EXECUTING COMPUTER-INTENSIVE DATABASE USER-DEFINED PROGRAMS ON AN ATTACHED HIGH-PERFORMANCE PARALLEL COMPUTER

摘要

The invention belongs to a kind of system and method, it is used to dispatch and execute for high performance what is once adhered to, the intensive part of the calculating-of the workflow of the data base querying on parallel computing platform. Performance cost, mobile required data and result between database platform and high performance computing platform are executed at for wherein workload, in several ways, such as by for accelerating to be amortized on calculating-intensive calculations parallel computing platform using the parallel and superior hardware performance of particulate, by using the data structures of memory, it is to parallel computing platform, cached data combination between its inquiry to a series of time lags in identical data, so that these inquiries can handle without other data transmission daily expenditure, pass through the replicate data within parallel computing platform, so that the independent parallel subregion that high performance computing platform can be used in a variety of independent inquiries on identical target data set is handled simultaneously. ;A specific embodiment of the invention is related to gene using the graceful algorithm of Smith-water in the Database Systems connected via an Ethernet local area network to a parallel supercomputer and matches with protein sequence for disposing raw information applications.