Source: Own illustration
From March 13 – March 17, 2023, we will continue the tradition and open our doors again for the Parallel Programming in Computational Engineering and Science seminar, PPCES for short, as well as the Introduction to High-Performance Computing 2023.
Until 2019, the week-long seminar was held at the IT Center every spring for 19 years. Unfortunately, the Corona pandemic interrupted that tradition for a few years.
However, the patience of waiting is over!
This year, the PPCES and also the “Introduction to High-Performance Computing 2023” will again take place as usual as face-to-face events at the IT Center.
Since there is a constant need for training in parallel programming on high-performance computers, the PPCES will focus on the teaching of content from the field and on the use of the RWTH high-performance computer.
The event is aimed at users who have little experience in parallel computing as well as at already experienced HPC users. Registration for the event is also open to internal and external interested parties.
But before the PPCES 2023 starts, the one-day “Introduction to High-Performance Computing 2023” will take place on March 7, 2023.
HPC Introduction 2023
This introduction provides basic information on High Performance Computing (HPC) at the IT Center for different target groups and needs. It lays the groundwork for the follow-up event, PPCES 2023. The HPC Introduction focuses on CLAIX-2018, the cluster installed at the IT Center in late 2018 and expanded in 2019.
The first part of the event, in the morning, provides basic information about HPC resources and is aimed not only at HPC users but also at decision makers*. Participants will receive information that will enable them to make strategic decisions about the use of HPC
resources and the submission of proposals for computing projects that consume an appropriate amount of resources.
No prior experience is required to attend.
The second part of the event, in the afternoon, is aimed at those who (want to) use HPC resources and those who (want to) program HPC resources.
In particular, participants will learn the most important details of how to use CLAIX-2018. Those who run their compute applications on local computers will learn how to run these applications on the RWTH Compute Cluster, provided the necessary software licenses are available. In general, this event is also intended to provide the HPC basics necessary to participate in our more advanced PPCES event, and to help you decide which part of PPCES might be of interest to you.
Basic knowledge of Linux and experience with its applications on other computers is assumed.
Overview of the topics
Morning:
- Fundamentals of high performance computing architecture
- CLAIX 2018
- Memory for HPC
- Application software availability
- Access, project categories and submission
Afternoon:
- Brief introduction to parallel programming
- Overview of OpenMP, MPI and GPU programming
- Performance metrics and measurements
- Compiler optimizations and switches
- Batch job submission
- Process and thread allocation for batch jobs
- Outlook for follow-up sessions: PPCES and aiXcelerate
PPCES 2023
PPCES 2023 will cover the basics of parallel programming with OpenMP and MPI in Fortran and C/C++ and a first step towards performance tuning, as well as current topics in AI/machine learning.
Practical exercises are provided for each topic.
The course content is general, but will be specific to CLAIX, the compute cluster currently installed at the IT Center.
It may be helpful to review the information given at the “Introduction to High-Performance Computing 2023” on March 07.
This is especially true if CLAIX is to be actively used after PPCES 2023.
OpenMP is a widely used approach to programming shared memory architectures that is supported by most compilers today. The basics of the programming paradigm will be covered as well as some advanced topics such as programming NUMA machines. The nodes of the RWTH Compute Cluster contain an increasing number of cores, so shared memory programming is considered an important alternative for applications that cannot be easily parallelized with MPI. In addition, a growing number of application codes are expected to combine MPI and OpenMP for clusters of nodes with a growing number of cores.
The Message Passing Interface (MPI) is the de facto standard for programming large HPC systems. During PPCES 2023, the basic concepts will be introduced and an overview of some advanced features will be given. It will also cover hybrid parallelization, i.e., the combination of MPI and shared memory programming, which is becoming increasingly popular as the number of cores per cluster node grows.
Have we piqued your interest?
Are you particularly interested in one of the above topics, or have you always wanted to dive into the HPC world, get some initial information and didn’t know how to take the first step?
Maybe you are already using the cluster and want to expand your knowledge?
In any case, you’ve come to the right place!
IT Center RWTH Aachen University
Kopernikusstraße 6
52074 Aachen
Registration for the events in advance is mandatory.
On the IT Center events pages you will find the link to register for the “Introduction to High-Performance Computing 2023” and PPCES 2023.
Both events will be held in person only and the event language is English.
Responsible for the content of this article is Dunja Gath.
Source: PPCES
