Distributed and Cluster Computing

Key Points

The demand for computational speed

  • Computational speed is critical in solving humanity’s growing complex problems across all aspects of society.
Introduction to paralel and distributed computing

  • There various parallel computing architecture/model

  • Users need to find the one that give the best speedup/efficiency measures.

  • Speed/efficiency depends on how well programs are parallelized.
Introduction to OpenMP
Introduction to XSEDE
OpenMP: parallel regions and loop parallelism

  • Parallel for allows simplification of code
OpenMP: Work sharing and controlling thread data
Introduction to MPI
MPI: point-to-point, data types, and communicators
MPI: Functional parallelism and collectives
MPI: pleasantly parallel and workload allocation
Introduction to CloudLab
Deploying compute nodes for a supercomputer
Distributed File Systems
Schedulers for cluster of computers
Data intensive computing
High throughput computing
Introduction to IOT