A multithreaded CPU & GPU-based facet imager

Authors: Benjamin Hugo, Oleg Smirnov, Cyril Tasse, James Gain

Copyright (C) 2014-2015 Rhodes Centre for Radio Astronomy Techniques and Technologies, Department of Physics and Electronics, Rhodes University, Artillery Road P O Box 94, Grahamstown, 6140, Eastern Cape, South Africa

Copyright (C) 2014-2015 Department of Computer Science, University of Cape Town, 18 University Avenue, University of Cape Town, Rondebosch, Cape Town, South Africa

Dependencies:

1. Python 2.7.6
2. casacore
3. numpy
4. pyrap
5. matplotlib
6. pyfits
7. gtk3.0 with engines and python dev packages and pycairo (and maybe any cairo-dev packages). Important: install package python-gi-cairo, otherwise the gtk drawing area's on draw is never called for some truely bizare reason.
8. (to edit gui: glade for gtk 3.0 is needed)
9. cmake
10. The GNU C++ compiler >= 4.8
11. libboost-all and libboost-dev
12. OpenMP
13. CUDA toolkit (nvcc,nvprof) >=5.0
14. Montage (available: http://montage.ipac.caltech.edu/index.html). Montage's bin folder must be in the PATH
15. CfitsIO
16. WcsLib

Build instructions (outputs CPU and GPU single and double precision libraries and a python wrapper for these)

• run: python setup.py install --user (this will install into the python user directory)

Run instructions

• Navigate to the bullseye/bullseye directory
• "./bullseye.py --help" to display a full list of options
• Enjoy

Toy GUI is available (proof of concept demonstrator for targeted faceting)

• "./bullseye_frontend.py"