Rafael Ponce (CIEMAT (Centro de Investigaciones Energéticas, Medioambientales y Tecnilógicas)), Miguel Cardenas (inst1), Juan José Rodriguez (inst1), Eusebio Sanchez (inst1), Ignacio Sevilla (inst1),
inst1: Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT), Madrid Spain
In this poster, we have explored the advantages and drawbacks of using
GPUs instead of CPUs in the calculation of a standard 2-point correlation
function (2pcf) algorithm, which is useful for the analysis of Large
Scale Structure of galaxies (LSS).
Taking into account the huge volume of data foreseen in our experiment, our main goal has been to accelerate significantly the analysis codes.
We find that GPUs offer a 100-fold increase in speed with respect to a single CPU without a significant deviation in the results. For comparison's sake, a MPI version was developed as well. Some issues, like code implementation, which arise from using this option are discussed.
Poster in PDF format
Paper ID: P121