Algorithmic, data access and data partitioning optimizations for multidimensional data processing

In this paper we present several optimization techniques for processing multidimensional data. The presented techniques belong to three categories: algorithmic techniques, data access techniques and data partitioning techniques. The algorithmic techniques improve the total number of operations performed during the data processing stage. The data access techniques improve the way the data is accessed during the processing stage and aim towards efficient cache usage. The data partitioning techniques split the data into logically independent groups, which can then be processed in parallel, on multiple cores, processors or machines.