Enhancing Recursive Brute Force Algorithm with Static Memory Allocation: Solving Motif Finding Problem as a Case Study

Abstract

Parallel Recursive Brute Force (PRBF) is one of the algorithms that need high memory. Memory allocation techniques play an important role and have a great effect on the performance of the PRBF. This paper proposes a modified implementation of the PRBF algorithm that uses the static memory allocation technique instead of dynamic memory allocation techniques. This is to avoid the memory management overhead and heap contention problems associated with dynamic memory allocation technique. This paper uses the Motif Finding Problem (MFP), one of the well-known computationally intensive problems in the field of bioinformatics, as a case study. The exponential memory requirements based on the problem size of the MFP make it very challenging to use static memory allocation. Experimental results show that the use of static memory allocation has achieved a significant speedup factor when using 16 MPI rank in comparison with the same implementation using dynamic memory allocation. The proposed R-BF modification scalability is also tested by using different number of MPI nodes and distributing the search space among them, and the results proved a significant reduction in the execution time by increasing the number of working MPI nodes.