Lorena A. Barba group


Submitted: Inexact Krylov iterations and relaxation strategies with fast-multipole boundary element method

sketch of quadrature-points clustering on boundary element method

Submitted: 19 June 2015   This paper presents a method to algorithmically accelerate boundary element methods. The fast multipole method (FMM) is already used in this context to obtain O(N) computational complexity while solving a dense linear system. Using relaxed-accuracy matrix-vector multiplications in a Krylov solver, we can achieve speed-ups in the order of 4x... Continue »

How will the fast-multipole method fare in the exascale era?

Roofline model of current hardware and computational intensity of various representative algorithms.

SIAM News (Vol. 46, issue 6) has published on the front page our piece about the future of algorithms in the exascale era. This piece was written by invitation on the aftermath of a very successful SIAM Conference on Computational Science and Engineering, held in Boston last February. At this conference several minisymposia focused on... Continue »

Barba awarded computer center grant

The Massachusetts Green High Performance Computing Center (MGHPCC) is a ground-breaking collaboration of five of the state’s most research-intensive universities, state government and private industry—the most significant such collaboration in the history of the Commonwealth.

Prof. Barba led a multi-institution bid to success in the call for research seed grants of the Massachusetts Green High-Performance Computing Center, MGHPCC. The award amounts to $130,487 for the duration of one year. This project is a collaboration among investigators in BU, Harvard and UMass aimed at creating an open and high-performance software infrastructure... Continue »

New paper published: CiSE journal

The Roofline model relates floating-point performance, operational performance and memory performance. This model distinctly quantifies the high operational intensity of fast N-body algorithms and reveals their advantage on many-core architectures.

A new paper authored by Rio Yokota and Lorena Barba has appeared (published online 3 Jan. 2012) in Computing in Science and Engineering, the joint publication of the IEEE Computer Society and he American Institute of Physics. This paper presents a new hybrid treecode/FMM formulation that has been implemented in the ExaFMM code, recently released... Continue »