Name: Hong Zhang
Pronouns:
Biography:
Hong Zhang received her B.S degree from Beijing Normal University in 1982 and Ph.D degree in computational mathematics from Michigan State University in 1989. She was a math professor at Clemson University of South Carolina and Louisiana State University before she joined the PETSc team at Argonne National Laboratory (ANL) in 1999. She is currently a research professor at the Department of Computer Science of Illinois Institute of Technology, and a consultant in the Mathematics and Computer Science Division of ANL. She conducts scalable (parallel) numerical algorithmic research and develops software for engineering simulations on extreme-scale computers. She is an experienced mentor.
Institution/Lab: Argonne National Laboratory
Website: https://www.mcs.anl.gov/~hzhang/
SRP Collaboration Topic/Title: Scientific Computing using the PETSc/TAO Library on Exascale Machines
Field or research area: HPC
Please select all the topical areas that apply to your project:
Computational Science Applications (i.e., bioscience, cosmology, chemistry, environmental science, nanotechnology, climate, etc.); High-Performance Computing
Brief Abstract:
Robust, efficient, and scalable numerical solvers for simulations based on partial differential equations (PDEs) and networks are at the heart of computational science. PETSc, the Portable, Extensible Toolkit for Scientific Computation (https://petsc.org), is a suite of data structures and routines for the scalable (parallel) solution of scientific applications modeled by PDEs, including related functionality for numerical optimization in TAO (Toolkit for Advanced Optimization). This project will focus on advances in PETSc/TAO composable solvers, which provide the core of scalable multiphysics and multiscale applications, including fusion, geosciences, power grids, nuclear energy, and more. Areas of potential work include developing efficient and scalable algorithms for linear, nonlinear, and timestepping solvers; creating example programs to demonstrate functionality and explore performance on extreme-scale architectures; and advancing new capabilities as motivated by the needs of data-driven computing and machine learning. Students are expected to gain hands-on numerical programming experience on state- of-the-art parallel computers. Students will apply the algorithms and techniques learned to a project in either their own field (particularly encouraged) or suggested by the mentor.
Desired relevant skills, background, or interests:
Mathematical understanding, computer coding skills, motivation and hard working
Other comments:
Do any special requirements apply? International OK
Other, specify:
Keywords:
parallel traffic network; data intensive; machine learning;
Lightning Talk Title: Scientific Computing using the PETSc/TAO Library on Exascale Machines with Application to Traffic Flow