skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Fixed-Rate Compressed Floating-Point Arrays

Publication Date:
Research Org.:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Org.:
OSTI Identifier:
Report Number(s):
Journal ID: ISSN 1077-2626
DOE Contract Number:
Resource Type:
Journal Article
Resource Relation:
Journal Name: IEEE Transactions on Visualization and Computer Graphics; Journal Volume: 20; Journal Issue: 12
Country of Publication:
United States

Citation Formats

Lindstrom, Peter. Fixed-Rate Compressed Floating-Point Arrays. United States: N. p., 2014. Web. doi:10.1109/TVCG.2014.2346458.
Lindstrom, Peter. Fixed-Rate Compressed Floating-Point Arrays. United States. doi:10.1109/TVCG.2014.2346458.
Lindstrom, Peter. 2014. "Fixed-Rate Compressed Floating-Point Arrays". United States. doi:10.1109/TVCG.2014.2346458.
title = {Fixed-Rate Compressed Floating-Point Arrays},
author = {Lindstrom, Peter},
abstractNote = {},
doi = {10.1109/TVCG.2014.2346458},
journal = {IEEE Transactions on Visualization and Computer Graphics},
number = 12,
volume = 20,
place = {United States},
year = 2014,
month =
  • ZFP is a library for lossy compression of single- and double-precision floating-point data. One of the unique features of ZFP is its support for fixed-rate compression, which enables random read and write access at the granularity of small blocks of values. Using a C++ interface, this allows declaring compressed arrays (1D, 2D, and 3D arrays are supported) that through operator overloading can be treated just like conventional, uncompressed arrays, but which allow the user to specify the exact number of bits to allocate to the array. ZFP also has variable-rate fixed-precision and fixed-accuracy modes, which allow the user to specifymore » a tolerance on the relative or absolute error.« less
  • Rotating electric fields are used to compress electron plasmas confined in a Penning-Malmberg trap. Bifurcation and hysteresis are observed between low-density and high-density steady states as a function of the applied electric field amplitude and frequency. These observations are explained in terms of torque-balanced fixed points using a simple model of the torques on the plasma. Perturbation experiments near the high-density fixed point are used to determine the magnitude, frequency, and voltage dependence of the drive torque. The broader implications of these results are discussed.
  • A microprogrammed multiple data stream digital signal processor is introduced. This floating-point processor is capable of implementing optimum Wiener filtering of signals, in general, and images in particular. Generalised spectral analysis transforms such as Fourier, Walsh, Hadamard, and generalised Walsh are efficiently implemented in a bit-slice microprocessor-based architecture. In this architecture, a microprogrammed sequencing section directly controls a central floating-point signal processing unit. Throughout, computations are performed on pipelined multiple complex data streams. 12 references.