Related Experiment Videos
Three-dimensional finite-difference bidomain modeling of homogeneous cardiac tissue on a data-parallel computer
H I Saleheen1, P D Claessen, K T Ng
1Concurrent Technologies Corporation, Johnstown, PA 15904 USA.
IEEE Transactions on Bio-Medical Engineering
|February 1, 1997
Summary
This study demonstrates how data-parallel computing effectively solves large finite-difference bidomain problems. Parallel processing significantly reduces computation time and memory requirements for complex simulations.
Area of Science:
- Computational science
- Applied mathematics
- Computer engineering
Background:
- Solving large-scale finite-difference bidomain problems requires substantial computational resources.
- Traditional computing methods face limitations in memory and processing time for such complex problems.
Purpose of the Study:
- To present a data-parallel computing approach for efficient solution of large finite-difference bidomain problems.
- To detail the algorithm and evaluate its performance on parallel architectures.
Main Methods:
- Utilized a data-parallel computer architecture (Connection Machines CM-200 and CM-5).
- Mapped the finite-difference grid directly to virtual processors (one node per processor).
- Developed and implemented a data-parallel finite-difference algorithm.
Main Results:
- Successfully solved finite-difference bidomain problems with over 2 million nodes.
- Achieved significant reduction in computer time and memory usage.
- Demonstrated effective utilization of parallel processing capabilities.
Conclusions:
- Data-parallel computing is a viable and efficient method for tackling large-scale bidomain problems.
- The presented algorithm offers a scalable solution for complex computational simulations.
- Parallel architectures provide the necessary resources for advancing bidomain problem-solving.