Related Experiment Videos
On the parallelization of linkmap from the LINKAGE/FASTLINK package.
A Rai1, N Lopez-Benitez, J D Hargis
1Department of Computer Science, Texas Tech University, Lubbock, 79409, USA.
Computers and Biomedical Research, an International Journal
|October 6, 2000
Summary
Parallel computing speeds up genetic linkage calculations. A dynamic task allocation strategy for the Linkmap program significantly improved performance, nearing maximum speedup potential.
Area of Science:
- Computational Biology
- Genetics
- High-Performance Computing
Background:
- Genetic linkage calculations are computationally intensive, especially with large pedigrees.
- Existing software like LINKAGE/FASTLINK faces performance limitations due to exponential time increases with pedigree size.
Purpose of the Study:
- To develop a parallel implementation of the Linkmap program for efficient genetic likelihood calculations.
- To evaluate static and dynamic task allocation strategies in a heterogeneous computing environment.
Main Methods:
- Parallel implementation of the Linkmap program.
- Utilized static and dynamic task allocation strategies for distributing computations across a heterogeneous platform.
Main Results:
- The parallel implementation demonstrated significant performance improvements.
- The dynamic strategy achieved performance gains close to the theoretical maximum speedup.
Conclusions:
- Parallel processing is effective for accelerating complex genetic linkage analyses.
- Dynamic task allocation offers superior performance for Linkmap in heterogeneous environments.