.png){kind=icon}
.png){kind=icon}
Return
Parallel implementation of fast multipole method based on JASMIN
.png){kind=icon}
PRE
AI
DOI：10.1007/s11432-011-4181-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Fast multipole method (FMM) may reduce the complexity of N-body problems from O(N (2)) to O(N logN) or O(N). It was applied in problems ranging from electromagnetic scattering to dislocation dynamics. FMM can be divided into two parts: commonness and individuality. A parallel solver of FMM commonly used in various applications has been designed and implemented in JASMIN infrastructure. The solver encapsulates the commonness. Furthermore, it supplies users with abstract interfaces required to implement the individuality with serial mode. The commonness contains distributed storage of multi-levels, intra-level and inter-level data communication, and arrangement of computation, etc. The individuality contains various expansion and translation operators. We give here two applications that have used the solver. Scalability was demonstrated with a parallel efficiency above 80% on 1024 processors.
Keywords:
fast multipole method
N-body problems
JASMIN infrastructure
parallel solver
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
7.6
Papers:
4.9K
Citations:
8.9K