Projects

• Radiance Cascades on WebGPU

  Exploring real-time global illumination using radiance cascades implemented on WebGPU.

• an inquiry into the peformance of GEMM in WebGPU

  Benchmarking and analyzing matrix multiplication kernels to understand WebGPU’s performance limits.

• SPH fluid simulation on the GPU

  Smoothed Particle Hydrodynamics simulation running fully on the GPU for interactive fluid effects.

• Optimizing Gaussian Splatting for the Jetson Nano

  Adapting and optimizing Gaussian Splatting techniques for constrained hardware like the Jetson Nano.

• TinyGrad: an inquiry into performance benefits of fusing kernels

  Investigating how kernel fusion in TinyGrad affects performance, memory bandwidth, and execution time.

• CUDA Ray Tracing

  A basic real-time ray tracer built with CUDA, focusing on efficient memory access and parallelism.