At DIPSA, we study, evaluate, and design high-performance algorithms and systems for data-intensive applications. Our projects are, mainly, divided into the following categories:
– Optimizing Runtime Environments
– Transprecise Computing
To optimize the time and energy costs of long applications, we study the impacts of reducing accuracy in favour of accelerating the execution. We use the results of these studies to design flexible algorithms and systems for signal and image processing, machine learning and computer vision, and multi-physics simulations. For more details please visit Transprecise Computing Projects.
– Multi-Agent Distributed Computing
– High-Performance Graph Processing
Graph algorithms find usage in different fields of industry, science, and humanities; however, the large and fast-growing size of datasets imposes very long execution times. We study and design high-performance graph algorithms and systems with optimized utilization of hardware resources and optimized memory locality, work-efficiency, load balance . For further information, please visit our High-Performance Graph Processing Projects.