Code
PetRBF
PetRBF is a portable, extensible toolkit for the fast solution of radial basis function (RBF) interpolation using Gaussian bases. The software is written using the PETSc library, and it is distributed as open source. It can be used either in serial or parallel mode.
Authors: Rio Yokota, Matthew G Knepley, Lorena A Barba
First release: September, 2009.
The first release corresponds to the research code used for production of the tests and scalability studies presented in the preprint, below. Both strong scaling and weak scaling studies have been performed, in different hardware. On the Blue Gene/L of BU’s Center for Computational Science, the code achieved 84% parallel efficiency for 128 processors (strong scaling) and 78% on 1024 processors (weak scaling).
Publications:
-
‣“PetRBF–A parallel O(N) algorithm for radial basis function interpolation”, Rio Yokota, L A Barba, Matthew G Knepley. Submitted.
[arXiv]
Download:
-
‣Please visit our page on Google Code >>
The software library is open source under the conditions of the MIT license.
Contact:
If you have interest or any questions about PetRBF, join the Mailing list and post there —traffic is very low and you can unsubscribe at any time:
http://barbagroup.bu.edu/mailman/listinfo/petrbf
