MRI:"Acquisition" of Rice Computational Research Cluster (RCRC)

点击翻译按钮获取中文摘要

This project, acquiring a large scale (512 processors) Linux computational cluster, provides a research instrument for computing intensive research and education projects that can help advance knowledge and understanding in a diversity of fields, while enabling technologies in systems and compilers, multiscale computational biology, biophysics, bioengineering, earth sciences, and psychology. Sharing benefits of terascale computational power with 32 investigators and 50 NSF-funded projects, this advanced high performance computing instrument, also promotes cross-disciplinary collaboration between tool-building teams and scientific computing teams in bio-computing, geodynamics, and magnetic resonance analysis. Software tool-building addresses performance issues, caching and clustering schemes for dynamic web sites, and designing improved TCP/IP interfaces. Eliminating the need for many, separate ad-hoc clusters in different departments, the cluster services, among others, the following specific computational research projects:* Studying software and performance issues for heterogeneous Grid systems.* Caching and clustering schemes for web sites with dynamic content,* Designing improved TCP/IP interfaces for cluster computing,* Multiscale computational biology, biophysics, and bioengineering,* Modeling of tissues, cells, and cell populations,* Modeling of flow and hemolysis in implantable blood pumps,* Computer-aided drug discovery,* Algorithms for PDE constrained optimization,* Novel algorithms for large-scale numerical linear algebra,* Using computational fluid dynamics to model thermal convection in the Earth's mantle,* Particle dynamic simulations of earth processes,* Multiscale seismological simulation,* Analysis of functional MRI (fMRI) data,* Source modeling of EEG and ERP signals, and* Modeling of visual ideal observers.Broader Impact: The facility, benefiting more than 100 PhD students, 50 undergraduates, and 30 Postdoctoral fellows, and directly impacting 32 investigators, will be used in classes at the undergraduate and graduate levels and also as a recruiting tool through existing alliance activity (AGEP: Alliance for the Graduate Education and the Professoriate and a "Spend a Summer with a Scientist" outreach program). Furthermore, the work exhibits strong emphasis on learning, communication, and interaction across disciplines.