Mid-Level servers with large memory for CFD and MDO development

点击翻译按钮获取中文摘要

The computational engineering group (GRMIAO) at École Polytechnique de Montréal is one of the largest Computational Fluid Dynamics (CFD) research unit in Canada. Its members are internationally recognized experts with very strong publication records in CFD. Their research is driven by the need for tools to analyze and improve the performance of mechanical products and systems. The flows around airplane wings, in hydraulic turbine components and in medical devices are but a few examples. The GRMIAO research covers the full spectrum of CFD and Multi-Disciplinary Optimization (MDO) activities. Its members are active on fundamental issues such as algorithm development, code prototyping and verification/validation activities. They also develop methodologies to attack fluid mechanics problems of industrial relevance and develop design optimization methods involving one or many disciplines to optimize the performance of these systems. This broad spectrum of activities places special demands on the computational environment. Mid-level servers are requested which are flexible enough to efficiently run commercial software while providing a productive environment for code development and prototyping of flow solvers. This proposal requests the funds to acquire four mid-level servers to replace our old and obsolete servers and provide our group (25-30 graduate students and 7-9 research assistants) with sufficient computing power and memory to maintain international leadership in CFD. The present grant application is the third and last phase of a structured strategy to renew our computational infrastructure. The present RTI is for the central components of our computational infrastructure and is the cap stone to two other sources of funds secured prior to submitting this application. A CFI Team award (Garon as PI) has funds earmarked for acquisition of substantial amounts of high speed disks to store PIV measurement and CFD prediction data files. We have also earmarked $30,000 or about 18% of the total cost of the requested servers through an FQRNT Team equipment grant (Guibault as PI). These actions reflect the critical importance of these machines for our research and the high level of our commitment