Esmond G. Ng
Esmond Ng has more than 35 years of experience in research and development in high-performance numerical algorithms and scientific computing, including 30 years of working with and at the Department of Energy national laboratories.
His research interests are scientific computing, including sparse matrix computation, numerical linear algebra, large-scale computing, high-performance, and parallel computing; mathematical software development and software engineering; and computational complexity and combinatorial scientific computing.
Sparse Matrix Computation
Ng is one of the leading experts in sparse matrix computation and is well known for his work on the theoretical aspect of sparse matrix computation and the development of efficient sparse matrix algorithms. He was a key contributor to the well-known SPARSPAK package, which was one of the first efficient and reliable software packages for solving large sparse systems of linear equations. The package was used widely in industry (such as Boeing and NASA) and universities.
His research includes combinatorial aspect of sparse matrix computation, design and implementation of reliable and efficient sparse matrix algorithms for advanced computer architectures, and complexity analysis. He is heavily involved in computational science and engineering and has been engaging in the development and application of sparse matrix techniques for solving large-scale scientific problems, particularly.
Ng’s research has provided a theoretical understanding of structural properties and issues in sparse matrix factorizations. He was one of the first to study the effect of, and the interplay between, row and column permutations in sparse orthogonal factorization and sparse Gaussian elimination with partial pivoting. His theoretical work has provided a foundation for subsequent sparse matrix research, and the algorithmic development has provided insight that influences subsequent development of sparse matrix factorization algorithms.
The sparse matrix algorithms that Ng and his collaborators have developed are well known and have been used in a large variety of scientific and engineering applications, such as structural analysis, numerical optimization, computational fluid dynamics, and finite element calculations.
Computational Sciences and Engineering
Ng has been actively involved in the application of sparse matrix techniques to scientific and engineering applications for the last 15 years. In particular, he has been a leader in DOE’s Scientific Discovery through Advanced Computing (SciDAC) program since 2001. He has led a team of computational mathematicians that work closely with domain scientists and develop and apply sparse matrix techniques to solve challenging large-scale DOE scientific problems.
His efforts have accelerated scientific discoveries, which include detecting design defects in accelerator cavities through modeling and simulation for the upgrade of the DOE Continuous Electron Beam Accelerator Facility (CEBAF) at Thomas Jefferson National Accelerator Laboratory; predicting the properties of the Fluorine-14 isotope through simulation before it was detected experimentally in 2010 at the Texas A&M University’s Cyclotron Institute; and understanding the reason why the Carbon-14 isotope (which is used for carbon dating) has the long, useful lifetime. All three examples required the solution of large-scale sparse matrix problems (sparse systems of linear equations and sparse eigenvalue calculations) at the heart of the computation.
Ng is the co-author of more than 50 peer-reviewed technical papers, more than 35 conference papers, and written chapters for nine books on computation. He co-authored the book “Parallel Algorithms for Matrix Computations,” which was published by the Society for Industrial and Applied Mathematics (SIAM).
Prior to joining Berkeley Lab in 1999, Ng led the Computational Methods Group at Oak Ridge National Laboratory, where he also held a joint appointment as an adjunct professor at the University of Tennessee in Knoxville. Ng earned his Ph.D., master’s, and bachelor’s degrees in computer science from the University of Waterloo in Ontario, Canada.
Levermann, A., Winkelmann, R., Albrecht, T., Goelzer, H., Golledge, N. R., Greve, R., Huybrechts, P., Jordan, J., Leguy, G., Martin, D., Morlighem, M., Pattyn, F., Pollard, D., Quiquet, A., Rodehacke, C., Seroussi, H., Sutter, J., Zhang, T., Van Breedam, J., Calov, R., DeConto, R., Dumas, C., Garbe, J., Gudmundsson, G. H., Hoffman, M. J., Humbert, A., Kleiner, T., Lipscomb, W. H., Meinshausen, M., Ng, E., Nowicki, S. M. J., Perego, M., Price, S. F., Saito, F., Schlegel, N.-J., Sun, S., van de Wal, R. S. W, "Projecting Antarctica’s contribution to future sea level rise from basal ice shelf melt using linear response functions of 16 ice sheet models (LARMIP-2)", Earth System Dynamics, February 14, 2020, 11:35–76, doi: 10.5194/esd-11-35-2020
R. Van Beeumen, O. Marques, E.G. Ng, C. Yang, Z. Bai, L. Ge, O. Kononenko, Z. Li, C.-K. Ng, L. Xiao, "Computing resonant modes of accelerator cavities by solving nonlinear eigenvalue problems via rational approximation", Journal of Computational Physics, 2018, 374:1031-1043, doi: 10.1016/j.jcp.2018.08.017
Meiyue Shao, Hasan Metin Aktulga, Chao Yang, Esmond G. Ng, Pieter Maris, James P. Vary, "Accelerating nuclear configuration interaction calculations through a preconditioned block iterative eigensolver", Computer Physics Communications, 2018, 222:1--13, doi: 10.1016/j.cpc.2017.09.004
Hasan Metin Aktulga, Md. Afibuzzaman, Samuel Williams, Aydın Buluc, Meiyue Shao, Chao Yang, Esmond G. Ng, Pieter Maris, James P. Vary, "A High Performance Block Eigensolver for Nuclear Configuration Interaction Calculations", IEEE Transactions on Parallel and Distributed Systems (TPDS), November 2016, doi: 10.1109/TPDS.2016.2630699
- Download File: ieeetpds-mfdn-lobpcg-rev.pdf (pdf: 889 KB)
S.L. Cornford, D.F.Martin, V. Lee, A.J. Payne, E.G. Ng, "Adaptive mesh refinement versus subgrid friction interpolation in simulations of Antarctic ice dynamics", Annals of Glaciology, September 2016, 57 (73), doi: 10.1017/aog.2016.13
J. Brabec, C. Yang, E. Epifanovsky, A.I. Krylov, and E. Ng, "Reduced-cost sparsity-exploiting algorithm for solving coupled-cluster equations", Journal of Computational Chemistry, January 24, 2016, 37:1059–1067, doi: 10.1002/jcc.24293
S. L. Cornford, D. F. Martin, A. J. Payne, E. G. Ng, A. M. Le Brocq, R. M. Gladstone, T. L. Edwards, S. R. Shannon, C. Agosta, M. R. van den Broeke, H. H. Hellmer, G. Krinner, S. R. M. Ligtenberg, R. Timmermann, D. G. Vaughan, "Century-scale simulations of the response of the West Antarctic Ice Sheet to a warming climate", The Cryosphere, August 18, 2015, doi: 10.5194/tc-9-1579-2015, 2015
H. M. Aktulga, L. Lin, C. Haine, E. G. Ng, C. Yang, "Parallel Eigenvalue Calculation based on Multiple Shift-invert Lanczos and Contour Integral based Spectral Projection Method", Parallel Computing, December 6, 2013, in press,
H. M. Aktulga, C. Yang, E. G. Ng, P. Maris, J. P. Vary, "Improving the Scalability of a Symmetric Iterative Eigensolver for Multi-core Platforms", Concurrency and Computation: Practice & Experience, September 12, 2013, online, doi: 10.1002/cpe.3129
S.L. Cornford, D.F. Martin, D.T. Graves, D.F. Ranken, A.M. Le Brocq, R.M. Gladstone, A.J. Payne, E.G. Ng, W.H. Lipscomb, "Adaptive mesh, finite volume modeling of marine ice sheets", Journal of Computational Physics, 232(1):529-549, 2013,
- Download File: cornfordmartinJCP2012.pdf (pdf: 1 MB)
Julian Bellavita, Mathias Jacquelin, Esmond G. Ng, Dan Bonachea, Johnny Corbino, Paul H. Hargrove, "symPACK: A GPU-Capable Fan-Out Sparse Cholesky Solver", 2023 IEEE/ACM Parallel Applications Workshop, Alternatives To MPI+X (PAW-ATM'23), ACM, November 13, 2023, doi: 10.1145/3624062.3624600
Sparse symmetric positive definite systems of equations are ubiquitous in scientific workloads and applications. Parallel sparse Cholesky factorization is the method of choice for solving such linear systems. Therefore, the development of parallel sparse Cholesky codes that can efficiently run on today’s large-scale heterogeneous distributed-memory platforms is of vital importance. Modern supercomputers offer nodes that contain a mix of CPUs and GPUs. To fully utilize the computing power of these nodes, scientific codes must be adapted to offload expensive computations to GPUs.
We present symPACK, a GPU-capable parallel sparse Cholesky solver that uses one-sided communication primitives and remote procedure calls provided by the UPC++ library. We also utilize the UPC++ "memory kinds" feature to enable efficient communication of GPU-resident data. We show that on a number of large problems, symPACK outperforms comparable state-of-the-art GPU-capable Cholesky factorization codes by up to 14x on the NERSC Perlmutter supercomputer.
Gianina Alina Negoita, James P. Vary, Glenn R. Luecke, Pieter Maris, Andrey M. Shirokov, Ik Jae Shin, Youngman Kim, Esmond G. Ng, Chao Yang, Matthew Lockner, Gurpur M. Prabhu, "Deep Learning: Extrapolation Tool for Ab Initio Nuclear Theory", CoRR, November 10, 2018,
Mathias Jacquelin, Esmond G Ng, Barry W Peyton, "Fast and effective reordering of columns within supernodes using partition refinement", 2018 Proceedings of the Seventh SIAM Workshop on Combinatorial Scientific Computing, 2018, 76--86,
Ariful Azad, Mathias Jacquelin, Aydin Bulu\cc, Esmond G Ng, "The reverse Cuthill-McKee algorithm in distributed-memory", Parallel and Distributed Processing Symposium (IPDPS), 2017 IEEE International, January 2017, 22--31,
- Download File: RCM-ipdps17.pdf (pdf: 1.1 MB)
Mathias Jacquelin, Yili Zheng, Esmond Ng, Katherine Yelick, "An Asynchronous Task-based Fan-Both Sparse Cholesky Solver", August 23, 2016,
Systems of linear equations arise at the heart of many scientific and engineering applications. Many of these linear systems are sparse; i.e., most of the elements in the coefficient matrix are zero. Direct methods based on matrix factorizations are sometimes needed to ensure accurate solutions. For example, accurate solution of sparse linear systems is needed in shift-invert Lanczos to compute interior eigenvalues. The performance and resource usage of sparse matrix factorizations are critical to time-to-solution and maximum problem size solvable on a given platform. In many applications, the coefficient matrices are symmetric, and exploiting symmetry will reduce both the amount of work and storage cost required for factorization. When the factorization is performed on large-scale distributed memory platforms, communication cost is critical to the performance of the algorithm. At the same time, network topologies have become increasingly complex, so that modern platforms exhibit a high level of performance variability. This makes scheduling of computations an intricate and performance-critical task. In this paper, we investigate the use of an asynchronous task paradigm, one-sided communication and dynamic scheduling in implementing sparse Cholesky factorization (symPACK) on large-scale distributed memory platforms. Our solver symPACK relies on efficient and flexible communication primitives provided by the UPC++ library. Performance evaluation shows good scalability and that symPACK outperforms state-of-the-art parallel distributed memory factorization packages, validating our approach on practical cases.
P. Maris, H. M. Aktulga, S. Binder, A. Calci, U. V. Catalyurek, J. Langhammer, E. G. Ng, E. Saule, R. Roth, J. P. Vary, C. Yang, "No Core CI calculations for light nuclei with chiral 2- and 3-body forces", J. Phys. Conf. Ser., IOP Publishing, August 1, 2013, 454:012063, doi: 10.1088/1742-6596/454/1/012063
P. Maris, H. M. Aktulga, M. A. Caprio, U. V. Catalyurek, E. G. Ng, D. Oryspayev, H. Potter, E.
Saule, M. Sosonkina, J. P. Vary, C. Yang, Z. Zhou,
"Large-scale Ab-initio Configuration Interaction Calculations for Light Nuclei",
J. Phys. Conf. Ser.,
December 18, 2012,
Z. Zhou, E. Saule, H. M. Aktulga, C. Yang, E. G. Ng, P. Maris, J. P. Vary, U. V. Catalyurek, "An Out-of-core Eigensolver on SSD-equipped Clusters", 2012 IEEE International Conference on Cluster Computing (CLUSTER), Beijing, China, September 26, 2012, 248 - 256, doi: 10.1109/CLUSTER.2012.76
Z. Zhou, E. Saule, H. M. Aktulga, C. Yang, E. G. Ng, P. Maris, J. P. Vary, U. V. Catalyurek, "An Out-Of-Core Dataflow Middleware to Reduce the Cost of Large Scale Iterative Solvers", 2012 41st International Conference on Parallel Processing Workshops (ICPPW), Pittsburgh, PA, September 10, 2012, 71 - 80, doi: 10.1109/ICPPW.2012.13
H. M. Aktulga, C. Yang, P. Maris, J. P. Vary, E. G. Ng, "Topology-Aware Mappings for Large-Scale Eigenvalue Problems", Euro-Par 2012 Parallel Processing Conference, Rhode Island, Greece, August 31, 2012, LNCS 748:830-842, doi: 10.1007/978-3-642-32820-6_82
H. M. Aktulga, C. Yang, U. V. Catalyurek, P. Maris, J. P. Vary, E. G. Ng, "On Reducing I/O Overheads in Large-Scale Invariant Subspace Projections", Euro-Par 2011: Parallel Processing Workshops, Bordeaux, France, August 29, 2011, LNCS 715:305-314, doi: 10.1007/978-3-642-29737-3_35
E. G. Ng, J. Sarich, S. M.Wild, T. Munson, H. M. Aktulga, C. Yang, P. Maris, J. P. Vary, N. Schunck, M. G. Bertolli, M. Kortelainen, W. Nazarewicz, T. Papenbrock, M. V. Stoitsov, "Advancing Nuclear Physics Through TOPS Solvers and Tools", SciDAC 2011 Conference, Denver, CO, July 10, 2011, arXiv:1110.1708,
H. M. Aktulga, C. Yang, P. Maris, J. P. Vary, E. G. Ng, "Large-scale Parallel Null Space Calculation for Nuclear Configuration Interaction", 2011 International Conference on High Performance Computing and Simulation (HPCS), Istanbul, Turkey, July 8, 2011, 176 - 185, doi: 10.1109/HPCSim.2011.5999822
Xiaoye S. Li, Meiyue Shao, Ichitaro Yamazaki, Esmond G. Ng, "Factorization-based sparse solvers and preconditioners", (SciDAC 2009) Journal of Physics: Conference Series 180(2009) 012015, 2009, doi: 10.1088/1742-6596/180/1/012015
E. Saule, H. M. Aktulga, C. Yang, E. G. Ng, U. V. Catalyurek, "An Out-of-core Task-based Middleware for Data Intensive Scientific Computing", Handbook on Data Centers, in press, (Springer: February 1, 2014)
Daniel F. Martin, Stephen L. Cornford, Esmond G. Ng, Impact of Improved Bedrock Geometry and Basal Friction Relations on Antarctic Vulnerability to Regional Ice Shelf Collapse, Americal Geophysical Union Fall Meeting, December 15, 2021,
Anne M. Felden, Daniel F. Martin, Esmond G. Ng, SUHMO: An SUbglacial Hydrology MOdel based on the Chombo AMR framework, American Geophysical Union Fall Meeting, December 13, 2021,
Daniel F. Martin, Stephen L. Cornford, Esmond G Ng, Effect of Improved Bedrock Geometry on Antarctic Vulnerability to Regional Ice Shelf Collapse, European Geosciences Union 2020 General Assembly, May 5, 2020,
- Download File: EGU2020-10033-presentation.pdf (pdf: 467 KB)
Dan Martin, Brent Minchew, Stephen Price, Esmond Ng, Modeling Marine Ice Cliff Instability: Higher resolution leads to lower impact, AGU Fall Meeting, December 12, 2018,
- Download File: Martin-AGU-2018-1.pdf (pdf: 2.8 MB)
Daniel Martin, Xylar Asay-Davis, Stephen Cornford, Stephen Price, Esmond Ng, William Collins, A Tale of Two Forcings: Present-Day Coupled Antarctic Ice-sheet/Southern Ocean dynamics using the POPSICLES model., European Geosciences Union General Assembly 2015, April 16, 2015,
- Download File: Martin-EGU-2015.pdf (pdf: 5.3 MB)
Daniel Martin, Peter O. Schwartz, Esmond G. Ng, Improving Grounding Line Discretization using an Embedded-Boundary Approach in BISICLES, 2015 SIAM Conference on Computational Science and Engineering, March 14, 2015,
- Download File: Martin-CSE-3-15.pdf (pdf: 3.5 MB)
Daniel Martin, Xylar Asay-Davis, Stephen Price, Stephen Cornford, Esmond Ng, William Collins, Response of the Antarctic ice sheet to ocean forcing using the POPSICLES coupled ice sheet - ocean model, Twenty-first Annual WAIS Workshop, September 25, 2014,
Esmond Ng, Katherine J. Evans, Peter Caldwell, Forrest M. Hoffman, Charles Jackson, Kerstin Van Dam, Ruby Leung, Daniel F. Martin, George Ostrouchov, Raymond Tuminaro, Paul Ullrich, Stefan Wild, Samuel Williams, "Advances in Cross-Cutting Ideas for Computational Climate Science (AXICCS)", January 2017, doi: 10.2172/1341564
- Download File: AXICCS-Report.pdf (pdf: 4 MB)
W. Kramer, J. Carter, D. Skinner, L. Oliker, P. Husbands, P. Hargrove, J. Shalf, O. Marques, E. Ng, A. Drummond, K. Yelick, "Software Roadmap to Plug and Play Petaflop/s", 2006,
Horst D. Simon, William T.C. Kramer, David H. Bailey, Michael J. Banda, E. Wes Bethel, Jonathan T. Carter, James M. Craw, William J. Fortney, John A. Hules, Nancy L. Meyer, Juan C. Meza, Esmond G. Ng, Lynn E. Rippe, William C. Saphir, Francesca Verdier, Howard A. Walter, Katherine A. Yelick, "Science-Driven Computing: NERSC's Plan for 2006-2010", 2005,
Courtney Shafer, Daniel F Martin and Esmond G Ng, "Comparing the Shallow-Shelf and L1L2 Approximations using BISICLES in the Context of MISMIP+ with Buttressing Effects", AGU Fall Meeting, December 13, 2021,
Hans Johansen, Daniel Martin, Esmond Ng, "High-resolution Treatment of Topography and Grounding Line Dynamics in BISICLES", AGU 2019 Fall Meeting, December 13, 2019,
D.F. Martin, H.S. Johansen, P.O. Schwartz, E.G. Ng, "Improved Discretization of Grounding Lines and Calving Fronts using an Embedded-Boundary Approach in BISICLES", European Geosciences Union General Assembly, April 10, 2019,
- Download File: Martin-EGU2019-final.pdf (pdf: 1.2 MB)
E.G. Ng, D.F. Martin, X. S. Asay-Davis , S.F. Price , W.D. Collins, "High-resolution coupled ice sheet-ocean modeling using the POPSICLES model", American Geophysical Union Fall Meeting, December 17, 2014,
- Download File: Ng-AGU2014.pdf (pdf: 815 KB)
D.F. Martin, X.S.Asay-Davis, S.F. Price, S.L. Cornford, M. Maltrud, E.G. Ng, W.D. Collins, "Response of the Antarctic ice sheet to ocean forcing using the POPSICLES coupled ice sheet-ocean model", AmericanGeophysical Union Fall Meeting, December 17, 2014,
- Download File: Martin-AGU2014.pdf (pdf: 1000 KB)