Scientific Computing Masterclass: Parallel and Distributed
What you'll learn
- Learn about Supercomputing
- HPC system's basic components
- HPC software stack
- HPC job schedulers and batch systems (Slurm and PBS Pro)
- Introduction to parallel programming concepts: Open MP and MPI
- GPU programming: NVIDIA CUDA and AMD HIP ROCm
- AWS HPC Deployment and Run Codes
Requirements
- Linux/ Unix command line
- Computer programming skills in any language
Description
Welcome to the First-ever High Performance Computing (HPC) Systems course on the Udemy platform. The goal main of this course is to introduce you with the HPC systems and its software stack. This course has been specially designed to enable you to utilize parallel & distributed programming and computing resources to accelerate the solution of a complex problem with the help of HPC systems and Supercomputers. You can then use your knowledge in Machine learning, Deep learning, Data Sciences, Big data and so on.
HPC clusters typically have a large number of computers (often called ‘nodes’) and, in general, most of these nodes would be configured identically. Though from the out side the cluster may look like a single system, the internal workings to make this happen can be quite complex. This idea should not be confused with a more general client-server model of computing as the idea behind clusters is quite unique. Cluster computing utilize multiple machines to provide a more powerful computing environment perhaps through a single operating system.
WHAT DO YOU LEARN?
A Little bit of Supercomputing history, Supercomputing examples, Supercomputers vs. HPC clusters, HPC clusters computers, Benefits of using cluster computing.
Components of a High Performance Systems (HPC) cluster, Properties of Login node(s), Compute node(s), Master node(s), Storage node(s), HPC networks and so on.
Introduction to PBS, PBS basic commands, PBS `qsub`, PBS `qstat`, PBS `qdel` command, PBS `qalter`, PBS job states, PBS variables, PBS interactive jobs, PBS arrays, PBS MATLAB example
Introduction to Slurm, Slurm commands, A simple Slurm job, Slurm distrbuted MPI and GPU jobs, Slurm multi-threaded OpenMP jobs, Slurm interactive jobs, Slurm array jobs, Slurm job dependencies
OpenMP basics, Open MP - clauses, worksharing constructs, OpenMP- Hello world!, reduction and parallel `for-loop`, section parallelization, vector addition,
MPI - hello world! send/ receive and `ping-pong`
Parallel programming - GPU and CUDA: Finally, it gives you a concise beginner friendly guide to the GPUs - graphics processing units, GPU Programming - CUDA, CUDA - hello world and so on! We understand that CUDA is a difficult API, particularly the memory models. We have added some easy to understand CUDA lessons with examples to make your life easy and comfortable to grasp the basics fast!
Parallel programming - AMD GPU and HIP (New! Aug 2023): Learn parallel programming on AMD GPU's with ROCm and HIP from basic concepts to advance implementations. We will start our discussion by looking at basic concepts including AMD GPU programming, execution model, and memory model. Then we will show you how to implement algorithms using ROCm and HIP.
AWS HPC: With the recent advantage of the faster Cloud technologies, AWS provides the most elastic and scalable cloud infrastructure to run your HPC applications. With virtually unlimited capacity, engineers, researchers, and HPC system owners can innovate beyond the limitations of on-premises HPC infrastructure. We have added lectures to show and tell you on how to build a AWS HPC cluster and how to run codes -easily!
Based on your earlier feedback, we are introducing a Zoom live class lecture series on this course through which we will explain different aspects of the Parallel and distributed computing and the High Performance Computing (HPC) systems software stack: Slurm, PBS Pro, OpenMP, MPI and CUDA! Live classes will be delivered through the Scientific Programming School, which is an interactive and advanced e-learning platform for learning scientific coding. Students purchasing this course will receive free access to the interactive version (with Scientific code playgrounds) of this course from the Scientific Programming School (SCIENTIFIC PROGRAMMING IO). Instructions to join are given in the additional contents section.
DISCLAIMER
We created here a total of one university semester worth of knowledge (valued USD $2500-6000) into one single video course, and hence, it's a high-level overview. Don't forget to join our Q&A live community where you can get free help anytime from other students and the instructor. This awesome course is a component of the Learn Scientific Computing master course.
Who this course is for:
- Students, researchers and programmers from any discipline
- Software developers and Big data analysts
Instructors
The Scientific Programming Instructor Team helps you to learn the use of scientific programming languages, such as CUDA, Julia, OpenMP, MPI, C++, Matlab, Octave, Bash, Python Sed and AWK including RegEx in processing scientific and real-world data. The teamed is formed by PhD educated instructors in the areas of Computational Sciences.
Scientific programming is a rapidly growing multidisciplinary field that uses advanced computing capabilities to understand and solve complex problems.
The Scientific Programming School, with 60,000+ students is an awesome e-education start-up initiative to provide professional training and practice courses for Scientific Coding, Linux, and Big Data. It is also an interactive and advanced e-learning platform that gives you the opportunity to run scientific codes/ OS commands as you learn with playgrounds and Interactive shells inside your browser. Scientific Programming Instructors specialize on Linux, Devops, HPC and Data Science coding with scientific programming. Currently we support three OS (Ubuntu, RHEL and SuSE) and 50+ programming languages including the commercial ones like Matlab. At the Scientific Programming School you start learning immediately instead of fiddling with OS, VMs, SDKs and/ IDEs setups. It‘s all setup with Docker on the cloud.