Project maintained by oadrianHosted on GitHub Pages — Theme by mattgraham
Schedule
Week 1 11/04 - 11/11
We researched the topic of SPH more thoroughly exploring different aspects of the algorithm, from how to calculate pressure and densities, to what type of smoothing kernels are typically used.
Installed CUDA on our main work PCs and ensured compatibility with our GPUs
Outlined a brute force O(N^2) algorithm for SPH
Week 2 11/11 - 11/18
Started writing some OpenGL to draw 3D objects on screen.
Designed Particle datastructure
We realized that we were spending too much time writing/learning OpenGL so we decided to build our project on top of one of NVIDIA’s Particle Demo which already had a graphical solution in OpenGL.
Stripped out the Particle System out of the NVIDIA particle demo, removed modules that weren’t useful to us
Week 3 11/18 - 11/25
Wrote brute force SPH solver ontop of the NVIDIA particle demo
Lots of debugging/parameter tuning to get particles to move correctly
More research into how to make our simulation more stable numerically
Added some features to the particle demo such as increasing/decreasing size of simulation box
Week 4 11/25 - 12/2
Thanksgiving Break
More debugging to get particle solver to work
Added Particle Collision
Added Z-indexing sorting
Added OpenMP support for bottleneck for loops
Week 5.0 12/2 - 12/05
Debug particle collisions using OpenMP - Oscar
Write GPU kernels for density and pressure calculations - Logan
Week 5.5 12/05 - 12/09
Write GPU kernels for force calculations - Oscar
Write GPU kernels for velocity and position integration - Logan
Parameter tweaking - Oscar
Week 6.0 12/09 - 12/12
Implement GPU kernel for z-index sorting of particles - Logan