alpaka and openPMD Workshop & Hackathon
Dear HPC Enthusiasts,
We are excited to invite you to a dynamic three-day workshop and hackathon focused on two powerful tools for high-performance computing: alpaka and openPMD. This event will bring together developers, researchers, and enthusiasts from diverse backgrounds to collaborate, learn, and innovate.
Event Dates: 23-25 October 2024
UPDATE: The pre workshop session has been moved forward by one hour to 9 AM CEST on 22 October. Sorry for the inconvenience.
We have a pre-workshop session on the 22nd of October from 10 AM to 12 PM, to help you familiarize and set up your development environment on the LUMI supercomputer.
Registration deadline is 11 October. Please note: Registration by this deadline is required to provide access to HPC resources, as account setup takes time. NO late registrations will be allowed as we cannot provide access to compute resource on short notice.
For a brief description of these libraries, see below.
What to Expect:
- Hands-on Tutorials: Participate in guided, hands-on sessions to set up your environment and run your first alpaka and openPMD applications.
- Engagement with Experts: Engage with leading experts of these libraries who will share insights, best practices, tips and their experiences.
What we will cover:
- Introduction to alpaka: Learn about the alpaka library, its architecture, and how it enables portable and performant parallel programming across various hardware backends.
- Optimizing with alpaka: Discover advanced techniques for optimizing kernels and managing memory in alpaka to achieve maximum performance.
- Getting Started with openPMD: Dive into the openPMD standard, exploring how it facilitates scalable I/O for scientific data in simulations and experiments.
- Integrating openPMD with Simulation Codes: Learn best practices for integrating openPMD with your existing simulation frameworks and how to leverage its full potential.
- Showcasing use of alpaka and openPMD with PIConGPU: PIConGPU is a user of both these libraries and leverages them to do exascale simulations. Take a short look inside PIConGPU to see how these libraries are used in a large project and run a test simulation with PIconGPU.
Join us for an engaging and productive event where innovation meets collaboration. Whether you are new to alpaka and openPMD or an experienced user, this workshop and hackathon will provide valuable insights and opportunities to advance your skills and projects and at the end you will be ready to use these powerful libraries in your own projects.
Register now to secure your spot. We look forward to seeing you there!
Hybrid Event Format
Our alpaka and openPMD hackathon will be hosted in a hybrid format. This means you can join us either in-person at HZDR or virtually from the comfort of your own home or office.
We will provide help and support, whether you join us online or in-person. We aim to provide a smooth and seamless experience for all participants.
Getting to HZDR
You can get to HZDR from Dresden with public transport or by car. Detailed instruction can be found on this webpage.
Video Conference Link
The link to the video conference will be shared with the registered participants close to the date of the conference.
Organization
This hackathon is organized by HZDR/CASUS with the Plasma-PEPSC CoE in collaboration with ENCCS and EuroCC@GCS, High-Performance Computing Center Stuttgart (HLRS), Jülich Supercomputing Centre (JSC), Leibniz Supercomputing Centre (LRZ)—National Competence Centres for High-Performance Computing.
alpaka
alpaka (Abstraction Library for Parallel Kernel Acceleration) is a powerful C++ library designed to enable portable and efficient parallel programming across a variety of hardware platforms. It provides a uniform API that abstracts the underlying hardware, allowing developers to write parallel code that can be executed on different backends, such as CPUs, GPUs, and FPGAs. By using alpaka, developers can achieve high performance and maintain portability, making it an ideal choice for high-performance computing (HPC) applications.
Key features of Alpaka include:
- Hardware Abstraction: Write once, run anywhere. alpaka abstracts the specifics of different hardware architectures.
- Performance: Optimize your applications to run efficiently on any supported backend.
- Flexibility: Support for various parallel programming paradigms and patterns.
- Scalability: Suitable for both small-scale and large-scale HPC environments.
GitHub: https://github.com/alpaka-group/alpaka
openPMD
openPMD (Open Standard for Particle-Mesh Data) is a community-driven metadata standard for describing and storing scientific data in a portable and scalable manner. It provides a FAIR data format for particle and mesh data, enabling blazingly fast and easy IO, sharing, analysis, and visualization of large datasets. The openPMD standard ensures consistency and interoperability across different software and hardware platforms, promoting reproducible research and collaboration.
Key features of openPMD include:
- Standardization: Defines a common format for metadata and data, facilitating data exchange and analysis.
- Scalability: Efficient handling of large datasets, suitable for HPC environments.
- Interoperability: Seamless integration with various simulation codes and data analysis tools.
- Extensibility: Flexible and extensible to accommodate specific needs of different scientific domains.
GitHub: https://github.com/openPMD/openPMD-standard
