A framework for parallel unstructured grid applications on GPUs
Oxford University Mathematical Institute, Oxford eResearch Centre
SIAM Conference on Parallel Processing for Scientific Computing 2010
PDEs are important in a whole variety of applications. Want a suitable level of abstraction to separate the user’s specification of the app from the details of the parallel implementation. Aim to achieve code longevity and near-optimal performance through re-targeting the back-end to different hardware.
June 5, 2011 by hgpu
