Distributed Computing on GPUs

Welcome to the Distributed Computing on GPUs Community!

This community covers the applications of distributed computing on GPUs for academia and industries. Group members are invited to post links to recent publications, links to calls for participation for journals and conferences, and exchange example codes and ideas. Particular examples of research areas of interest include:

Volunteer computing on GPUs (global)
Distributed computing on GPUs (local)
Computational biology applications of distributed computing
Physics applications of distributed computing
Problems related to running on GPU grids
Production-level GPU parallelization of widely used algorithms and tools.

As community interests manifest themselves, new forum topics can be created to focus participation. The intention is to make our community a focal point for quality discussions on problems and solutions that can be used as a searchable knowledge base to benefit everyone interested in the use of GPUs for distributed computing.

Gianni De Fabritiis, University of Pompeu Fabra, http://multiscalelab.org.

Featured Stories and Papers

Experiences with the Sparse Matrix-Vector Multiplication on a Many-core Processor (IEEE)

Abstract:

_______________________________________________________________________________________________

Parallel Acceleration of SAM Algorithm and Performance Analysis (IEEE)

Abstract:

_______________________________________________________________________________________________

Work Stealing on Hybrid Architectures (IEEE)

Abstract:

Parallel computing systems have been based on multicore CPUs and specialized coprocessors, like GPUs. Work-stealing is a scheduling technique that has been used to distribute and redistribute the workload among resources in an efficient way. This work aims to propose, implement and validate a scheduling approach based on work stealing in parallel systems with CPUs and GPUs simultaneously. Results show that our approach, called WORMS, presents competitive performance when compared to reference tool for multicore CPUs (Cilk). In hybrid scenario, WORMS with multicore+GPU outperforms WORMS and Cilk with multicore only and also the GPU reference tool (Thrust).

_______________________________________________________________________________________________