Hpc dtu

Can you list me the High Performance Computing (HPC) Courses & Trainings offered by HPC Clusters – DTU? Also provide me the contact detail of HPC Clusters – DTU?

[Rajkumar Agarwal]

The following High Performance Computing (HPC) Courses & Trainings are offered by HPC Clusters – DTU:

HPC Introduction – tailored for new HPC users
It is dedicated to beginners in High Performance Computing. No previous knowledge of HPC and even UNIX is required.

List of Topics
1. Introduction to High Performance computing
• Where does the need for HPC come from?
o A few examples from DTU
o HPC: specialized hardware or cluster of “commodity hardware”?
2. HPC cluster at DTU: interactive usage
• Overview of the system, hardware and software.
• Multi-user environment:
o What you get, your share of resources
o How to login: Graphical User Interface and/or Command Line
o Interactive usage: running programs, command line basics
• Hands-on
3. HPC cluster at DTU: batch jobs
• Running programs on the HPC: the Resource Manager, the idea behind batch jobs.
• Job script:
o Preparing a job script
o Submitting a job
o Managing jobs
• Hands-on
4. Simple introduction to Parallelization
• Basic overview on parallelism, Parallel programming models:
o Distributed, shared
o General concepts: communication, costs, parallel overhead, balancing
• Running parallel programs on HPC
• Examples


Advanced HPC
It is dedicated to potential power-users of the HPC clusters to make the users aware of the possibility that HPC at DTU gives to developers.
The course gives an overview of the development tools available (compilers, debuggers, performance analyzers, and Integrated Development Environments), and an overview of parallelization and of the most common parallel programming models. A basic knowledge of HPC environment (UNIX commands, basic programming skills, how the batch system works) is required.


Accelerator Courses
The DTU HPC Competence Center periodically offers half- and one-day special courses on developing scientific applications for accelerators.
The courses consider the practical development of algorithms and associated coding in the CUDA programming environment for general purpose computing with GPUs, and/or in OpenMP for the Intel Xeon Phi co-processors.
These courses are primarily targeted for expert programmers that are already developing their own scientific applications in an HPC environment.


Contact Detail:
DTU Compute
Building 324, Room 280
Technical University of Denmark
2800 Kgs Lyngby
Denmark