Interactive Mandelbrot Visualization

Mandelbrot was developed during a two semester Project at the CAPS chair of TUM. It shows the possible speedup which can be achieved by parallelization on a cluster of independent compute nodes.

Getting Started

The Application is split into a backend computing the mandelbrot set and a frontend displaying that data and serving as a user interface. A hosted version of this can be found here.

Running the application

1. Start the backend either locally using the provided docker image from from backend/run_docker.sh (requires Docker, python to be installed)

   This starts the backend server with 37 processes at `ws://localhost:9002

   OR

   If you have access to the himmuc cluster from the CAPS chair, you can also use that as the backend:

   backend/run_himmuc.sh -b -p odr [rechnerkennung] [processes] [nodes]

   This will build the backend and start a job on the himmuc with #processes on #nodes. The server will then be port forwarded to ws://localhost:9002

2. Start the frontend either locally with npm (see doc/Ausarbeitung for more details) or use the hosted version here.

   If you have started the backend at another url than 'localhost:9002' you can specify the backend to use with the backend= html get parameter in the mandelbrot URL.

   https://eragp.github.io/mandelbrot?backend=ws://[URL]:[PORT]
   

More Info

For more info see the full documentation of the project (in german) at doc/Ausarbeitung Mandelbrot ERAGP - Finale Ausarbeitung.pdf