CPS 5320 (18688), Advanced Scientific Computing

MW 1:30-2:50, QUIN 202

Office hours: MTW 9:00-9:50, BELL 200

Prerequisite: CPS 5310

There are two recommended (but definitely not required) texts:

"Introduction to Computation and Modeling for Differential Equations," Lennart Edsberg, John Wiley & Sons, 2008
"Computational Methods of Linear Algebra, 3rd Edition," Granville Sewell, World Scientific Publishing Company, 2014.

The first 4 problem sets involve solving a range of PDE applications, using PDE2D, which will be made available to you. Book references in these problems are to the Edsberg book.

The last 3 problem sets, involving parallelizing linear algebra Fortran90 codes using MPI, are taken from Chapter 6 of the Sewell book. We will occasionally also have guest lecturers to discuss other topics such as computer visualization. (See Chris Sewell's slides on Visualization and Data Analysis for HPC here, and see Vladik Kreinovich's paper on interval computations here. )

Projects (subject to change during the semester!!)
Topics date due (click on date for problem set)
Beam bending, heat conduction/convection, math finance Sept 14
Elasticity and fluid flow Sept 26
Schrodinger and minimal surface equations Oct 10
Diffusion (3D and axisymmetric) Oct 17
Parallelizing linear algebra codes (Problems 1,2: linear systems) Oct 31
Parallelizing linear algebra codes (Problems 3,6: linear least squares) Nov 14
A parallel Monte-Carlo simulation Nov 30

Projects (subject to change during the semester!!)
Topics	date due (click on date for problem set)
Beam bending, heat conduction/convection, math finance	Sept 14
Elasticity and fluid flow	Sept 26
Schrodinger and minimal surface equations	Oct 10
Diffusion (3D and axisymmetric)	Oct 17
Parallelizing linear algebra codes (Problems 1,2: linear systems)	Oct 31
Parallelizing linear algebra codes (Problems 3,6: linear least squares)	Nov 14
A parallel Monte-Carlo simulation	Nov 30

The Fortran codes required for the MPI problems can be downloaded here

Derivation of equations of

Grades

80% based on projects
For PDE2D and MPI projects, email me (sewell@utep.edu) your Fortran programs, I will re-run them and see the output.
Please work INDEPENDENTLY. If you have questions, ask me, not other students.
20% based on one exam, the last week of classes (Nov 30)

Grades
90-100%	guaranteed A
80-89%	guaranteed B or better
70-79%	guaranteed C or better
below 70%	no guarantees

Note: Drop Deadline is October 28

Useful links

MPI (Fortran) Tutorial

(Very) Detailed MPI Documentation

UTEP Research Cloud

Review of Multivariate Calculus

PDE2D Video

Old Tests

Fall 2016: Practice Exam | Final