Login to EPCC SAFE SAFE 

ARCHER2 Weekly Newsletter

By ARCHER2 Service on October 28, 2026

Tags: newsletters 

ARCHER2 Enabled…

As the ARCHER2 service draws to a close, we would like to capture all the wonderful things that ARCHER2 enabled our community to achieve.

We invite you to share with us one (or more!) things that having access to ARCHER2 enabled in your work.

Complete the survey

Extending containers to reference or include additional software or data

Webinar, Wednesday 4th November 2026 15:00 - 16:00

This webinar will explain how an existing singularity container can be extended to access or incorporate additional data or software. It will cover accessing external data directly or encapsulated in a filesystem image and how data/applications can be incorporated into a new container.

The steps outlined can be accomplished by a non-privileged user and are applicable to apptainer and podman.

Full details and join link

Building Hybrid HPC/AI workflows using RHAPSODY, radex, and DragonHPC

Webinar, Wednesday 11th November 2026 15:00 - 16:00

Many of the underlying techniques that drive today’s AI applications can be repurposed for advancing scientific simulations.

This webinar focuses on two classes of hybrid AI/simulation workflows: AI-in-the-loop where the AI component is coupled tightly within the simulation’s inner loop (e.g. for sub grid-scale modelling) and AI-around-the-loop where AI techniques are used to drive a simulation (e.g. for parameter exploration).

To address the challenges of defining and executing workflows and in-memory data transfer within a worklow, this webinar introduces RHAPSODY and radex, tools co-developed by HPE and the RADICAL lab at Rutgers University, and backed by the DragonHPC distributed runtime.

Full details and join link

Containers for Reproducible Research: Introduction to Podman and Apptainer

Newcastle University, 12 November 2026 09:30 - 16:00

This course aims to introduce the use of containers with the goal of using them to effect reproducible computational environments. Such environments are useful for ensuring reproducible research outputs and for simplifying the setup of complex software dependencies across different systems. The course will introduce the use of Podman and Apptainer containers but the material will be of use for whatever container technology you plan to, or end up, using. On completion of this course attendees should:

  • Have an understanding of what Podman/Apptainer containers are, why they are useful and the common terminology used
  • Have a working Podman installation on your local system to allow you to use containers
  • Understand how to use existing container images for common tasks
  • Be able to build your own Podman/Apptainer container images by understanding both the role of a Contianerfile recipe in building container images, and the syntax used in Contianerfiles
  • Understand how to manage Podman/Apptainer containers on your local system
  • Appreciate decisions that need to be made around containerising research workflows
  • Understand the differences between Podman and Apptainer containers and why Apptainer is often more suitable for multi-user systems (e.g. HPC)
  • Appreciate issues around reproducibility in software, understand how containers can address some of these issues and what the limits to reproducibility using containers are

Full details and registration

Porting HPC applications to GPUs: choosing a programming model

Webinar, Wednesday 18th November 2026 14:00 - 15:00 (Note earlier than usual start time)

This webinar will survey various programming models that can be used to accelerate applications using GPUs. It will cover language parallelism, library usage, directive-based models (OpenMP/OpenACC), grid-based models (CUDA/HIP), and mention high-level frameworks such as Kokkos.

Full details and join link

PowerSched trial

From 27th October (10:00 GMT) to 3rd November (16:00 GMT), CSE will trial a new power-management technology called PowerSched. This is a follow-up to a postponed trial originally scheduled for early September.

In simple terms, PowerSched eliminates the need to control manually CPU frequencies , and has other interesting power-management features we may wish to exploit on future systems.

CSE has been testing PowerSched over recent weeks and, working with the developers at HPE and HLRS (Stuttgart), CSE has identified a configuration that we believe will provide a very similar behaviour to the current manual CPU frequency control.

During the trial, users should see little change in queue times, although people interested in benchmarking software should be aware that jobs may run slightly slower or slightly faster than usual during the trial.

Subject to progress of the trial, on the final day (3rd November), we may adjust the PowerSched configuration aiming to reduce overall system power draw. This will lead to a modest increase (5-10 percent) in job runtimes, but will allow us to gather important information on PowerSched’s ability to reduce system power draw dynamically.

ARCHER2 Calendar 2027

As in previous years, we plan to produce a printed calendar featuring some of the images from the ARCHER2 Image and Video 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.

[image]

ARCHER2 end of service: 17:00 GMT Friday 20th November 2026

To help users, we have compiled a set of documentation specifically covering the end of the ARCHER2 service

The documentation covers:

  • Timeline for the end of the ARCHER2 service
  • Data on ARCHER2: what do I need to do to save data I have stored on ARCHER2 file systems?
  • EPCC SAFE: what happens to my personal data in SAFE?
  • HPC access beyond ARCHER2

Key impacts for ARCHER2 users:

  • Data on home, work, solid state scratch file systems will not be accessible beyond the end of the ARCHER2 service
  • RDFaaS will continue beyond the lifetime of ARCHER2 but users will need to transfer data from /epsrc and /general to a different local mount point on ARCHER2 - more details will be provided as soon as they are available
  • No login access will be available beyond the end of the ARCHER2 service

If you have questions about the end of service that are not answered by the documentation, then please contact the ARCHER2 service desk

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

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.

Twitter

Recordings of past courses

Recordings of past virtual tutorials