ARCHER2 Weekly Newsletter
By ARCHER2 Service on October 16, 2024
Tags:
- Autumn 2024 UKRI Access to HPC Call open (deadline 28 Nov 2024):
- Exploring new computational frontiers with SmartSim, webinar, Thursday 24th October 2024 15:00 - 16:00
- Single Node Performance Optimisation, Edinburgh, 12 - 13 November 2024 09:30 - 17:30 Day 1, 09:30 - 16.30 Day 2
- Message Passing programming with MPI, 26 - 27 November 2024 09:00 - 17:00, Birmingham
- Intermediate Research Software Development, 26 - 28 November 2024, 09:30 - 17:00, Queen’s University, Belfast
- ARCHER2 Image Competition results - and sign up for your 2025 calendar
- Accessibility of EPSRC Large Scale Facilities - User survey
- Recently added known issues
- Upcoming ARCHER2 training
Autumn 2024 UKRI Access to HPC Call open
Deadline: 28 November 2024 4:00pm UK time
No TA form required at application stage
Webinar for potential applicants
UKRI will hold a webinar on 7 November from 11:00am to 12:00 noon. This will provide more information about the funding opportunity and a chance to ask questions.
Exploring new computational frontiers with SmartSim
Webinar, Thursday 24th October 2024 15:00 - 16:00
Over the last two decades, Machine Learning (ML) and Artificial Intelligence (AI) techniques have been gaining popularity among high performance computing (HPC) scientists. Enhancing state-of-the-art numerical workflows with AI components can lead to two main improvements: the discovery of new results and the optimisation of the computational resources needed to obtain them. The integration of ML workflows and more traditional numerical software can be difficult due to differences in how these computational paradigms are developed and deployed on large systems.
In this webinar, we introduce SmartSim, a portable open-source framework maintained by HPE, which can be used to couple HPC applications and ML toolkits. Using exemplar use cases from Computational Fluid Dynamics and Climate Modelling, we will show how these new AI-enhanced workflows can be used for scientific discovery. First, we will describe how SmartSim has been used on ARCHER2’s small GPU development platform and then present the experience of one research group who are coupling machine learning and Xcompact3D to improve the power output of wind farms by introducing dynamic wake steering.
Single Node Performance Optimisation
12 - 13 November 2024 09:30 - 17:30 Day 1, 09:30 - 16.30 Day 2, Edinburgh
This course covers techniques for improving the performance of parallel applications by optimising of the code that runs within each node.
Modern HPC systems such as ARCHER2 are being constructed using increasingly powerful nodes, with larger and larger numbers of cores and enhanced vector capabilities. To extract maximum performance from applications, it is therefore necessary to understand, and be able to overcome, on-node performance bottlenecks. This course will cover the main features of modern HPC nodes, including multiple cores, vector floating point units, deep cache hierarchies, and NUMA memory systems. We will cover techniques for efficient programming of these features, using batch processing options and compiler options as well as hand tuning of code. The course will also contain an introduction to the use of Cray performance analysis tools. Prerequisites:
Participants must be familiar with software development on ARCHER2, or any other HPC facility, using C, C++ or Fortran.
This course is targeted at users interested in optimising the performance of their own applications, e.g. through compiler options or code changes.
Further information and registration
Message Passing programming with MPI
26 - 27 November 2024 09:00 - 17:00, Birmingham, Tuesday 3rd December 14:00 to 16:30 online
The world’s largest supercomputers are used almost exclusively to run applications which are parallelised using Message Passing. The course covers all the basic knowledge required to write parallel programs using this programming model, and is directly applicable to almost every parallel computer architecture.
Parallel programming by definition involves co-operation between processors to solve a common task. The programmer has to define the tasks that will be executed by the processors, and also how these tasks are to synchronise and exchange data with one another. In the message-passing model the tasks are separate processes that communicate and synchronise by explicitly sending each other messages. All these parallel operations are performed via calls to some message-passing interface that is entirely responsible for interfacing with the physical communication network linking the actual processors together. This course uses the de facto standard for message passing, the Message Passing Interface (MPI). It covers point-to-point communication, non-blocking operations, derived datatypes, virtual topologies, collective communication and general design issues.
Intermediate Research Software Development
26 - 28 November 2024 09:30 - 17:00 Queen’s University, Belfast
This course aims to teach a core set of established, intermediate-level software development skills and best practices for working as part of a team in a research environment using Python as an example programming language (see detailed Learning objectives below). The core set of skills we teach is not a comprehensive set of all-encompassing skills, but a selective set of tried-and-tested collaborative development skills that forms a firm foundation for continuing on your learning journey.
ARCHER2 Image Competition results - and sign up for your 2025 calendar
Thank you to everyone who submitted images and videos to this year’s competition. The outstanding range of entries once again made judging a most enjoyable experience with so many great demonstrations of the important work being done using ARCHER2 .
We are pleased to announce the winners as follows:
- Winning Early Career Researcher entry : Dr Chandan Bose, Aerospace Engineering, School of Metallurgy and Materials, University of Birmingham with their image ‘Separated Vortex Ring above a Three-Dimensional Porous Disc’
- Winning Video: Dr Sébastien Lemaire, EPCC University of Edinburgh with their video ‘Effects of “wake steering” on wind turbine flows’.
- Winning Image and Overall Winner: Eric Breard, School of Geoscience, University of Edinburgh with their image ‘Turbulent Times’
Congratulations!
View the gallery of all the submitted images and videos
ARCHER2 Calendar 2025
As in previous years, we plan to produce a printed calendar featuring some of the images from the competition – you are most warmly invited to sign up to receive a copy, and these will be posted out in time for the new year.
Accessibility of EPSRC Large Scale Facilities - User survey
We have been asked by EPSRC to circulate this survey to all users to collect the experiences of facility users.
With this survey their aim is not to single out any individual facilities but instead to identify and share what is working well. Once they have analysed the responses, the aggregated results will be shared with all the facilities, so we can jointly explore ways in which the range of facilities can become more inclusive, e.g. through good practice sharing.
EPSRC are keen to reach as many users as possible, to help them understand what barriers, if any, exist to providing greater accessibility.
Please do complete the survey.
Recently added known issues
The “Known Issues” page of the ARCHER2 Documentation https://docs.archer2.ac.uk/known-issues/ lists all current open known issues including a description of the issue, its symptoms and any work-arounds.
No recent issues.
Upcoming ARCHER2 Training
- Message-passing Programming with MPI, Online, Always open - self-service
- Shared Memory Programming with OpenMP, Online, Always open - self-service
- QM/MM with GROMACS + CP2K, Online, Always open - self-service
- Hands-on Introduction to HPC, Online, Always open - self-service
- Introduction to LAMMPS, Online, 17 October 2024 10:00 - 16:30
- Data Analysis and Visualisation in Python, Online, 21 - 22 October 2024 09:30 - 16:30
- Exploring new computational frontiers with SmartSim, Online webinar, Thursday 24th October 2024 15:00 - 16:00
- Single Node Performance Optimisation, Edinburgh, 12 - 13 November 2024 09:30 - 17:30 Day 1, 09:30 - 16.30 Day 2
- HPC Carpentry, University of Newcastle, 18 - 19 November 2024 09:30 - 16:00 (Fully booked)
- Intermediate Research Software Development, Queen’s University, Belfast, 26 - 28 November 2024 09:30 - 17:00
- Message Passing programming with MPI, 26 - 27 November 2024 09:00 - 17:00 in Birmingham, Tuesday 3rd December 14:00 to 16:30 online
Further details of upcoming training
We always welcome researchers wishing to present their work in a webinar - please contact the Service Desk if you would be interested in presenting your work.