Project 4: Dominik Kau

[dominikkau]

• Repo Link
• Live Demo
• Features
  • Naive Renderer
  • Forward+ Renderer
  • Clustered-Deferred Renderer
• Extra Credit
  • (+10) Optimized G-Buffer (I decided against a compute shader in the end, after having implemented it. In order to write to the canvas texture from a compute shader, the 'bgra8unorm-storage' feature has to be available, which it wasn't on my GPU and I didn't want to sacrifice performance in all other renderers by using 'rgba8unorm'. Source: https://gist.github.com/greggman/295e38eeedf5957ac50179308666d98b. I hope this doesn't mean a deduction in points, because the compute shader was actually even simpler than the vertex+fragment shader setup.)
• Feedback: I struggled a lot with this one. Without any background in how graphics piplines work, there was a lot to catch up on. That's especially true, because I want to understand why I should be doing the things that I can find in the slides or online. It was however still a very fun project and I'm glad I spent so much time learning about everything involved!