CyberTraining: Pilot: Employing Proper Orthogonal Decomposition (POD) and High-Performance Computing (HPC) in Advanced CI

Part of NSF Initiative on Workforce Development for Cyberinfrastructure (CyberTraining)

Contact

mailDaqing Hou
dhou@clarkson.edu

Syllabus

Instructors
See list of instructors here
Teaching Assistants
  • Ahmad D. Suleiman's Picture Ahmad D. Suleiman (2024)

    Ph.D. Student - Electrical and Computer Engineering Department.

  • Anthony Dawling's Picture Anthony Dowling (2022)

    Ph.D. Student - Electrical and Computer Engineering Department.

  • Lin Jiang's Picture Lin Jiang (2022)

    Ph.D. Student - Electrical and Computer Engineering Department.

  • Martin Veresko's Picture Martin Veresko (2022, 2024)

    MS Student - Electrical and Computer Engineering Department.

Course Descriptions

This training covers principles and practices of high-performance computing (HPC) programming, which includes concepts of parallel and distributed computing, multicore CPU architecture, GPGPU (General purpose GPU) architecture, NVIDIA CUDA programming, and MPI (Message Passing Interface) programming. Also, this training provides intensive, integrated instruction on using open-source computing platforms to solve Ordinary/Partial Differential Equations (ODEs/PDEs), develop Proper Orthogonal Decomposition (POD) models using HPC. Team-based interdisciplinary projects offer an effective approach based on the data-driven POD leaning algorithm for computationally intensive multiphysics simulation problems in various science and engineering disciplines.

Course Structure
Module Topic Brief Description Instructor
1 HPC & Programming Basics Linux & cluster environments (ACRES/XSEDE , SLURM); Review of C/C++ (essential programming constructs, compile-build-debug); HPC computer architectures & programming models: MPI and general-purpose GPU Hou, Liu
2 PDEs for Science & Engineering Review of PDEs; FEM & FDM intro (boundary condition, governing equation, basis function, weak form, mesh generation) Yao (2024), Liang (2022), Welland (2022), Prudil (2022)
3 Open-Source Software Programming & POD Open-source toolkit - FEniCS; PETSc, SLEPc and their scalability with HPC; Proper Orthogonal Decomposition (POD) method, procedure to derive POD modes, and applications Hou, Cheng, Liu
4 Projects POD applications in heat transfer, electromagnetic wave propagation and quantum eigenvalue problems; Hand-on projects with the support from PIs and the TA Hou, Cheng, Liu