GPU Computing and Deep Learning Homework and Project
SJTU CS473 course, 2021 Spring
❗ ❗ ❗ Any forms of copying code is strictly prohibited.
Homework
HW1
The homework is about implementing RBFKernel in CUDA, and try to optimize it.
Source code and report can be found here.
HW2
The homework is about to draw a three-color-triangle.
Source code of CUDA and C++, OpenGL
HW3
The homework is about to draw the Voronoi diagram and Delauny triangulation.
Source code of CUDA: naive and jump flood algorithm; draw Delauny triangles.
Source code of OpenGL: naive algorithm
Project
The first part is to change the mesh segmentation network structure to get higher performance. The final version of the network can be found here.
The second part is to do subdivision of the mesh. I reference ozkanyumsak's in my work.
Report can be found here (Only Chinese version).