Performance analysis of incremental data partitioning data mining algorithm for Relational Database on multicore processor

摘要

In this paper, we have proposed and? implemented? parallel Online Analytical Processing for the relational database (MySQL) on the cores of the processors with a heterogeneous incremental partition for the? database record in such way that the processor cores can be maximum utilize for the computation and the computation time for large database can be minimized. ? The database is partitioned horizontally in a heterogeneous size, and the size of the record allocated to a particular core is increased with a multiple of the increment variable and this partitioned database record is allocated to the particular core of the processor asynchronously. In this way a processor with small data record will complete its job first and then a new job is allocated to it, until the processor core with the larger size of data record? completed its job. ? This implementation is implemented with PHP, SQL and MySQL and this is new implementation in the data mining and data clustering with multicore processor. Through this implementation, we will prove that the computation time with this partition is better than the previous parallel implementations.?