Edmond Chow

TL;DR: This work presents several new algorithms and implementation techniques that significantly accelerate parallel MD simulations compared with current state-of-the-art codes, including a novel parallel decomposition method and message-passing techniques that reduce communication requirements, as well as novel communication primitives that further reduce communication time.

TL;DR: Anton's performance when executing actual MD simulations whose accuracy has been validated against both existing MD software and experimental observations is reported, allowing the observation of aspects of protein dynamics that were previously inaccessible to both computational and experimental study.

TL;DR: Newton, "global," and column-oriented algorithms, and options for initial guesses, self-preconditioning, and dropping strategies are discussed, and some limited theoretical results on the properties and convergence of approximate inverses are derived.

TL;DR: This paper presents a distributed breadth- first search (BFS) scheme that scales for random graphs with up to three billion vertices and 30 billion edges, and develops efficient collective communication functions for the 3D torus architecture of BlueGene/L that take advantage of the structure in the problem.

TL;DR: A better practical understanding is gained of ILU preconditioners and how these problems can sometimes be circumvented through pivoting, reordering, scaling, perturbing diagonal elements, and preserving symmetric structure.