Show The Graduate Center Menu
 
 

Scientific Parallel Computing

Rationale

Computationally complex problems cannot be solved on a single computer.
They need to be run in an environment of 100 to 1000 processors or more.
Designing algorithms to efficiently execute in such a parallel computation
environment requires a different thinking and mindset than designing algo-
rithms for single processor computers. This course is designed to give the
students the parallel computation perspective using the MPI framework.

Description

Computationally complex problems cannot be solved in a single computer
either because they are combinatorially complex (NP-Hard) or because they
are large involving much data such as very large matrices or much compu-
tation. The framework we use to solve these kinds of problems in parallel
is called MPI, short for Message Passing Interface. We examine combina-
torial problems such as Boolean Satis_ablility, Set Partitioning, Traveling
Salesman and large problems such as might be in matrix multiplication or
simulated annealing.

Topic List

  • MPI Tutorial

  • Amdahl's and Gustfason's Laws

  • Matrix Multiplication

  • Boolean Satis_ability

  • Set Partitioning

  • Simulated Annealing

  • Graph Coloring

  • Graph Betweenness

  • Large Optimization Problems

  • Student Presentations of Papers and their Programming Results

Learning Goals

_ Learn how to design algorithms in parallel environments
_ Learn how to use MPI in a parallel program
_ Learn how to use MPI in solving
     -Clustering Problems
     -The Traveling Salesman Problem
     -The Set Partitioning Problem
     -Matrix Multiplication
     -Simulated Annealing
     -Optimization Problems
     -Graph Coloring
     -Graph Betweenness

Assessment
Every student will work on two di_erent MPI programs to solve computa-
tionally complex problems of their own choosing. In the second half of the
course they will report on their algorithms and the results of their programs.
Grades will be based entirely on their programs and presentations.