High-precision QCD Meets High-performance Computing

高精度QCD遇见高性能计算

• 批准号: 1520916
• 负责人: Francis Petriello
• 金额: $ 13.95万
• 依托单位: Northwestern University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-08-01 至 2019-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1520916&HistoricalAwards=false
• 关键词: precision; QCD; Meets; performance; Computing

Our understanding of Nature at the smallest scales is encoded in the Standard Model (SM) of particle physics, which explains the properties of the elementary particles making up most of known matter in the Universe. Experiments such as those at the Large Hadron Collider (LHC) at the CERN laboratory in Switzerland are now searching for subtle deviations from SM predictions that would indicate a new theory beyond the SM. This effort requires increasingly intricate and precise calculations in order to determine whether the predictions of the SM are indeed obeyed at the percent-level. The goal of this project is to produce precision simulation programs that can take full advantage of modern parallel computing systems in order to enable the successful pursuit of the program of discovery at the LHC. The graduate student working with the investigator on this project will be well-positioned for a future career at the cutting edge of scientific research and high-performance computing.The goal of this research is to produce parallel versions of several heavily-used numerical programs that incorporate higher-order corrections in the perturbative expansion of Quantum Chromodynamics (QCD), the theory of the strong interactions. Several representative codes written by the PI will be restructured to utilize a hybrid MPI+openMP parallelization scheme. The resulting programs will be scalable to take advantage of the ever-increasing number of nodes available on modern computing platforms. In addition to allowing otherwise intractable phenomenology to be performed, through these codes we will also gain insight into how future QCD calculations can be successfully structured to take full advantage of high-performance computing.

我们在最小尺度上对自然的理解被编码在粒子物理学的标准模型（SM）中，它解释了构成宇宙中大多数已知物质的基本粒子的特性。瑞士欧洲核子研究中心（CERN）实验室的大型强子对撞机（LHC）等实验正在寻找与SM预测的细微偏差，这些偏差可能会表明一种超越SM的新理论。这项工作需要越来越复杂和精确的计算，以确定SM的预测是否确实在百分比水平上得到遵守。这个项目的目标是产生精确的模拟程序，可以充分利用现代并行计算系统，以便在大型强子对撞机上成功地追求发现项目。在这个项目上与研究者一起工作的研究生将为未来在科学研究和高性能计算领域的前沿职业生涯做好准备。本研究的目标是产生几个广泛使用的数值程序的并行版本，这些程序在强相互作用理论量子色动力学（QCD）的微扰扩展中包含高阶修正。由PI编写的几个代表性代码将被重组以利用混合MPI+openMP并行化方案。由此产生的程序将可扩展，以利用现代计算平台上可用的节点数量不断增加的优势。除了允许执行难以处理的现象学之外，通过这些代码，我们还将深入了解如何成功地构建未来的QCD计算，以充分利用高性能计算。