Accelerating In-Memory Graph Database traversal using GPGPUS
University of Virginia, High Performance Computing Laboratory
University of Virginia, 2012
The paper aims to provide a comparitive analysis on the performance of in memory databases as opposed to a customised graph database written ground up whose joins(searches) are performed on a GPGPU. This is done primarily to serve as a proof of concept on how databases that are represented as graphs can benefit by fostering the raw parallell processing power of GPGPUS.
May 25, 2012 by hgpu
