CIS565: Project 3: Simulations ------------------------------------------------------------------------------- Fall 2013 -------------------------------------------------------------------------------
INTRODUCTION: ------------------------------------------------------------------------------- This project is a two part project. The first project a 3D visualization of an N-Body system simulated using CUDA and OpenGL shaders. The second part was a simple flocking simulation based largely on this paper: http://www.red3d.com/cwr/boids/
PART I: N-BODY SIM -------------------------------------------------------------------------------
This part of the project is rather straightforward. Here are a few screen shots.
 
PART II: FLOCKING SIM -------------------------------------------------------------------------------
For this part of the project, I implemented a simple flocking simulation. I added in alignment, separation, cohesion, and a spherical boundary avoidance force that kept the particles within the frame of the window. Here are a few screenshots:
 
Here's a video of my flocking simulation in action. I apologize for the poor quality: http://www.youtube.com/watch?v=Ypkwl0yjSD8
PERFORMANCE REPORT: -------------------------------------------------------------------------------
Here's a table with some performance analysis that I conducted on my code. I recorded how many seconds for one update/frame of the simulation to occur when I used naive force accumulation over all the planets in the system compared to using shared memory.
| Number of Planets | Without Shared Memory | With Shared Memory |
|---|---|---|
| 1000 | 0.001 sec | 0.001 sec |
| 2000 | 0.003 sec | 0.002 sec |
| 3000 | 0.005 sec | 0.004 sec |
| 4000 | 0.008 sec | 0.008 sec |
| 5000 | 0.012 sec | 0.010 sec |
| 6000 | 0.015 sec | 0.013 sec |
| 7000 | 0.019 sec | 0.018 sec |
| 8000 | 0.025 sec | 0.022 sec |
| 9000 | 0.042 sec | 0.038 sec |
As you can see, shared memory is faster but only slightly.