CUDA Projects Repository

This repository contains a collection of CUDA programs exploring various aspects of GPGPU computing, from basic memory management to complex cellular automaton simulations.

Project Structure

The source code is located in the src/ directory:

File	Description
src/00-cudaTest.cu	Standard Vector Addition: Basic implementation using std::vector and CUDA C++ APIs. Demonstrates host-device memory transfer.
src/01-cudaTest.cu	Algorithmic Vector Addition: Generates large datasets on the host using a custom formula before processing on the GPU. Uses C-style arrays.
src/02-cudaTest.cu	Kinematics Simulation: A 3D particle physics simulation. Uses curand to initialize random positions and velocities for multiple bodies and updates their states in parallel.
src/03-trafficSimulation.cu	Traffic Simulation (Nagel-Schreckenberg): A cellular automaton model simulating traffic flow on a single-lane circular road. Features acceleration, collision avoidance, and random braking logic to generate emergent traffic jams.
src/04-cudaTest.cu	Experimental Structures: Definitions for audio/device control structures and enums. Currently serves as a prototype/scaffold.

Getting Started

Prerequisites

• NVIDIA GPU with CUDA Compute Capability.
• CUDA Toolkit (nvcc).
• Linux OS (tested on Linux).

Compilation and Running

You can compile any of the source files using nvcc.

Running the Traffic Simulation

This is the feature project of the repository.

1. Compile:

   nvcc src/03-trafficSimulation.cu -o traffic

2. Run:

   ./traffic

   You will see an ASCII visualization of the traffic flow in your terminal:

   |.....>........>.....>>.>...>..>...>|
   

Running Vector Addition

nvcc src/00-cudaTest.cu -o vector_add
./vector_add