2010
DOI: 10.1016/j.jmmm.2009.09.033
|View full text |Cite
|
Sign up to set email alerts
|

Simple recursive implementation of fast multipole method

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
28
0

Year Published

2010
2010
2022
2022

Publication Types

Select...
7
1

Relationship

0
8

Authors

Journals

citations
Cited by 23 publications
(28 citation statements)
references
References 22 publications
(31 reference statements)
0
28
0
Order By: Relevance
“…Finally, there is another possibility of vortex conversion into the uniform in-plane state, not captured by the expression (13). At very small thicknesses the vortex may become unstable with respect to lateral shift.…”
mentioning
confidence: 99%
See 1 more Smart Citation
“…Finally, there is another possibility of vortex conversion into the uniform in-plane state, not captured by the expression (13). At very small thicknesses the vortex may become unstable with respect to lateral shift.…”
mentioning
confidence: 99%
“…Magnetism@home project numerically computed magnetostatic energy (including that of volume, face and side magnetic charges, as well as their mutual interaction) of magnetization distributions (1) in a 4-dimensional unithypercube with normalized coordinates g = g/(1 + g), c = c/(1 + c), a = a/(1 + a), p = 1/p, which covers all the magnetization configurations it describes. The densities of magnetic charges were computed analytically, while the magnetostatic energy itself was evaluated using fast multipole method [13] on a dense 50000 non-uniform finite elements mesh, perfectly covering the circular cylinder's boundary, with exact analytical treatment for Z dependence of demagnetizing field. This computation (including preliminary testing runs) took about a year to complete, using idle processor time of a several tens of thousands of computers on the Internet, communicating via Berkeley Open Infrastructure for Network Computing (BOINC) protocol.…”
mentioning
confidence: 99%
“…Over the 25 years of its existence, the FMM underwent several changes and improvements in order to reduce the complexity coefficients, make it practical in 3D, apply it to various kernels, using different expansions, and using different data structures and spatial adaptivity to reduce the amount of work needed [11][12][13]. Also, the implementation of the FMM saw many advances from sequential to massively parallel, and optimizations on heterogeneous CPU-GPU platforms [14][15][16].…”
Section: The Fast Multipole Methodsmentioning
confidence: 99%
“…From a source point (or field point, resp.) perspective the truncation error is proportional to (α/2) p+1 where α is the opening angle 2r/d with r being the diameter of source (or field) cell and d the distance between centers [21]. Plugging the expansion into the expression for the j-th volume source (eq.…”
Section: Source Expansionmentioning
confidence: 99%