Hannah Morgan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Hannah Morgan has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
L. Ridgway Scott
3 shared publicationsPatrick Sanan
ETH Zurich, Zurich, Switzerland
3 shared publicationsMatthew G. Knepley
Department of Computational and Applied Mathematics Rice University Houston 77005 TX USA
2 shared publicationsMatthew Knepley
2 shared publications- Alp Dener
Rensselaer Polytechnic Institute
1 shared publication Barry F. Smith
1 shared publication
Latest Publications
- 1 December 2021
- Published by Elsevier in Parallel Computing
- Vol. 108
Toward performance-portable PETSc for GPU-based exascale systems
- 1 December 2021
- Published by Elsevier in Parallel Computing
- Vol. 108
Toward performance-portable PETSc for GPU-based exascale systems
- Open Access
- 20 October 2020
- Published by SAGE Publications in The International Journal of High Performance Computing Applications
- Vol. 35 (1) , 47-59
Understanding performance variability in standard and pipelined parallel Krylov solvers
In this work, we collect data from runs of Krylov subspace methods and pipelined Krylov algorithms in an effort to understand and model the impact of machine noise and other sources of variability on performance. We find large variability of Krylov iterations between compute nodes for standard methods that is reduced in pipelined algorithms, directly supporting conjecture, as well as large variation between statistical distributions of runtimes across iterations. Based on these results, we improve upon a previously introduced nondeterministic performance model by allowing iterations to fluctuate over time. We present our data from runs of various Krylov algorithms across multiple platforms as well as our updated non-stationary model that provides good agreement with observations. We also suggest how it can be used as a predictive tool.
- 20 October 2020
- Published by SAGE Publications in The International Journal of High Performance Computing Applications
- Vol. 35 (1) , 47-59
Open AccessUnderstanding performance variability in standard and pipelined parallel Krylov solvers
In this work, we collect data from runs of Krylov subspace methods and pipelined Krylov algorithms in an effort to understand and model the impact of machine noise and other sources of variability on performance. We find large variability of Krylov iterations between compute nodes for standard methods that is reduced in pipelined algorithms, directly supporting conjecture, as well as large variation between statistical distributions of runtimes across iterations. Based on these results, we improve upon a previously introduced nondeterministic performance model by allowing iterations to fluctuate over time. We present our data from runs of various Krylov algorithms across multiple platforms as well as our updated non-stationary model that provides good agreement with observations. We also suggest how it can be used as a predictive tool.
- 1 January 2018
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Scientific Computing
- Vol. 40 (1) , A130-A141
Towards a Unified Finite Element Method for the Stokes Equations
We present an algorithm for solving the Stokes system based on the Scott--Vogelius technique together with the iterated penalty method. We show that by projecting the Scott--Vogelius pressure onto a continuous function space we can recover an accurate, continuous pressure solution while retaining a divergence-free velocity. This algorithm could be of particular interest when modeling incompressible complex fluids where there is a strong dependence on the pressure. We also show that it is more efficient than the Taylor--Hood approach in some cases. Full details of our implementation in the automated finite element software FEniCS are described, and numerical results are given.
- 1 January 2018
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Scientific Computing
- Vol. 40 (1) , A130-A141
Towards a Unified Finite Element Method for the Stokes Equations
We present an algorithm for solving the Stokes system based on the Scott--Vogelius technique together with the iterated penalty method. We show that by projecting the Scott--Vogelius pressure onto a continuous function space we can recover an accurate, continuous pressure solution while retaining a divergence-free velocity. This algorithm could be of particular interest when modeling incompressible complex fluids where there is a strong dependence on the pressure. We also show that it is more efficient than the Taylor--Hood approach in some cases. Full details of our implementation in the automated finite element software FEniCS are described, and numerical results are given.
- 31 March 2016
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 28 (18) , 4532-4542
A stochastic performance model for pipelined Krylov methods
Summary: Pipelined Krylov methods seek to ameliorate the latency due to inner products necessary for projection by overlapping it with the computation associated with sparse matrix‐vector multiplication. We clarify a folk theorem that this can only result in a speedup of 2× over the naive implementation. Examining many repeated runs, we show that stochastic noise also contributes to the latency, and we model this using an analytical probability distribution. Our analysis shows that speedups greater than 2× are possible with these algorithms. Copyright © 2016 John Wiley & Sons, Ltd.
- 31 March 2016
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 28 (18) , 4532-4542
A stochastic performance model for pipelined Krylov methods
Summary: Pipelined Krylov methods seek to ameliorate the latency due to inner products necessary for projection by overlapping it with the computation associated with sparse matrix‐vector multiplication. We clarify a folk theorem that this can only result in a speedup of 2× over the naive implementation. Examining many repeated runs, we show that stochastic noise also contributes to the latency, and we model this using an analytical probability distribution. Our analysis shows that speedups greater than 2× are possible with these algorithms. Copyright © 2016 John Wiley & Sons, Ltd.
- Open Access
A Stochastic Performance Model for Pipelined Krylov Methods
Pipelined Krylov methods seek to ameliorate the latency due to inner products necessary for projection by overlapping it with the computation associated with sparse matrix-vector multiplication. We clarify a folk theorem that this can only result in a speedup of $2\times$ over the naive implementation. Examining many repeated runs, we show that stochastic noise also contributes to the latency, and we model this using an analytical probability distribution. Our analysis shows that speedups greater than $2\times$ are possible with these algorithms.Open AccessA Stochastic Performance Model for Pipelined Krylov Methods
Pipelined Krylov methods seek to ameliorate the latency due to inner products necessary for projection by overlapping it with the computation associated with sparse matrix-vector multiplication. We clarify a folk theorem that this can only result in a speedup of $2\times$ over the naive implementation. Examining many repeated runs, we show that stochastic noise also contributes to the latency, and we model this using an analytical probability distribution. Our analysis shows that speedups greater than $2\times$ are possible with these algorithms.