Aparna Chandramowlishwaran

Latest

• A CPU:GPU hybrid implementation and model-driven scheduling of the fast multipole method
• Towards a communication optimal fast multipole method and its implications for exascale
• On the limits of GPU acceleration