2011 | OriginalPaper | Buchkapitel
Multi-core CPU Based Parallel Cube Algorithms
verfasst von : Guoliang Zhou, Han Zhang
Erschienen in: Advanced Research on Computer Science and Information Engineering
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
In recent years, computer hardware technology has greatly developed especially large memory and multi-core, but algorithm efficiency is not beneficial from the development of hardware. The fundamental reason is that there is insufficient utilizing CPU cache, as well as the limitations of single-thread programming. In the field of data warehousing and OLAP, data cube computing is an important and time-consuming operation, how to improve efficiency of data cube calculation is continuing to pursue goals. Based on the characteristics of modern CPU, we have proposed two parallel algorithms TASK_PMW and DATA_SSMW, TASK_PMW is task-based division of the parallel algorithm, each CPU core is responsible for one
Cuboid
; DATA_SSMW is data partition, and scanned sharing raw data, ensure load balancing, has good scalability and high efficient. Through experiments on dual-core CPU, TASK_PMW improve 1/3, DATA_SSMW 2/3 than the original algorithm.