OpenMP Seminar 2025

Hands-on introduction to parallel programming with OpenMP

View Examples · Quick Start

---

About the Project

• This repository contains materials and code examples from a lecture I delivered about OpenMP, as part of a seminar I attended on Parallelization and Program Optimization.
• The lecture covered OpenMP fundamentals step by step:
  • Parallel regions
  • Synchronization (critical, atomic, barrier)
  • Loop scheduling and reductions
  • Worksharing constructs (for, sections, single, master)
  • Tasks and advanced usage
• In addition, the repository includes a MATMUL folder with matrix multiplication implementations (classic and Strassen) parallelized with OpenMP tasks.

---

Examples

Hello World & Pi Approximation

Getting started with OpenMP and exploring scaling issues

Features:

• hello_world.c – simple parallel region with thread IDs
• pi_spmd_simple.c – parallel numerical integration (SPMD style)
• pi_spmd_final.c – fixes false sharing, demonstrates critical & atomic sections
• pi_loop.c – loop-based parallelism with scheduling & reduction

Synchronization & Worksharing

Understanding control of data races and workload distribution

Features:

• Critical vs. atomic vs. barrier
• Worksharing constructs: for, single, sections, master
• Scheduling policies: static, dynamic, guided, runtime

Matrix Multiplication (MATMUL)

Parallel matrix multiplication with tasks

Features:

• matmul.c – classic matrix multiplication with OpenMP
• Strassen’s matrix multiplication with divide-and-conquer
• OpenMP tasks for recursive subproblems
• Time analysis and performance comparison

Explore the MATMUL examples

---

Quick Start

1. Clone the repository:

   git clone https://github.com/your-username/openmp-seminar.git
   cd openmp-seminar

2. Compile examples with GCC and OpenMP flag:

   gcc -fopenmp hello_world.c -o hello_world
   gcc -fopenmp pi_loop.c -o pi_loop
   gcc -fopenmp MATMUL/matmul.c -o matmul

3. Run executables:

   ./hello_world
   ./pi_loop
   ./matmul

4. Control the number of threads:

   export OMP_NUM_THREADS=4

---

References

• OpenMP Official Tutorial
• Parallel programming with OpenMP by Mr Orian Louant
• Seminar slides: Hands-on Introduction to OpenMP