Math 627 - Introduction to Parallel Computing
Spring 2009 - Matthias K. Gobbert
Presentations of the Class Projects
Friday, May 15, 2009, 02:30 p.m., SOND 206
-
02:30-02:45
Long-Time Simulation of Calcium Waves in a Heart Cell to
Model Wave Self-Organization
Zana Coulibaly, Michael Muscedere, Matthias K. Gobbert,
and Bradford E. Peercy, Department of Mathematics and Statistics
A model for the flow of calcium on the scale of one heart cell is
given by a system of time-dependent reaction-diffusion equations coupled by
nonlinear reaction terms. The model has previously been validated to recreate
the conditions of an experiment and can be used to to perform simulations. A
coefficient function of the calcium current through a calcium release unit
determines the calcium distribution level and thus influences wave
self-organization. Previous experiments determined extreme values of
this parameter for which a wave self-organizes or not.
Running the simulation multiple times for multiple values allowed us
to determine a smaller, more precise range of the parameter
for which a wave will self-organize or not.
-
02:50-03:05
Parallel Performance Studies with COMSOL Multiphysics
via Scripting and Batch Processing
Noemi Petra and Matthias K. Gobbert,
Department of Mathematics and Statistics
COMSOL Multiphysics is an extremely powerful and versatile
finite element package for the solution of partial differential equations.
While the graphical user interface (GUI) offers a friendly environment
for solving small problems, for the solution of a more demanding problem
with correspondingly larger memory requirements and longer run times, it
is often desirable to explore the script-based modeling capabilities, as
well as the benefits of running COMSOL in parallel. This work gives
step-by-step instructions on how to use m-files in conjunction with MATLAB
as scripting tool and COMSOL's own binary format for batch processing as
well as guidance on how to run COMSOL in parallel on the cluster hpc in
the UMBC High Performance Computing Facility. We also report on the shared
memory parallel performance of COMSOL using all cores available
on a compute node. The results show that the speedup is not in proportion
to the number of cores used.
-
03:10-03:25
Parallel Performance Studies for a Maximum Likelihood Estimation
Problem with TAO
Andrew Raim, Nagaraj K. Neerchal, Matthias K. Gobbert, and Jorge G. Morel,
Department of Mathematics and Statistics
In this report we present an application of parallel computing to an
estimation procedure in statistics. The method of maximum likelihood
estimation (MLE) is based on the ability to perform maximizations of
functions. The optimizations are often carried with numerical methods
on computer in practice, and may be time consuming for some likelihood
functions. We consider one such likelihood function based on the
Finite Mixture (FM) Multinomial distribution. We conduct estimation
for this problem in parallel on a cluster of computers. Our study
utilizes the High Performance Computing Facility (HPCF) at University
of Maryland, Baltimore County (UMBC), and uses the Toolkit for
Advanced Optimization (TAO) software library. We study how the
resource requirements change as problem sizes vary, and demonstrate
that scaling up the number of processes for larger problems decreases
wall clock time significantly.
-
03:30-03:45
Parallel Numerical Study for a 3-D Poisson Problem
Guan Wang and Matthias K. Gobbert,
Department of Mathematics and Statistics
In this talk, we want to find the numerical solution to the 3-D
Poisson problem using Jacobi method by parallel computing. After
checking the correctness of our computation, we will compare the
speed of computation by blocking communication with
MPI_Ssend/MPI_Recv and by non-blocking communication with
MPI_Isend/MPI_Irecv. Also, we need to compare the speedup and
efficiency of these two different communications and verify
non-blocking case is more efficient.
-
03:50-04:05
Parallel Studies for Chemically Reacting Systems
Yushu Yang and Muruhan Rathinam,
Department of Mathematics and Statistics
Parallel computing code can be applied to solve chemically
reacting systems. In a well stirred chemical system, the number of
molecules for each species can be solved by implicit tau method,
and the histogram of the method can be compared with the exact
simulation called SSA. In this report, the parallel code will be
implemented in both SSA and implicit tau, with the discussion that
how random number generator will be applied in the parallel code
from large number of sample simulations. Moreover, the studies of
analyzing the number of molecules from large samples will be
provided.
Copyright © 2001-2009 by Matthias K. Gobbert. All Rights Reserved.
This page version 1.0, May 2009.