CS590C: Introduction to Scientific Computing
Fall 1996.
MWF, 10:30 AM, REC 309
Professor:
Ananth Grama
164D, Computer Science Building;
Office Hours: Monday 11:30 AM - 12:30 PM, Wednesday 11:30 AM - 12:30 PM.
(Office hours can also be arranged by appointment).
Teaching Assistant:
David Lutterkort
lutterdc@cs.purdue.edu, 494-4359
Office Hours: Monday, Friday 2:00 PM - 3:00 PM, Tuesday 11:00 AM - 12:00 PM.
Texts:
The course will be drawing on the following sources and will be supplemented
with papers and lecture notes.
-
The CSEP book at Oak Ridge
or another site at
Oak Ridge, or
Vanderbilt, or
Colorado State.
-
Iterative Methods for Sparse Linear Systems, Yousef Saad,
PWS Publishing, 1996.
-
Introduction to Linear Algebra, Gilbert Strang, Wellesley-Cambridge Press,
1993.
-
Scientific computing : an introduction with parallel computing,
Gene Golub, James M. Ortega, Boston : Academic Press, c1993.
-
Other web
resources, class notes from other universities
(PHYS594
at UTK) and class notes.
Course Handouts:
The course handouts will be available electronically at the URL
http://www.cs.purdue.edu/homes/ayg/CS590C/www/cs590c.html.
Students are expected to check this page periodically for supplementary
material, assignments and projects. If you do not have access to the web,
contact your professor.
Course-work:
-
4 Homework sets 20%
(no late homeworks accepted)
-
Group project 20%
(Project spans the length of the semester)
-
1 Midterm 25% (in-class; date to be announced)
-
Final 35%
Class Outline:
-
Introduction
-
Hardware and software evolution.
-
Review of elementary mathematical background
(Basic Linear Algebra, Differential Equations, Mathematical Modeling
of Physical Systems)
-
Using computational tools for problem solving
(Maxima, Matlab, Mathematica)
-
Using libraries (LINPACK, LAPACK, EISPACK, ScaLAPACK, BLAS)
-
Differential equations
(Finite difference / finite element methods); Sparse linear systems
from discretized differential equations.
-
Solving sparse linear systems
-
Iterative methods (ITPACK)
-
Direct methods (Cholesky / LU factorization)
- Accelerating convergence of iterative methods (preconditioning).
-
Integral equations
(Boundary element methods); Dense linear systems from discretized
integral equations.
-
particle dynamics simulations
-
Solution of dense linear systems.
-
Eigenvalue problems.
-
Introduction to structured matrix computations, problem characteristics
leadign to structured matrices (Toeplitz and Block-Toeplitz methods)
-
Computational envorinments: Languages and tools, scientific visualization,
performance analysis.
- High Performance Computing, HPC programming environments, Grand
Challenge Problems.