A MATLAB implementation of the smoothed particle hydrodynamics based on the Philip Mocz's paper, in both CPU and GPU.
By using a GPU, we can speed up the computation to about 10 times. Here, I am using the NVidia Tesla K20m. It is probably the cheapest Tesla card in second-hand markets. ;-) A normal GPU also works.
Please read Philip Mocz's paper here:
https://pmocz.github.io/manuscripts/pmocz_sph.pdf
Although the paper stated MATLAB implementation, I can only find the Python implementation in his Github repository:
https://github.com/pmocz/sph-python
These are some examples in this repository:
Example 1:
Free-falling Universitas Pertamina's logo
Example 2:
Free-falling particles arranged in a grid formation
Example 3:
More free-falling particles also arranged in a grid formation ;-)
Example 4:
Waaayy more free-falling particles also arranged in a grid formation ;-) ;-)
Additional notes:
SPHDemo2D_Ex2_NeighbourSearch_CPU.m uses a very simple neighbour search mechanism to reduce the execution time. We first sort all nodes based on their distances to the origin. Hence, the adjacent nodes are neighbours.
Contact:
manurunga@yadex.com