CUDA RayTracing Project

This project implemented ray tracing on GPU (NVIDIA CUDA). It also includes our implementations of various optimization techniques for benchmarking performance and identifying best practices.

Rendering spheres with circulating light source

Real-time interactive with TriangleMesh rendering

Table of Contents

• Prerequisites
• Build Instructions
• Targets
• Usage

Prerequisites

Before building the project, ensure you have the following installed:

• NVIDIA CUDA Toolkit
• GCC (for compiling the CPU code)
• NVCC (NVIDIA CUDA Compiler)

Build Instructions

To build the project, you can use the provided Makefile. The Makefile includes various targets for building different versions of the rendering programs.

To build all targets, simply run:

make all

Targets

The Makefile defines several targets for building different versions of the CUDA rendering and optimization programs:

• realtime: Compiles realtime_render.cu for real-time rendering.
• global: Compiles global_launcher.cu for global memory access.
• array_bvh: Compiles different-versions/array_bvh.cu for BVH with array storage.
• optimized: Compiles optimized.cu for general optimizations.
• optimized_vertices-in-shared: Compiles different-versions/optimized_vertices-in-shared.cu for shared memory optimization.
• optimized_non-coalesced: Compiles different-versions/optimized_non-coalesced.cu for non-coalesced memory access optimization.
• optimized_bvh-tree: Compiles different-versions/optimized_bvh-tree.cu for optimized BVH tree.
• optimized_bvh-texture: Compiles different-versions/optimized_bvh-texture.cu for optimized BVH using textures.
• optimized_recursive: Compiles different-versions/optimized_recursive.cu for optimized recursive BVH.
• cpu: Compiles cpu_launcher.cpp for CPU rendering using OpenMP.

Usage

After building the targets using the make command (e.g., make all or make <target>), you can run the respective executables. Here are examples of how to run some of the programs:

./realtime      # Run the real-time rendering program
./global        # Run the global memory access program
./array_bvh     # Run the BVH with array storage program
./optimized     # Run the general optimizations program

Notes

The code has only been tested on Linux machines (Ubuntu) with CUDA and OpenGL installed.

Authors

Project Members: Cong VU, Minh PHAM, The NGUYEN