Technical Report CSI-0042

Algebraic Approach to Compositing Hypercubic Data Transforms in Memory and Abstract Storage

K. A. Hawick

Archived: 2015

Abstract

Specifying and transforming multi dimensional hypercubic data are common to many programming systems and the goal of abstracting out dimensionality remains difficult to implement portably. We develop an algebra for specifying, transforming and compositing hypercubic data in memory or other storage media based on often occurring operations including: shifting; translating; reducing and expanding dimensionality; padding; culling; reflecting; and reordering dimensions, using both fixed and periodic or wrapping boundary conditions. We develop both an abstract specification and test implementations as well as discussing how such a system aids implementation on distributed memory systems such as parallel computers and clusters with array or node-distributed memory systems. We illustrate our k-indexing nomenclature with examples in hypercubic system simulation models.

Keywords: hypercube; data transforms; data algebra; k-indexing; multi-dimensional data; platform independence

Full Document Text: Not yet available.

Citation Information: BiBTeX database for CSI Notes.

BiBTeX reference:

@TechReport{CSI-0042,
        Title = {Algebraic Approach to Compositing Hypercubic Data Transforms in Memory and Abstract Storage},
        Author = {K. A. Hawick},
        Institution = {Computer Science, University of Hull},
        Year = {2015},
        Address = {Cottingham Road, Hull, HU6 7RX, UK},
        Month = {September},
        Number = {CSI-0042},
        Type = {CSI},
        Keywords = {hypercube; data transforms; data algebra; k-indexing; multi-dimensional data; platform independence},
        Owner = {kahawick},
        Timestamp = {2017.08.23}
}