Parallel-Programming

The first excercise of the course of Parallel Programming. The objective is to create a knnsearch function in c and optimize it using parallel programming.

Components

The project is divided in 4 directories:

1. cilk_pthreads This directory contains the implementation of both simple and parallel knnsearch using the OpenCilk and Pthreads libraries. Both a source file (.c) and a header file (.h) are included and needed. The source file contains the main function while the header one has all the other functions. To compile it manually (not recommended, a Makefile is provided, see details below) in a terminal, the following command should be used, with the necessary adjustments for your system: /yourPath/clang -I/usr/include/openblas -L/usr/lib/openblas -o cilk_pthreads cilk_pthreads.c -lopenblas -lm -fopencilk -O3
2. omp_pthreads Similarly, this directory contains the implementation of both simple and parallel knnsearch using the OpenMP and Pthreads libraries. To compile it manually in a terminal, the following command should be used, with the necessary adjustments for your system: gcc -fopenmp -I/usr/include/openblas -L/usr/lib/openblas -o omp_pthreads omp_pthreads.c -lopenblas -lm
3. test_files This directory was created mostly for testing and having a natural flaw when coding. It reveals the whole process of the project
4. txtDataset This directory contains all the .txt files used, although for your convenience they have already been placed inside the directories.

How to compile the project

This project contains a Makefile to compile two different versions of a program: one using OpenMP and the other using OpenCilk. The Makefile automates the process of compiling both versions, each with the appropriate flags and compilers.

Prerequisites

Before running the Makefile, make sure you have the following installed:

• GCC: For the OpenMP version.

• OpenCilk: For the OpenCilk version of the program.

  • You must have OpenCilk installed on your system, and you should be using the OpenCilk-enabled clang compiler (not the system default clang).

  To install OpenCilk, follow the instructions on the OpenCilk GitHub page.

  If you already have OpenCilk installed, make sure the clang compiler provided by OpenCilk is available in your system PATH or directly referenced in the Makefile.

• OpenBLAS: Make sure OpenBLAS is installed and the required headers and libraries are available:

  • Headers: /usr/include/openblas
  • Libraries: /usr/lib/openblas

Quick Start

1. Clone the repository:

   git clone https://github.com/NicKylis/Parallel-Programming.git
   cd Parallel-Programming

2. Check that your clang points to the OpenCilk version:

   Run the following to ensure the OpenCilk clang is being used:

   which clang
   clang --version

   If the output doesn't mention "OpenCilk," make sure to update your PATH to include the OpenCilk clang:

   export PATH=~/opencilk/opencilk-YOUR_VERSION_HERE/bin:$PATH

   To make this change permanent, add the export PATH=... line to your ~/.bashrc (or ~/.zshrc if you're using Zsh) and run:

   source ~/.bashrc

3. Run the Makefile to compile both versions of the program:

   make

   This will compile two executables:

   • omp_program in the omp_pthreads directory (compiled with OpenMP).
   • cilk_program in the cilk_pthreads directory (compiled with OpenCilk).

4. Clean up compiled files:

   To remove the compiled executables, run:

   make clean

   This will delete both omp_program and cilk_program.

Troubleshooting

• If you encounter an error like unknown argument: '-fopencilk', it means you're using the wrong cilk compiler.

AUTHOR

Nikolaos Kylintireas