NCSI: A Study for Implementation of Parallel Algorithms for High Fidelity Mixtures of Biofilm and Solvent

Status	Completed
Mentor Name	Jung-Han Kimn
Mentor's XSEDE Affiliation	XSEDE Investigator / Educator
Mentor Has Been in XSEDE Community	4-5 years
Project Title	A Study for Implementation of Parallel Algorithms for High Fidelity Mixtures of Biofilm and Solvent
Summary	We are developing a massively parallel 3D code to simulate high fidelity mixtures of biofilm and solvent. Biofilm growth has important industrial and medical ramifications, including corrosion in oil and gas pipelines and possible causal links to a host of health issues. Despite this, there have been few high fidelity 3D simulations of biofilms in the research literature. One reason is that a high fidelity simulation of the partial differential equations that govern biofilms requires the use of parallel computing (e.g., supercomputers), which necessitates more sophisticated computational techniques. Such techniques include message passing between processors, parallel linear solvers, and domain decomposition preconditioning methods
Job Description	As a part of this project, an undergraduate will develop base code for the "computational infrastructure," or the base computational tools that we will use to build the full 3D, parallel code. Successful outcomes will include demonstration of parallel communication with the Message Passing Interface (MPI), implementation of parallel linear solvers using the Portable, Extensible Toolkit for Scientific Computation (PETSc, developed at Argonne National Laboratory), and visualization of decomposed, 3D data using Paraview (developed by Sandia and Los Alamos National Laboratories). The undergraduate will demonstrate these capabilities through the numerical solution of a model problem, namely the 3D advection equation
Computational Resources	In the first stage, the undergraduate student will develop basic parallel codes using South Dakota State University cluster machine to debug and validate with small size problem. Once the code will be validated, we will use XSEDE resource to expand the result to more complicate and larger size problems to check scalability, speed, and accuracy.
Contribution to Community
Position Type	Apprentice
Training Plan	Understanding of basic Parallel Computing.Ideas Understadning of basic knowledge of Biofilm and Fluid Dynamics. Understanding of basic numerical methods for Computational Fluid Dynamics Proficiency in PETSc, MPI, KSP, Implementation Basic Understanding of Numerical Methods including KSP iterative methods and PC (preconditioner)
Student Prerequisites/Conditions/Qualifications
Duration	Semester
Start Date	01/01/2019
End Date	05/31/2019