Grep parallelization with MPI

This project consists in a simplified implementation of a string matching algorithm, with the introduction of parallel computation through the use of Microsoft MPI APIs.

Code Description

The main files that build up the project are:

• grep.h: class that defines the grep namespaces, which includes data types and functiions that are useful for the implementation of the algorithm.
• grep-main.cpp: the entry point fo the program, which implements the string matching algorithm, with the management of multiple processes.

Language and APIs

The code is entirely written in C++ programming language, with the use of the following libraries and APIs (omitting the standard ones):

• fstream: used for parsing the text file
• cstring: used for copying the content of vectors of strings into an unique array of chars, that is necessary as a consistent input for MPI functions
• mpi.h: which introduces parallel computation in the algorithm, through a message passing mechanism between processes.

Performance analysis

Check the analysis of the performances of the implementation here

How to run the code (Windows)

1. Install MinGw64
2. Install CMake version ≥ 3.23
3. Create folder for building project

  mkdir build
  cd build

4. Generate the makefiles

  cmake -G “MinGW Makefiles” ..

5. Build the project

  cmake --build .

6. Run the program

  mpiexec -np <num_proc> .\parallel_grep.exe <word_to_search> ..\resources\input_file.txt

where <num_proc> needs to be replaced by a positive integer representing the number of processes that will execute the code, and <word_to_search> needs to be replaced by a string representing the word to be found