This repository contains assignments for the course CS633: Parallel Computing executed on the PARAM Rudra cluster using MPI.
Assignment1/– Point-to-point MPI communicationAssignment2/– 3D stencil and halo exchange
- MPI (OpenMPI / Intel MPI)
- SLURM scheduler
- PARAM Rudra supercomputer
All experiments were run as per assignment specifications. Results include execution time analysis and scalability observations.