Modern On-Site High Performance Computing System

现代现场高性能计算系统

• 批准号: 10431417
• 负责人: Scott Bunnell
• 金额: $ 57.79万
• 依托单位: SUNY DOWNSTATE MEDICAL CENTER
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-07-01 至 2023-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10431417
• 关键词: Advisory Committees; Biophysics; Cardiac; Computer Systems; Computers; Data; Data Set; Equipment; Funding; High Performance Computing; Hour; Infrastructure; Logistics; Maintenance; Minor; Modernization; Occupations; Policies; Process; Research; Resources; Services; Site; Speed; System; Time; United States National Institutes of Health; Wages; Wait Time; computational neuroscience; computerized data processing; cost; data storage site; instrument; operation; psychogenetics; simulation; success

Project Summary / Abstract This proposal is for the acquisition of a computational instrument (CI) that will enable effective on-site storage, creation, and processing of datasets that are currently stored, created, and processed almost exclusively off- site, which is both logistically non-optimal and monetarily profligate. Said instrument will directly support a group of NIH awardees (also referred to as the Major Users), in addition to a subsidiary group of secondary users (also referred to as the Minor Users), whose continued success at SUNY Downstate is highly dependent on the capabilities of the proposed equipment. The research breadth of the Major Users comprises computational neuroscience, cardiac biophysics, psychiatric genetics, and neurodynamics, and the depth comprises thousands of simulation hours and millions of data points, whose realization and analyses are beyond the processing power afforded a standard desktop computer, and whose storage necessitates a robust, dedicated, and capacious centralized system. Off-site data storage and processing often is seen as an attractive option, but, there are always data-transport-speed concerns, access-to-resource concerns, and continued and unexpected, accrued, associated costs, the end result being both unnecessarily long job-completion wait times and high per-job expenses: the proposed CI will effectively reduce job-completion wait times and per-job expenses, and be absolutely essential for timely experimental completions. As SUNY Downstate currently lacks any sort of modern high-performance-computing (HPC) processing- storage offering, the requested CI will form the basis of a such a service. Day-to-day operations and technical administration of the CI will be the responsibility of the PI, with support from the HPC Advisory Committee, which will provide guidance on usage and operational policies. SUNY Downstate is fully committed to immediate and on-going support for the proposed CI, including PI-salary and infrastructure-maintenance funds, and is keen on establishing effective on-site HPC.

项目总结/摘要 该提案是为了获得一个计算仪器（CI），这将使有效的现场存储， 创建和处理当前存储、创建和处理的数据集，几乎完全是在 网站，这是既后勤非最佳和金钱挥霍。所述仪器将直接支持 一组NIH获奖者（也称为主要用户），除了一个附属的二级用户组， 用户（也称为次要用户），他们在SUNY Downstate的持续成功高度依赖于 所建议的设备的能力。 主要用户的研究范围包括计算神经科学，心脏生物物理学， 精神病学遗传学和神经动力学，深度包括数千小时的模拟和数百万小时的模拟。 数据点的实现和分析超出了标准桌面提供的处理能力 计算机，其存储需要一个强大的，专用的和容量大的集中式系统。非现场 数据存储和处理通常被视为一个有吸引力的选择，但是，总是有数据传输速度 关注，获取资源的关注，以及持续的和意外的，应计的，相关的成本，结束 结果是不必要的长时间工作完成等待时间和高额的每项工作费用：拟议的CI将 有效地减少了工作完成等待时间和每项工作的费用， 实验完成。 由于SUNY Downstate目前缺乏任何现代高性能计算（HPC）处理， 存储提供，所请求的CI将形成这种服务的基础。日常业务和技术 CI的管理将由PI负责，并得到HPC咨询委员会的支持， 该指南将提供使用和操作策略方面的指导。SUNY Downstate完全致力于 为拟议的CI提供即时和持续的支持，包括PI工资和基础设施维护资金， 并热衷于建立有效的现场HPC。