Reducing average access time of a parallel memory in a database environment by data permutation
Paul T. Hulina, Ali R. Hurson · 2002
Optimum use of existing memory bandwidth is needed to reduce the access gap and hence to improve the performance of computer systems in handling I/O-bound applications. For this goal the authors introduce ideas based on data permutation of the tuples in a relational database, which allows parallel access to the database tuples by efficient utilization of the existing memory bandwidth. As a result, several keys can be accessed simultaneously and the validity of several tuples in parallel can be investigated.>