HPC for neutrino detection (with Nikhef)

From Master Projects
Jump to: navigation, search


About HPC for neutrino detection (with Nikhef)

  • This project has not yet been fulfilled.
  • This project fits in the following Bachelor programs: {{#arraymap:|, |xXx|bachelorproject within::xXx|,}}
  • This project fits in the following masterareas: {{#arraymap:Parallel and Distributed Computer Systems, High Performance Distributed Computing|, |xXx|project within::xXx|,}}


Description

The ANTARES detector is a neutrino telescope located in the Mediterranean Sea, 40 km off the coast of Toulon (France). The scientific goal of the ANTARES neutrino telescope is to detect high-energy cosmic neutrinos. These high-energy neutrinos are believed to be produced by astrophysical sources, such as gamma-ray bursts, microquasars and supernova remnants. Neutrinos allow observing what cannot be observed with other detectors that study astrophysical sources, such as astronomy experiments and cosmic ray experiments. This is due to the neutrino characteristics, which enable the neutrino to escape from regions that other radiation cannot escape, and to come from cosmological distances. Neutrino astronomy will thus open up a new window on the Universe.

The Nikhef contribution to ANTARES has been significant from the beginning, and one of its showpieces is the ANTARES data filter, implemented in software. The data filter separates the physics signal from the background in real time, reducing the data stream from 1 GB/s to 10 kB/s. This data filter has been applied during data collection since the beginning of the experiment, it runs 24/7, and has high signal efficiency and high purity.

The successor of ANTARES, the KM3NeT neutrino telescope, is currently in the design phase, and is expected to exceed the size of ANTARES by two orders of magnitude. It is planned that also for this telescope the data will be filtered in real-time. However, it is unclear how the existing data filter can be used for KM3NeT, as some stages in the data filtering.

In this project, we investigate the performance requirements for the data filter in the context of ANTARES and KM3NeT. We aim to design and implement a first prototype of the data filter using massive parallelism on GPUs and multi-core CPUs, and analyze its performance and portability.

This project will require a moderate programming workload, as well as significant experimental work.