MPI-OpenMP-CUDA Project

This project is a parallel computing implementation using MPI, OpenMP, and CUDA for sequence alignment.

Overview

Sequence alignment is a fundamental problem in bioinformatics, where two sequences of DNA, RNA, or protein are compared to identify regions of similarity. This project implements a parallel solution for sequence alignment using multiple parallel computing paradigms:

• MPI (Message Passing Interface): For parallelizing tasks across multiple processes.
• OpenMP: For parallelizing tasks within each MPI process using shared-memory multiprocessing.
• CUDA (Compute Unified Device Architecture): For offloading compute-intensive tasks to the GPU.

Directory Structure

• src/: Contains the source code for the project.
  • parallel_seq_alignment.c: Main program file implementing the sequence alignment algorithm.
  • cuda_part.cu: CUDA file containing GPU-accelerated functions for sequence alignment.
  • Other source files for MPI and OpenMP implementations.
• inputs/: Contains input data files for testing the program.
• outputs/: Contains output files generated by the program.
• Makefile: Makefile for compiling and running the program.

Dependencies

• MPI implementation (e.g., Open MPI)
• CUDA Toolkit
• Compiler supporting OpenMP directives
• Standard C and C++ libraries

Usage

1. Clone the repository:

git clone https://github.com/your-username/mpi-openmp-cuda-project.git

2. Navigate to the project directory:

cd mpi-openmp-cuda-project

3. Compile the program:

make

4. Run the serial version:

make run_s

5. Run the parallel version:

make run_p

Authors

• Yosi Ben Shushan
• Noam Ben Binyamin