Week 2 — Coding Projects

Core

Apply linear transforms to point sets.

• NumPy: Implement matrix-vector and matrix-matrix multiply. Build rotation, scaling, shear, and reflection matrices. Transform a square point cloud.
• Metal: Pass transform matrices into a shader and animate object transforms. · Reading: MBT — uniforms, matrices, coordinate transforms, per-frame updates, shader parameter passing.
• Vulkan: Use uniform buffers or push constants for transforms and animate over frames. · Reading: Vulkan Book — descriptor sets, uniform buffers, transform data flow, frame resources.
• CUDA: Apply a matrix transform to a large point cloud in parallel. · Reading: CUDA Book — device memory layout, grid-stride loops, simple linear algebra kernels.
• Stretch: Separate model, view, and projection matrices cleanly. Add unit tests for composition order.
• Verify: Composition order visibly matters · Identity transform leaves points unchanged · Inverse restores original points when invertible.