Performance Analysis of an OpenACC Many-Core Pulsar Pipeline

From Master Projects
Jump to: navigation, search


About Performance Analysis of an OpenACC Many-Core Pulsar Pipeline

  • 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:High Performance Distributed Computing, Multimedia, Parallel and Distributed Computer Systems|, |xXx|project within::xXx|,}}


Description

Pulsars are rapidly rotating and highly massive neutron stars which signal is received on Earth only at periodic intervals. They are not just fascinating objects, but also scientifically important: in the last 50 years, two Nobel prizes in physics have been awarded for pulsars related discoveries. However, finding pulsars is difficult and a computationally intensive task.

In an effort to help astronomers and increase the rate of scientific discoveries, we developed a many-core accelerated pulsar pipeline that can run on multiple devices: multi-core CPUs, AMD and NVIDIA GPUs and even the Intel Xeon Phi. The goal of this masters project is to compare the performance of our pipeline with an OpenACC accelerated one, that you will have to design and implement. OpenACC makes the development of many-core accelerated software, and the porting of preexisting sequential code to many-core devices, much faster, but how does it compare with highly tuned scientific software? This is the question that you will have to answer.