Pipelining the Fast Multipole Method over a Runtime System

  title={Pipelining the Fast Multipole Method over a Runtime System},
  author={Emmanuel Agullo and B{\'e}renger Bramas and Olivier Coulaud and Eric F Darve and Matthias Messner and Toru Takahashi},
Fast Multipole Method (FMM) are a fundamental operation for the simulation of many physical problems. The high performance design of such methods usually requires to carefully tune the algorithm for both the targeted physics and the hardware. In this paper, we propose a new approach that achieves high performance across architectures. Our method consists of expressing the FMM algorithm as a task flow and employing a state-of-the-art runtime system, StarPU, in order to process the tasks on the… CONTINUE READING


Publications referenced by this paper.
Showing 1-10 of 32 references

Data-Driven Execution of Fast Multipole Methods

Concurrency and Computation: Practice and Experience • 2014
View 1 Excerpt

Flexible Development of Dense Linear Algebra Algorithms on Massively Parallel Architectures with DPLASMA

2011 IEEE International Symposium on Parallel and Distributed Processing Workshops and Phd Forum • 2011
View 1 Excerpt

LU factorization for accelerator-based systems

2011 9th IEEE/ACS International Conference on Computer Systems and Applications (AICCSA) • 2011
View 1 Excerpt

QR Factorization on a Multicore Node Enhanced with Multiple GPU Accelerators

2011 IEEE International Parallel & Distributed Processing Symposium • 2011
View 2 Excerpts

Scalable fast multipole methods on distributed heterogeneous architectures

2011 International Conference for High Performance Computing, Networking, Storage and Analysis (SC) • 2011
View 2 Excerpts

Similar Papers

Loading similar papers…