MPI Hybrid Distributed Memory Systems NUMA Parallel & Distributed Programming

Parallel Algorithms for Distributed Hybrid Memory systems using MPI.

Implemented Algorithms:

• 2d Heat Diffusion
• Concurrent Wave Equation
• Gaussian Eliminaiton (Normal and Cyclic Stripped)
• Jacobi Iteration
• Pi Calculation
• Prime Number Generation
• Sorting
• Matrix Vector Multiplication (TBD)
• Matrix Matrix Multiplication (TBD)

Benchmarking:

• Assessed Amdahl’s law and Gustafson’s Law using Strong Scaling and Weak Scaling respectively
• Speed Up
• Efficiency
• FLOPS