##
Computer Science 290I:

Tools and Methods in Parallel Computational Modeling

**
Mon/Wed 9:00-10:50 **

Phelps 1401

Enrollment code 18226

Course outline

References

Midterm presentations

Final projects

Two trends in computational science and engineering are
the increasing importance of combinatorial methods,
and the rapid spread of parallel cluster computers.

This course will explore the area where these trends
intersect. Our topics will be scientific computations
in which combinatorial or discrete algorithms play an
important role. Our experimental testbed will be a
parallel cluster, on which we will use and extend an
interactive software environment for scientific
computing.

Combinatorial scientific computing includes such topics
as graph models and algorithms for sparse matrix
comptuation, partitioning and scheduling for parallel
irregular computations, geometric algorithms for
generating and manipulating finite element meshes, etc.

In this course we will...

- study methods in combinatorial scientific computing;
- use and further develop Matlab*P, an experimental
software environment for interactive parallel computing;
- study some applications of parallel modeling in
which discrete or combinatorial aspects are important.

Students will do a course project in which they either...

- extend the infrastructure of the Matlab*P system;
- implement a parallel modeling problem of interest
to the student; or
- experiment with parallel algorithms in combinatorial
scientific computing.