Skip to content

zhengliii/Project1-CUDA-Flocking

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

34 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

University of Pennsylvania, CIS 565: GPU Programming and Architecture, Project 1 - Flocking

  • Li Zheng
  • Tested on: Windows CUDA10, i5-3600 @ 3.59GHz 16GB, RTX 2060 6GB (personal computer)

Boids Animation

Boids

Performance Analysis

FPS vs Boids When the number of boids is less than about 10000, the brute force method has a higher frame rate. In case the number of boids is more than about 10000, the scatter and coherent has better performance. The scatter method is generally better than the coherent method. The brute force method calculates velocity change naively based on all the boids. But the other two methods do many additional works. They should calculate the uniform grids and sort them. So, when the number of boids is relatively large, the scatter and coherent method has a better performance.

With visualization disabled, the frame rate is munch better. Obviously, displaying the calculating result on the screen takes time to render.

FPS vs Block Size
The change of block size grid doesn’t obviously affect the performance.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Cuda 41.3%
  • C++ 39.0%
  • CMake 17.7%
  • Other 2.0%