A Cost-effective, COTS-based and Standards-based, Scalable Message-Passing Multicomputer

In the Phase I effort, in order to build a cost-effective multicomputer, we will refine our proposed system-level design, specify the appropriate combination components which meet the desired goals, procure and assemble the hardware components and operating system, install the other systems software with which to perform parallel programming, and demonstrate the multicomputer. The key concept behind our proposal idea is the utilization of only commodity components - no custom hardware or systems software will be created. This concept is beneficial because it yields modularity and flexibility in the system-level design, and creates an evolutionary path of upgrades and increased capability with respect to the implementation. Because the result of SBIR research and development will be an affordable, high-performance multicomputer, many more users than before will be able to acquire a dedicated machine to solve their problems.