Skip to content

A fun small project used to practice some CUDA. And to see some sweet fractal ass (don't quote me on that)

License

Notifications You must be signed in to change notification settings

TendTo/MandelbrotSetCUDA

Repository files navigation

Mandelbrot set CUDA

The aim of this project is just for me to practice CUDA programming while trying some fun ideas.

👀 Preview

Commands:

Move Zoom in/out Increase/Decrease max iteration Reset Hide/Show HUD
L-click or WASD or Arrows Mousewheel or +/- E/Q R H

🔧 Requirements

The project written in c++ and uses:

🔟 On windows (See master branch)

Things to download:

The program uses the standard CUDA libraries and some SFML libraries.
All SFML libraries can be both static or dynamic and each has a debug version (just add a -d at the end).
You should use the debug version when compiling in DEBUG.

Before compiling, make sure that the paths to the SFML libraries and headers you downloaded are correct.

For the dynamic linking you will need the following libraries (in VS you need to add .lib):

  • sfml-graphics
  • sfml-system
  • sfml-window

The executable will need the following dll(s) to run:

  • sfml-graphics-2.dll
  • sfml-system-2.dll
  • sfml-window-2.dll

For the static linking you will need the following libraries (in VS you need to add .lib):

  • sfml-graphics-s
  • sfml-system-s
  • sfml-window-s
  • opengl32
  • winmm
  • gdi32
  • freetype

The following flag must be set:

  • SFML_STATIC

Some intersting findings that may act as troubleshooting

  • CUDA on windows needs the MSVC compiler, so you don't have the choice to use mingw
  • All the lib(s) and dll(s) in SFML have a debug version. You just need to add -d before the extension. You should use this version when compiling with the DEBUG flag
  • If you want to use the static linking, make sure that you are compiling in release mode, since otherwise the SFML libraries may conflict with the CUDA ones

🐧 On linux (See linux branch)

Things to download:

The program uses the standard CUDA libraries and some SFML libraries.
Make sure that the paths to the SFML libraries and headers you downloaded is accessible to the nvidia compiler.

On linux, using static libraries is usually discuraged, so SFML provides only the dynamic version. You can, however, compile your own static libraries (guide)

For the dynamic linking you will need the following libraries:

  • sfml-graphics
  • sfml-system
  • sfml-window

The executable will need the following .so to run:

  • sfml-graphics.so.2.5
  • sfml-system.so.2.5
  • sfml-window.so.2.5

Run sh start.sh -h to know what flags you may want to add, then run sh start.sh [flags] with the desired flags.
You can also run the nvcc compiler yourself.