Computing Connected Components on Multiple GPUs
Master thesis
Permanent lenke
https://hdl.handle.net/1956/18842Utgivelsesdato
2018-12-19Metadata
Vis full innførselSamlinger
Sammendrag
We look at multiple GPU programming, using the connected components algorithm as the basis when we look into libraries and runtime environment for multiple GPU programming. We also show that Rem is a fast parallel algorithm, that perform better than previously implemented algorithms.