This work was done during the High Performance Computing (HPC) course at ISAE-SUPAERO. The task was to explore the different ways to multiply two matrices in Cuda, and how to do effectively.
It notably involves the use of a shared memory by taking advantage of the Matrix Block Multiplication: