Project 1: Annie Qiu

[AnnieQiuuu]

Repo Link: https://github.com/AnnieQiuuu/Project1-CUDA-Flocking

Features

• Feature 0: Naive Boids Simulation with searching all the boids as neighbors.
• Feature 1: Scattered Boids Simulation using a scattered uniform grid for more efficient neighbor search.
• Feature 2: Coherent Boids Simulation where the uniform grid is further optimized for memory coherence, improving performance.
• Feature 3: 8 cells and 27 cells search in Scattered Boids Simulation and Coherent Boids Simulation
• Feature 4: Performance analysis comparing Naive, Scattered, and Coherent simulations under different conditions (boid count, block size, neighbor checks). Included graphs for:
  • Framerate change with increasing number of boids.
  • Framerate change with increasing block size.
  • Framerate comparison between 8 and 27 neighbor checks.

Feedback

• This project provides a solid introduction to CUDA programming. However, I found the performance analysis a bit confusing, as I wasn't sure what level of improvement to expect or what the performance rates should ideally look like when I was doing that. It would be helpful to have a recitation or further clarification on the expected performance results and the reasoning behind them.

###Clarification
Sorry for the late Pull request. I do not intend to use the late day for this assignment. I did complete my assignment on time before the due date and committed and pushed my work to Git Hub. However, I hadn’t used Pull Requests before, so I mistakenly thought it was the same as submitting the README. I’m wondering if this misunderstanding would result in the use of a late day or not, as the assignment itself was completed on time. Thank you for your consideration!