This project is a Mandelbrot fractal implementation in C++ on GPU using Nvidia CUDA.
We wanted to benchmark our results, so we implemented three versions :
- CPU single-threaded version
- CPU multithreaded version
- GPU version
Two binaries are produced by our CMake.
This one simply runs the algorithm and output a PNG file with the Mandelbrot fractal.
You can specify the version to run with the flag -m
and one of the following: {CPU
, MT
, GPU
}.
This one benchmark the three versions using Google Benchmark library.
Here is an example of our Mandelbrot fractal:
We ran our benchmark on multiple image size and multiple fractal iteration number to get more detailed results: