Wavelet-Based Compression of Volumetric CFD Data Sets

One of the major pitfalls of storing “raw” simulation results lies in the implicit and redundant manner in which it represents the flow physics. Thus transforming the large “raw” into compact feature- or structure-based data could help overcome the I/O bottleneck. Several compression techniques have already been proposed to tackle this problem. Yet, most of these so-called lossless compressors either solely consist of dictionary encoders, which merely act upon the statistical redundancies in the underlying binary data structure, or use a preceding predictor stage to decorrelate intrinsic spatial redundancies. Efforts have already been made to adapt image compression standards like the JPEG codec to floating-point arrays. However, most of these algorithms rely on the discrete cosine transform which offers inferior compression performance when compared to the discrete wavelet transform. We therefore demonstrate the viability of a wavelet-based compression scheme for large-scale numerical datasets.