CUDA Programming

This repository is a personal space for practicing CUDA programming and experimenting with GPU development projects.
While it is not primarily designed as a learning resource for others, anyone interested may find the examples useful.

---

Table of Contents

• Overview
• Getting Started
• Folder Structure
• Dependencies
• How to Use
• License

---

Overview

This repository includes a collection of CUDA projects and exercises for:

• Understanding CUDA threads, blocks, and grids.
• Practicing memory management techniques (shared, global, and constant memory).
• Experimenting with performance optimizations (e.g., memory coalescing).
• Building small applications such as matrix multiplication and basic image processing.

---

Getting Started

To get started, clone the repository to your local machine:

git clone https://github.com/mohammed1thabet/CUDA_Programming.git
cd CUDA_Programming

Ensure that you have a compatible NVIDIA GPU and the CUDA toolkit installed on your system.

---

Folder Structure

CUDA_Programming/
├── 001_project_name/    # Folder for the first project, includes source code and VS Studio project.
│   ├── 001_project_name.vcxproj  # Visual Studio project file.
│   ├── 001_project_name.sln  # Visual Studio solution file.
│   ├── kernel.cu        # Main CUDA source file.
│   ├── file1.cpp        # cpp file.
│   ├── file1.hpp        # hpp file.
│   ├── file2.cpp        # cpp file.
│   ├── file2.hpp        # hpp file.
.   .
.   .
.   .
├── 002_project_name/    # Folder for the second project.
│   ├── main.cu
│   ├── helper.cpp
│   ├── project.vcxproj
├── ...                  # Additional numbered project folders.
└── README.md            # Repository documentation.

Each project folder contains the following:

• CUDA source files (.cu).
• may contain C++ source and header files (.cpp, .hpp) for supporting code.
• A Visual Studio project file (.vcxproj) and solution file (.sln) for building the project.

---

Dependencies

• NVIDIA GPU with CUDA support.
• CUDA Toolkit (latest version recommended).
• GCC or Clang compiler.
• Optional: CMake for building more complex projects.

---

How to Use

1. Navigate to a Project: Go to the folder of the project you want to work on:

   cd 001_project_name

2. Build the Project:

   • If you're using Visual Studio, open the .vcxproj file and build the project.
   • If you prefer the command line, compile the .cu file using nvcc:

     nvcc main.cu -o project_name
     ./project_name

3. Experiment: Modify the source code to test different CUDA concepts, configurations, or algorithms. Use each project folder as a playground for trying new ideas.

4. Track Progress: Keep notes or comments within the code about what you learned or discovered while working on each project.

---

License

This repository is licensed under the MIT License. Feel free to use or reference the code, but keep in mind that it was primarily created for personal practice.

---