Describir: Towards Efficient HPC: Exploring Overlap Strategies Using MPI Non-Blocking Communication