next up previous contents index
Next: Conclusion Up: Computational technologies for astrophysics Previous: Grid Computing   Contents   Index

MARA: a parallel code for light curves

MARA (Multi light curve Analysis of Rotator Activity) is a code for modelling sequences of optical light curves of single or close binary stars with surface brightness inhomogeneities. The code takes into account ellipsoidicity, gravity-darkening and reflection effects as well as a distribution of cool spots in order to fit the light curve of a close binary system. The determination of the spot surface distribution is an ill-posed problem solved by introducing the Maximum Entropy and Tikhonov regularizations. A sequence of optical light curves of a close binary system can be analysed as a single data set to derive the best values of the photometric parameters, i.e., the luminosity ratio, the fractionary radii of the components and the inclination of the orbital plane of the system. If the light curve sequence spans an entire activity cycle, the systematic errors introduced by the cool spots in the determination of the photometric parameters can be significantly reduced. The code was developed using the MPI library.
Mara also can be freely used with the portal Astrocomp.
Figure 1.34: The Astrocomp mask to upload the set of parameters for each light curve analysis.
\begin{figure}\centerline{\psfig{file=becciani/P1-3-1-f2.ps,height=3.7cm}}\end{figure}
Figure 1.35: The visualization of a postscript file showing the best fit and the Mercator maps of the components of a close binary system for the Maximum Entropy solution (upper panels) and the best fit and the Mercator star maps for the Tikhonov model (lower panels), respectively.
\begin{figure}\centerline{\psfig{file=becciani/P1-3-1-f3.ps,height=3.7cm}}\end{figure}
The portal is a user-friendly application which allows to submit a job in a grid of computing systems. In order to use MARA through Astrocomp, a user should select the MARA code section and access the pages of the code that allow him to create all the input parameter files, upload his own light curve sequence, and create a job to submit the run in the available MPP systems (Fig. 1.34). The user can monitor the job status and download the output data files at the end of the run. Moreover, the light curve best fits and the corresponding spot maps can be visualized as postscript or jpeg files by means of a dedicated utility based on IDL (cf. Fig. 1.35).
Astrocomp makes it possible to use MARA without downloading and installing the code and provides some free computing time on MPP systems to run the code.


next up previous contents index
Next: Conclusion Up: Computational technologies for astrophysics Previous: Grid Computing   Contents   Index
Innocenza Busa' 2005-11-14