Light curve fitting on heterogeneous computers
| dc.contributor.advisor | Fussell, Donald | en |
| dc.creator | Luecke, Kevin J.P. | en |
| dc.date.accessioned | 2013-08-16T21:01:02Z | en |
| dc.date.available | 2013-08-16T21:01:02Z | en |
| dc.date.issued | 2013 | en |
| dc.description.abstract | The field of computer science is one of the most recently created and rapidly expanding sciences. In the years since its inception in the mid 1900s, it has revolutionized the other sciences, businesses, and our personal lives. With increasing capabilities, however, comes increasing complexity. In order to deal with this complexity, computer scientists are creating new unconventional hardware which is becoming more and more parallel. With different hardware on your computer it becomes harder to effectively use all this computational power. GAMA (GPU And Multi-core Aware) addresses this problem by providing a runtime system that determines the amount of work that should be sent to each available resource dynamically. In order to demonstrate the abilities of this approach, a sample program with real applications in astrophysics has been implemented. This program, Lcfit Theta, attempts to find the size of convection zones in white dwarf stars. Most white dwarfs pulsate or oscillates in temperature and density during their life cycles. The oscillations can be seen from earth telescopes which record a star’s brightness as a function of time in a graph known as a ’light-curve’. These light-curves can be reproduced fairly well by a simple linear sum of sine functions corresponding to the frequencies and amplitudes at which it the star is oscillating. However, in stars where a convection zone is present at the surface, non-linear combinations of these sine functions appear in the light-curve. Lcfit Theta uses a model of these white dwarfs proposed by Montgomery which includes the effects of these convective regions, to generate synthetic light curves (9). By comparing these synthetic light curves to the measured stellar data, we can find the input parameters to Mongomery’s model which create a synthetic light curve that matches the data. One of these ’fitted’ input parameters is the size of the convection zone. This application is the focus of this thesis. | en |
| dc.description.department | Computer Science | |
| dc.identifier.uri | http://hdl.handle.net/2152/21070 | en |
| dc.language.iso | eng | en |
| dc.subject | light curve fitting | en |
| dc.subject | heterogeneous computing | en |
| dc.subject | GPU | en |
| dc.subject | astronomy | en |
| dc.title | Light curve fitting on heterogeneous computers | en |
| dc.type | Thesis | en |
Access full-text files
Original bundle
1 - 1 of 1