Skip to content

Releases: OxfordSKA/OSKAR

OSKAR 2.8.3

26 May 20:10
Compare
Choose a tag to compare

Documentation

Change Log

  • Add option to use HARP beam library.
  • Fix problem trying to write Measurement Sets with tile sizes > 2 GiB.

OSKAR 2.8.2

19 Feb 15:35
Compare
Choose a tag to compare

Documentation

Change Log

  • Fix to make reference counter for HDF5 file handles thread-safe.

OSKAR 2.8.1

17 Feb 23:44
Compare
Choose a tag to compare

Documentation

Change Log

  • Allow HDF5 gain tables to be used in station and telescope models.

OSKAR 2.8.0

23 Nov 23:56
Compare
Choose a tag to compare

Documentation

Change Log

  • Fix read of ionosphere pixel size from FITS header.
  • Fix Measurement Set ANTENNA table to use absolute rather than relative coordinates.
  • Fix to ensure noise random numbers are different per channel.
  • Add PHASED_ARRAY table to Measurement Sets.
  • Use Ludwig-3 (instead of theta/phi) system to evaluate antenna responses.
  • Add capability to use different element data for X and Y polarisations.
  • Add capability to use custom Stokes parameters for beam pattern test source.
  • Allow visibility blocks to be tiled in frequency as well as time.
  • Add limited drift-scan mode (point sources only; no time-smearing).
  • Add capability to specify station type mapping for beam duplication, if using this mode.
  • Add option to treat phase screens as isoplanatic.
  • Add options to taper weights used by the imager.

OSKAR 2.7.6

20 Jan 13:54
Compare
Choose a tag to compare

Documentation

Example Data

This archive contains the input files used to run the example simulation.

Binary Packages

Singularity image for Linux

To use OSKAR with Singularity on Linux, download the pre-built Singularity image file. This contains the OSKAR command line applications as well as a basic Python environment, which includes the OSKAR Python interface.

macOS

Installation

To install the OSKAR package on macOS, download and open the disk image (DMG) file and drag the OSKAR.app bundle to your /Applications folder. After installation, double-click the OSKAR.app bundle to launch the GUI and set symbolic links to the applications in /usr/local/bin.

Uninstallation

To remove the OSKAR package on macOS, delete the OSKAR.app bundle from your /Applications folder, and delete symbolic links to the applications by typing rm -f /usr/local/bin/oskar* from a terminal prompt.

Windows

Installation

To install the OSKAR package on 64-bit Windows, download and run the installer executable and follow the on-screen instructions. After installation, the GUI can be launched using the shortcut on the Desktop or Start Menu.

If using a GPU on Windows, also be sure to install the latest driver for it.

Note that casacore is not available for Windows, so Measurement Sets cannot be used.

Uninstallation

To remove the OSKAR package on Windows, uninstall it from the list of software in Control Panel in the usual way.

Change Log

2.7.6

  • Fixed load of TEC screen settings.

2.7.5

  • Fixed crash when manually specifying which CUDA device to use.
  • Fixed display of timing information when using GPU devices.
  • Fixed compilation and installation on versions of Ubuntu >= 18.04.
  • Added optimised versions of cross-correlation kernels.
  • Added GPU-capable gridder to imager.
  • Added experimental OpenCL support.
  • Added option to specify cable length errors.
  • Added capability to use externally-generated spherical wave element patterns.
  • Added option to use an externally-generated ionospheric TEC screen.
  • Added option to ignore baseline W-components.
  • Added option to evaluate station beams at specified horizon coordinates.
  • Refuse to load FITS images with an unknown pixel size.

OSKAR 2.7.5

OSKAR 2.7.0

31 Oct 22:01
Compare
Choose a tag to compare

Documentation

Example Data

This archive contains the input files used to run the example simulation.

Binary Packages

macOS

Installation

To install the OSKAR package on macOS, download and open the disk image (DMG) file and drag the OSKAR.app bundle to your /Applications folder. After installation, double-click the OSKAR.app bundle to launch the GUI and set symbolic links to the applications in /usr/local/bin.

If using a GPU on macOS, also be sure to install the latest NVIDIA CUDA driver from
http://www.nvidia.com/object/mac-driver-archive.html

Uninstallation

To remove the OSKAR package on macOS, delete the OSKAR.app bundle from your /Applications folder, and delete symbolic links to the applications by typing rm -f /usr/local/bin/oskar* from a terminal prompt.

Windows

Installation

To install the OSKAR package on 64-bit Windows, download and run the installer executable and follow the on-screen instructions. After installation, the GUI can be launched using the shortcut on the Desktop or Start Menu.

If using a GPU on Windows, also be sure to install the latest NVIDIA driver from
https://www.geforce.com/drivers

Note that casacore is not available for Windows, so Measurement Sets cannot be used.

Uninstallation

To remove the OSKAR package on Windows, uninstall it from the list of software in Control Panel in the usual way.

Change Log

  • Removed telescope longitude, latitude and altitude from settings file.
    These are now specified using "position.txt" in the telescope model.
  • Made CUDA an optional dependency, and added option to specify whether
    or not to use GPUs for simulations. This makes it possible to run
    inside a virtual machine if necessary.
  • Removed Qt dependency from all command line applications, and updated
    the GUI to use Qt 5.
  • Changed GUI to only display settings for the selected application and to
    remember the last settings file used for each application.
  • Added option to change current working directory from GUI.
  • Removed oskar_settings_set, as it was unable to check whether settings
    were being set correctly. Settings files can now be updated using
    the --set command line option of the required application.
  • Added option to specify the number of compute devices used for
    simulations (CPU cores and/or GPU cards).
  • Added Python interface.
  • Re-wrote oskar_imager application to allow use of multiple GPUs
    when using a DFT. Visibility data are now loaded in blocks to reduce
    memory usage.
  • Added option to use a 3D DFT to make images.
  • Added option to use gridding and FFT to make images that are
    consistent with those from CASA.
  • Added option to use W-projection to make images that are
    consistent with those from CASA.
  • Added option to use natural, radial or uniform visibility weighting
    when making images.
  • Added option to imager to filter input by UV baseline length,
    in wavelengths.
  • Fix to ensure pixel coordinates are consistent between
    the DFT and FFT algorithms.
  • Added options to specify grid kernel parameters.
  • Added option to perform imaging in single or double precision,
    regardless of the precision of the input data.
  • Added capability to use multiple visibility files as input to the imager,
    with option to scale image normalisation accordingly.
  • Allowed visibility data to be read from a specified column of a
    Measurement Set.
  • Allowed negative source fluxes.
  • Allowed FITS images to be read if they have a diagonal CD matrix
    instead of CDELT* values.

OSKAR 2.6.1

22 Sep 12:52
Compare
Choose a tag to compare

OSKAR version 2.6.1

OSKAR 2.6.0

22 Sep 12:54
Compare
Choose a tag to compare

OSKAR version 2.6.0

...