MRI: Acquisition of an Infrastructure for Prototyping Next-Generation Algorithms for Large-Scale Visualization, Data Processing and Analysis

提案编号：12- 29185 PI（s）：Silva，Claudio; Freire，Juliana; Iacono，John; Suel，Torsten; Vo，Huy机构：纽约理工大学标题： MRI/采集：大规模可视化、数据处理和分析的下一代算法原型设计项目：该项目获得了一种用于大规模可视化、数据处理和分析的下一代算法原型设计的仪器，促进了利用不同硬件配置的下一代算法和系统的开发和评估。它包括一台具有384个内核和4个Tesla M2070 GPU的SGI UV 1000计算机形式的计算资源，以及一个由32个节点（共384个内核）组成的小型群集，每个节点具有2个Tesla M2070 GPU;以及固态存储卡和Isilon存储群集形式的磁盘资源。该工具支持各个领域的项目。计算机科学项目包括-数据流架构，-计算的理论模型，- Web搜索，-可视化和-数据挖掘;而应用领域的研究项目（与外部合作者进行）涉及-野火模拟，-鸟类迁移模拟和-气候科学。更广泛的影响：该团队计划将作为该项目的一部分开发的软件作为开源发布，这可能会产生更广泛的影响。研究生和本科生参与该项目提供了指导机会。

Proposal #: 12-29185PI(s): Silva, Claudio; Freire, Juliana; Iacono, John; Suel, Torsten; Vo, HuyInstitution: Polytechnic University of New YorkTitle: MRI/Acq.: Prototyping Next-Generation Algorithms for Large-scale Visualization, Data Processing and AnalysisProject Proposed:This project, acquiring an instrument for prototyping the next-generation algorithms for large-scale visualization, data processing and analysis, facilitates the development and evaluation of next-generation algorithms and systems that leverage different hardware configurations. It consists of computational resources in the form of an SGI UV 1000 machine with 384 cores and four Tesla M2070 GPUs, and a small cluster composed of 32 nodes (384 cores total), each with two Tesla M2070 GPUs; and disk resources in the form of solid-state storage cards and an Isilon Storage Cluster. The instrumentation supports projects in various domains. The computer science projects include - Dataflow architectures, - Theoretical models of computation, - Web search, - Visualization, and - Data mining; While the application area research projects (undertaken with external collaborators) relate to- Wildfire simulation, - Bird migration simulation, and - Climate science. Broader Impacts: The team plans to release the software developed as part of this project as open source, which carries potential for significant broader impact. The involvement of graduate and undergraduate students in the project provides mentoring opportunities.