Home
What is Virgo ?
People
CBwaves
GPU_inspiral
Computing activities
Analysis activities
Theory
Galery
Publications
Events
Useful links
Getting started
TDK/BSc/MSc/PhD
Archive
Contact
Other GW groups

Research topic proposals for TDK/BSc/MSc/PhD students,

The topic of the research is subject to a free discussion and we are open for any other new ideas. The listed proposals are here to provide an overview of the ongoing work and to give some ideas about the possible research areas. The order is randomly choosen.

Data Analysis

Hunting for gravitational waves produced by short gamma ray bursts.

Gamma Ray Bursts (GRBs) are interesting and still not perfectly understood events of astrophysical origins. It is beleived that a part of the GRBs also produce gravitational waves. The coincident detection of a gravitational wave and its optical counterpart originating from the same GRB in a relatively narrow time window would extremely increase our confidence when the first detection of gravitational wave will happen. Exactly for this reason the search for GW events in time windows in the neighborhood of GRB events has its unquestionable importance. The challenging data analysis methods, ideas and approaches are ideal to demonstrate all the basic principles of gravitational wave search pipelines currently in use.

Implementing and using spinning and eccentric templates for search for compact binary coalescensing objects

The widely used matched template filtering method is shown to be optimal in case of gaussian, stationary detector noise and known gravitational waveform. Despite its high sensitivity, the parametric instability of the algorithm is very high. As a consequance if physical signals are searched for with templates which does not describe them precisely a significant drop in detection statistics could occor, lowering the chances of the first observation. An example for this when a gravitational wave emitted by a spinning neutron star binary inspiral is searched for with non-spinning, circular template banks.The goal of this project is to determine those region of the multi dimensional parameter space (m1,m2,s1,s2,phi1,phi2,epsilon,etc...) where the use of spinning, eccentric gravitational wave templates is absolutely necessary for the reconstruction of the signals. The task involves theoretical calculations, programing, data analysis.

Detecting the gravitational waves produced by the/an ensable of rotating neutron stars in the Galaxy

While the individual signals of galactic neutron stars can be difficult to extract from the detector noise, in some circumstances it could be more feasible to detect the signal of the ensemble of neutron stars in the galaxy. This is a very challenging, innovative and so far untouched region of gravitational wave data analysis, despite its difficulty (or exactly because of it), every new achivement in this particular field will be of great interest ! The task is the practical implementation of the method described in the following paper.

Search for low frequency gravitational waves produced by pulsars.

The gravitational wave signal possibly emitted by rotating neutron stars is relatively simple, but iat the same time is very weak. In order to reconstruct such a signal one has to integrate long stretches of data, taking into account the a.) (sometime) unknown freuency spin-down of neutron star, b.) the doppler modulation due to earth rotation and orbiting around the Sun, c. ) the non-stationarity of the detector noise and a lot other physical and enviromental effects. There are various search algorithms in use. In 'targeted searches' the signals of known pulsars such as for example the Vela and Crab is searched for, while at the 'all sky' search looks for gravitational waves of various frequency and sky location. This later search is computationally bounded, so any speed-up in the search algorithm is of utmost importance ! The candidate could implement new search algorithms, improve existing ones or search for objects in different regions of the parameter space.

Grid, computing

Implementing GW data analysis softwares on the EGEE Grid

The Ligo and Virgo Collaboration are working together, sharing data, knowledge and computational resources. However due to the different computational model and infrastructure available in Ligo and Virgo, the interoperability has to be maintained. The task contains development of interfaces between the EGI (European Grid Infrastructure), OSG (Open Science Grid) and LDG (Ligo Data Grid) using (b)leading edge Grid technology to allow the change of jobs and data between these sites. The establishment and/or management of a user transparent, centrally managed Condor batch pilot pool on the top of EGI to receive and map the DAGs and DAXes of various gravitational wave search pipelines to the EGI infrastructure is also included. The candidate can extend his/her knowledge with a wide range of computing and storage technology including script languages, programing, web services, Grid technologies, virtualisation, batch systems and storage solution.

Developing the GPU accelerated data analysis software, implementing CUDA, OpenCL variants.

The emerging many-core computing technologies and hardwares (such as GPUs) provide unique possibility for the gravitational wave data analysis methods. A significant fraction of the algorithms in use can be massively paralellized or various chunks of data can be paralelly and independently processed in the same time. GPUs prooved to be an excellent solution for this. The task is the identification, benchmarking and efficient implementation of the analysis algorithms on GPU. The software development happens on our test GPU cluster containing 3 nVidia GTX 295, 3 ATI radeon 5970 and 2 Tesla C2050 card, while the full scale analysis can be performed on a much larger GPU cluster. OpenCL and CUDA programing languages are used. Some more details can be found on this link.

Theoretical studies


The RMKI Budapest Virgo Group


This topic: RmiVirgo > WebHome > VirgoStudies
Topic revision: r2 - 2010-12-15 - GergelyDebreczeni
 
This site is powered by the TWiki collaboration platform Powered by Perl This site is powered by the TWiki collaboration platformCopyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback