularly in cases where the computational workload is large and processing speed is slow,an often fail to meet practical demands.Therefore,enhancing the efficiency and accuracy ofγ-ray spectrum analysis has become an urgent problem to solve.This study aims to design aγ-ray spectrum analysis system accelerated by CUDA(Compute Unified Device Architecture)to enhance the processing speed and efficiency ofγ-ray spectrum analysis so as to meet the demands of high-throughput data processing.To achieve this goal,the research adopted an innovative system architecture that utilized the parallel computing capabilities of CUDA for fast processing ofγ-ray spectra.Specifically,the system implemented the key steps such as spectrum smoothing,peak detection,and nuclide identification,effectively handling multiple spectra simultaneously.By distributing these computational tasks across multiple processing cores,the system completed complex calculations that would typically require more time with traditional methods.Furthermore,to validate the effectiveness of the system,multiple tests were conducted in real-world application scenarios by collecting a large volume of data to ensure the system's stability and reliability under various conditions.The test results showed that the CUDA-basedγ-ray spectrum analysis system significantly improved the processing efficiency compared to traditional CPUbased methods,with a multiple-fold increase in speed.At the same time,the system maintained accuracy and stability.Overall,the CUDA-acceleratedγ-ray spectrum analysis system designed in this study effectively enhances data processing efficiency through parallel computing,meets the needs of high-throughput data processing,and maintains a high level of a

关 键 词：γ能谱分析 CUDA 加速计算 系统设计 

分 类 号：P619.1[天文地球—矿床学] TP391[天文地球—地质学]