Project 1 : Ling Xie

[Jack12xl]

Repo Link
In this project, we manage to fulfill all the requirements:

1. Baseline: Naively traverse all the existing boids
2. Uniform Grid: Add spatial uniform grid for searchings.
3. Coherent: Based on uniform grid, align the velocity and position attributes in memory for large memory bandwidth.

And extra points:

1. shared memory: try to speed up the updateVelocity kernel
2. grid loop optimization: use adaptable neighbor searching

[Jack12xl]

Feedback:
A decent proj to get familiar with CUDA