Projects

CUDA Matrix Multiplication

CUDA C++

CUDA matrix-multiply kernels using shared-memory tiling, register blocking, and vectorised loads, compared against cuBLAS.