Brownian tree is a simulation of diffusion-limited aggregation process. It also can be refered as Lichtenberg figure and it's perfect example of stochastic fractal.
Grey - moving particles
White - frozen particles
Blue - moving particles
Other - different time slices of frozen particles
$ pip install torch==1.5.0+cu101 torchvision==0.6.0+cu101 -f https://download.pytorch.org/whl/torch_stable.html
$ pip install -r requirements.txt
$ sudo apt install ffmpeg -y
| Configuration | Grid size | Particles percent | Iterations | Time (sec) |
|---|---|---|---|---|
| cpu+numpy | 301 | 10 | 1500 | 535 |
| cpu+pytorch+oldMoveFunc | 301 | 10 | 1500 | 2544 |
| gpu+pytorch+oldMoveFunc | 301 | 10 | 1500 | 7627 |
| cpu+pytorch | 301 | 10 | 1500 | 37 |
| gpu+pytorch | 301 | 10 | 1500 | 12 |
| gpu+pytorch | 501 | 10 | 1500 | 30 |
| gpu+pytorch | 1001 | 10 | 1500 | 128 |
| gpu+pytorch | 301 | 20 | 1500 | 19 |
| gpu+pytorch | 301 | 30 | 1500 | 19 |
| gpu+pytorch | 301 | 10 | 5000 | 36 |
| gpu+pytorch | 301 | 10 | 10000 | 47 |
- CPU: Intel(R) Core(TM) i5-6600K CPU @ 3.50GHz
- GPU: GeForce GTX 1070
This code is freely available for free non-commercial use, and may be redistributed under these conditions. Third-party librariers are subject to their respective licenses.