Facility for Integrated Experimental and Computational Turbulence Research

Abstract : The funds from the subject grant and the matching funds from the University of Illinois were used to purchase the equipment listed in tile table from the indicated vendors. As proposed, this equipment includes PlY data acquisition equipment (items 1-7) and computing equipinent (items 8-l9) used in experimental data analysis and numerical simulation. Also included are several added items to enable improved performance on the primary missions of the facility. These include a high speed (100 Mbit/sec) switch card for the in building net-work (item 16 for improved and isolated data transfer from the fluids lab to the computing facility) and a l44Gbyte RAID disk array (items 14 and 15 for high-speed and high capacity disk I/O for post-processing data analysis). The primary computing platform was selected to be a 5 node, 20 processor IBM SP system with a total of 12 Gigabytes of memory. Through a series of benchmarks using turbulence simulation codes, it was determined that the selected IBM system was superior in cost/performance to the proposed SGl system, or SGI's current offereing, for the applications of interest. The architecture is also more advantageous for use as a real-time PIV data analysis system, though it is marginally less convenient for large scale simulation. For these reasons the IBM system was selected over the originally proposed 16 processor SGI system.