This is the final blog post in the UNIVERSE-HPC Running a Training Pilot series. These posts contain excerpts from the Pilot Operations Guide published on Zenodo and developed by UNIVERSE-HPC.

In our previous posts we've looked at key areas to include in planning your pilot, the options for how it will be delivered, and obtaining feedback from both attendees and the training team. In this final post we'll focus on the practical aspects of promoting and delivering the training pilot, and provide tips for effectively marketing the pilot to the right audience to set clear expectations, and to help smooth delivery during the event itself. We'll also include advice on setting the tone and engaging learners from the start, and adapting to challenges during the pilot to ensure a positive learning experience.

Promoting a Pilot

Promoting your pilot workshop effectively is crucial to attracting the right participants and ensuring a successful event.

To maximise your chances of reaching your target audience, consider using a variety of advertising channels. You could reach out to groups who have previously attended your training sessions and send targeted emails to mailing lists that are relevant to the workshop’s subject matter. You could also post adverts in relevant Slack channels used by your target audience and include information about your workshop in newsletters that reach your potential participants.

Start promoting the course well in advance to ensure there is enough time for participants to learn about the workshop and sign up. Begin advertising the workshop as soon as the details are confirmed and send out promotional information multiple times to increase visibility. Something to remember is that promotion isn't a one-shot activity, so as the date approaches, consider sending reminder emails to potential participants.

Your promotional materials should be clear, concise, and informative. It’s useful to include the following key points:

• Workshop topic: provide a brief introduction to the subject matter, especially if the topic is not widely known.
• Course level and prerequisites: specify the level of the course (novice, intermediate, advanced) and any prerequisite knowledge. Consider including a short form or quiz to assess participants’ current knowledge.
• Location: indicate where the course will be held. If online, mention the platform (e.g., Zoom, Microsoft Teams). For in-person pilots, indicate whether the venue can accommodate certain accessibility requirements. If the venue has not yet been determined or could be changed, ask participants about their accessibility requirements when they sign up. 
• Dates and times: clearly state the workshop dates and times.
• Eligibility: specify who the course is open to (e.g., students from a particular university).
• Instructors and delivery method: introduce the instructors and outline the method of delivery.
• Expectations: set clear expectations regarding what will be taught, how it will be delivered, and that learner feedback is expected.
• Registration process: explain what will happen if too many people sign up. Indicate whether it’s first come, first served or if there will be a selection process.

When creating your advert, aim for brevity while ensuring it contains all necessary information. Keep the advert short and to the point, ensuring key details are prominent – for example, make critical details like the date, time, and topic bold and large. Alternatively, keep the advert short but provide a link to a more detailed explanation of the workshop - perhaps on an event website - for those who need more information.

Strategically promoting your workshop through various channels, starting early, and providing clear, concise information all help to attract the right participants.

Delivering a Pilot

How to Staff a Pilot?

From the teaching side, in addition to instructors who teach the course, you may want to consider having helpers who, whilst not involved in class teaching, assist learners with any issues or questions. Instructors often adopt the role of helper when not instructing. It's generally recommended to have a ratio of learners to helpers of between 1:5-1:10, depending on the course and how it's delivered. Helpers work particularly well with in-person courses, or those held online with the cohort separated across different virtual rooms (e.g. 5 learners per room with 1 helper in each); in both cases helpers are able to address issues with a minimum of crosstalk. They are less effective with online instructor-led courses however, since issues can't be vocally discussed without interrupting the instructor. Complex issues may be resolved in separate virtual rooms, although it becomes much easier for the learner to fall behind. Smaller issues can be resolved over text chat.

Helpers do not necessarily need to come from within the internal training team. For courses delivered to learners within a singular discipline, recruiting helpers from the same discipline who have applied the taught techniques is very effective. Groups that request training for their less experienced members may well have experienced staff happy to act as helpers throughout the course.

Counterintuitively, instructors who are not experts in the subject matter can be an excellent choice for novice courses, especially if they learnt the topic or skill recently. This is because their obstacles to learning the topic, and how they overcame them, are still fresh in their mind and can be applied to their teaching. Conversely, there's a tendency for experts to suffer from an "expert awareness gap", where their own struggles in learning the fundamentals are long forgotten and so are harder to appreciate, especially if they haven't taught the topic before. For less experienced instructors, short informal practice teaching sessions with co-instructors boosts confidence and helps identify areas for improvement.

How to Schedule a Pilot?

Scheduling a pilot workshop requires careful planning and flexibility to accommodate potential issues and gather valuable feedback.

Before the workshop begins, consider organising a pre-pilot software installation surgery, particularly if installation is complex. This session allows learners to ensure all necessary software is installed correctly on their devices and to familiarise themselves with the new software and tools that will be used during the workshop. It also provides an opportunity for learners to speak directly with instructors and resolve any technical issues. This preemptive step can prevent delays and disruptions on the first day of the workshop.

It can be helpful to allow larger time margins during a pilot workshop to handle unexpected delays without rushing through the content. This is particularly important on the first day where unforeseen technical challenges are most likely to occur. For example, you could start early on the first day to allow for extra time setting up and addressing any last-minute technical issues and have extra helpers available at this time to help with troubleshooting.

Consider scheduling half-day sessions rather than full-days for a pilot workshop. This approach gives you the chance to tweak materials between sessions if necessary and to make course corrections based on feedback and observations from the previous session. Half-day sessions also reduce the intensity of the workshop, making it easier for both learners and instructors to stay focused and engaged. Additionally, learners are likely to retain new information more effectively if the content is spread over multiple days.

Pilots can be more taxing than regular workshops for both learners and instructors. To ensure everyone remains attentive, schedule regular breaks throughout the sessions to prevent fatigue and maintain concentration. The instructors could also encourage participants to stretch, hydrate, and step away from their screens during breaks.

How to Run a Pilot?

Delivering a pilot workshop effectively requires careful planning, clear communication, and adaptability.

First, let’s consider the introduction to the workshop. You can enhance your workshop introduction with a few key elements to set yourself up for success in the rest of the workshop. Start by having the instructors and helpers introduce themselves. This initial interaction increases the likelihood that learners will engage throughout the workshop. From the start, establish an upbeat and exciting tone to create a positive and energetic learning environment. If the workshop will be long or cover difficult content, prepare learners for the challenge and remind them of the benefits of mastering the subject matter. Clearly indicate which aspects of the training are being piloted and which are well-established and emphasise the importance of receiving feedback from the learners on the piloted aspects of the workshop. In the introduction, it’s also useful to inform learners about the break schedule for the workshop so they can plan their time accordingly.

Throughout the workshop, regularly check in with the group to gauge the pace and give learners the opportunity to ask questions, and adjust the speed of the course as needed to ensure everyone is keeping up and note any questions or problems that arise. This feedback provides valuable insights for improving the materials and delivery in future sessions. If the workshop is instructor-led, consider sharing the delivery between multiple instructors. This approach helps keep the energy high and gives each instructor a rest. Make sure you adhere to the break schedule announced at the beginning, since this helps learners manage their time and maintain focus.

As this is a pilot workshop, it’s likely that not everything will go to plan. When something goes wrong, it's helpful to remember that this is the purpose of pilots. If appropriate and feasible, you could work with the learners to solve the problem in real time, or make notes of the issue so that you can fix it before the next workshop. It can be particularly helpful to identify and connect with learners who tend to encounter issues first. These "advance party" learners can help identify problems before the larger group is affected, and provide an opportunity to resolve or mitigate issues before they become larger problems experienced by the wider group. Encourage them to report issues promptly and consider offering incentives for their feedback.

Overall, when delivering a pilot workshop:

• Encourage an interactive atmosphere to foster engagement and create a dynamic learning experience.
• Provide clear instructions to minimise confusion.
• Be prepared to adapt your plan based on real-time feedback and observations.
• Continuously collect feedback throughout the workshop to refine and improve the content and delivery for future sessions, and ensure that the participants understand the importance and value of filling in the post-workshop survey.

In Conclusion

This post concludes the UNIVERSE-HPC Running a Training Pilot series, which we hope you've found useful. You can find the full training guide, on which these posts are based, on Zenodo.

This blog post is the third in the UNIVERSE-HPC Running a Training Pilot series. These posts contain excerpts from the Pilot Operations Guide published on Zenodo and developed by UNIVERSE-HPC.

In the previous two posts, we looked at key areas to include in planning your pilot, the development of training materials, and covered the various options for delivery. In this post, we'll look at feedback, an integral tool to ensure the success and continuous improvement of your training workshop. We'll take a brief look at helpful approaches and the types of feedback you can obtain to help you improve your course in the future, from both the learners and the training team.

Gathering Feedback

It's a good idea to align feedback with the goals of your pilot, so it’s important to clearly define what you want the pilot to achieve beforehand so that you can tailor your feedback collection to measure these specific goals. In particular, ensure you collect feedback on any aspects of your workshop that have never been trialled before.

As well as collecting feedback on the learning experience and training materials, consider collecting feedback on other aspects of the workshop which affect the instructors’ and learners’ experiences including the method of delivery and workshop organisation. It can also be helpful to collect feedback on how participants heard about the workshop and what attracted them to it, as this can help refine your marketing strategy for future events.

It's easy to focus on learner experiences and neglect the perspective of the instructors or helpers, but observations and suggestions from the teaching team may also provide valuable insights from the pilot workshop. One method of collecting feedback from instructors is to provide in-situ instructor notes, via a platform such as GoogleDocs, for instructors to capture experiences and suggestions. For a more structured method of collecting instructor feedback, you could use a specialised training platform such as Oxford's Gutenberg infrastructure, which allows both instructors and learners to add annotations/comments directly to learning material to highlight obstacles to learning or errors.

Surveys are a popular method of feedback collection. Feedback surveys can be used before the workshop to gauge participants' expectations and prior knowledge, at the end of each day to monitor ongoing satisfaction and address issues promptly, and after the workshop to assess overall effectiveness and areas for future improvement. Ensure these surveys are concise to avoid survey fatigue and encourage honest and thorough responses.

A particularly useful tool to determine the learning effectiveness of the workshop is to establish prior knowledge and post-workshop knowledge, by asking the same technical ability questions (based on the expected learning outcomes) in the pre-workshop and post-workshop surveys. For example, for a session teaching the Bash shell, if a learning outcome is to be able to write a Bash script to execute an operation over a number of files, an example question might be "Could you write a script to loop over all the files in a directory and output the first five lines of each file?", and adding this to both surveys. Developing similar key questions for each session or topic presents a number of benefits:

• The pre-workshop survey establishes a prior knowledge baseline across the cohort, allowing instructors to tailor the pilot according to cohort ability.
• Comparing these questions' results from pre- and post-workshop surveys indicates the extent to which learning took place across the pilot, and helps to highlight training material areas that need improvement.

The idea is to ensure that the questions are sufficiently representative of the learning outcome but to not have too many of them.

Additionally, at the end of each day or section it can be helpful to hold wrap-up sessions to facilitate immediate and open feedback. These sessions allow participants and instructors to voice their thoughts while they are still fresh.

By implementing a comprehensive and organised feedback collection process, you can ensure you gather as much useful information as possible from your pilot workshop.

Things to Consider when your Pilot has Finished

It's worth remembering that, particularly for a pilot, the workshop isn't complete as soon as it's been delivered. In order to get the most out of a pilot workshop, aim to collect and action as much useful feedback as possible. This can be done in a number of ways.

Firstly, in terms of the surveys:

• Don't neglect to ask any attendees who dropped out early to fill in the survey, since otherwise, your survey analysis is subject to survivorship bias, and it's important to understand why the training didn't connect with anyone who did not see it through to the end.
• Analyse results from post-workshop survey ability questions against the same results in the pre-workshop survey, which will help to understand the extent to which learning took place across the pilot, and helps to highlight training material areas that need improvement. For example, if learning effectiveness for a particular training episode was low, this may highlight issues with the content, its delivery, or its portrayal in the material. This analysis also provides evidence on the impact of these workshops which can be used to support funding of these activities in the future.
• Collate and analyse survey text responses for areas of success and improvement. This helps direct future improvement work and any positive comments provide qualitative support for the activities. Consider requesting longer testimonials from attendees who particularly liked the training, to publish in blog posts or future promotional material, for example.
• Consider a follow-up with attendees, perhaps between 3-6 months after the pilot, to ask how useful they have found the training in their work and any areas that were particularly valuable.

Having feedback from attendees is only one side of the feedback coin, so it's also important to obtain feedback from the training team. One way to do this is to hold a debrief (also sometimes referred to as a "retrospective" or "post-mortem"), where aspects of organisation, delivery and materials can be discussed to identify successes and areas of improvement from the perspective of the trainers. It also provides an opportunity to quickly prioritise issues and potentially assign them to individuals and schedule the remedial work.

What's Next?

The final post in this series will look at ways to promote your event and how to actually deliver the pilot, in terms of staffing, scheduling and running it.

If you would prefer to have all the guidance at once, take a look at the full guide on Zenodo.

This blog post is the second in the UNIVERSE-HPC Running a Training Pilot series. These posts contain excerpts from the Pilot Operations Guide published on Zenodo and developed by UNIVERSE-HPC.

In the first post of the series, we introduced key areas to consider when planning and preparing for a training pilot and also covered some tips for how to develop training materials. In this post, we will cover two decisions in more detail that greatly affect the shape of a pilot: how it will be delivered and whether it will be held online or in-person. The options for these decisions vary significantly and will greatly affect the event, so the advantages and disadvantages of each should be carefully considered.

Method of Delivery

We'll consider two key delivery models which are both organised around a scheduled event:

• Instructor-led (pedagogical), where learners follow the lead of an instructor in a shared learning environment, using techniques such as presentations or live coding.
• Supported self-learning (andragogical), where the responsibility and pace of learning are passed to the learner, although within an environment shared with other learners.

In both cases, helpers (or other instructors) are on hand to answer questions and help solve problems. Let's take a look at the pros and cons of each of these approaches.

For established training courses, switching to the other mode of delivery may still work well, although success depends largely on the nature of the material, an understanding of the typical challenges encountered with the material (and whether they are suited to a particular delivery style), and the technical level of the learners. Of course, for pilots, there are many more unknowns, particularly with what learners (and instructors) will find challenging, and to what extent there are undiscovered errors in the pilot materials. How problems are handled and communicated is therefore even more important.

Issues tend to surface much more quickly in instructor-led delivery, which are then resolved (with learners able to use solutions to common problems immediately). Capable instructors are also able to adapt the learning narrative in the event of unsolvable issues, using analogous examples to illustrate the same concepts. However with self-learning, as noted in the drawbacks, the self-pacing and distance learning aspects may mean that problems aren't always so readily identified and may lead to learners struggling with the material if they are unsure they are at fault. Plus, any solutions to problems must be communicated to other learners somehow, which is more challenging with a self-learning approach.

One way to make self-learning more engaging is to include pre-recorded videos of instructors in the material. This could be for course or section introductions, or for explaining fundamental or difficult concepts, although note that developing, recording, and editing course videos often takes a fair amount of effort, as does re-recording/re-editing them if the underlying material needs to change. However, if a self-learning course is planned to be developed from an existing in-person one, recording instructors' presentations for the self-learning version can mitigate this considerably.

There is also a third, hybrid model of both, where the training is initially instructor-led (for example, using live coding) then switches to supported self-learning, which has the benefit of high levels of engagement and feedback with introductory topics to give the training momentum, motivate learners, and bring them up to a baseline skill level, whilst then allowing them individual space to self-learn remaining topics at their own pace.

In-person vs Online

The rise of teleconference platforms, together with the pandemic, has seen online training events becoming increasingly popular. Let's look at the benefits of online compared to in-person training.

There is a common theme shared with supported self-learning in terms of their disadvantages: particularly with pilots, the identification and resolution of issues is often more difficult and time-consuming, and it's harder to gauge overall progress. Online delivery together with supported self-learning may therefore prove especially challenging for both instructor and learner, so for initial pilots and inexperienced training teams an in-person instructor-led event will likely make the most sense.

However, there are ways to mitigate issues with running online + self-learning training. Organisers often take advantage of "physical" virtualisation, with events splitting the cohort into manageable groups across multiple "breakout" rooms each with an instructor (or helper) to assist learners. This helps to reduce the communication "bandwidth" and cross-talk issues with using a single virtual room, although at the expense of a greater need for coordination between instructors and helpers in separate rooms. Shared documents and backchannel communication also greatly help to record and share issues and solutions found in separate virtual rooms.

What's Next?

The two remaining posts in this series will look into the following:

• How to collect and use feedback to improve your training pilot, from the perspectives of both the learner and training team
• Ways to promote your event and how to actually deliver the pilot, in terms of staffing, scheduling and running it

Alternatively, if you would prefer to have all the guidance at once, take a look at the full guide on Zenodo.

This blog post is the first in the UNIVERSE-HPC Running a Training Pilot series. These posts contain excerpts from the Pilot Operations Guide published on Zenodo and developed by UNIVERSE-HPC.

In his treatise The Art of War, Sun Tzu said that "every battle is won before it's even fought", and so it is with training pilots - early preparation is key. That's certainly not to say that a pilot is a battle against learners, or if it is a battle, it's a battle with learners against barriers to knowledge, both from the perspective of the learners wanting to improve their skills, and the training team wanting to understand how the course needs to improve.

But perhaps a better way to see a training pilot is as a collaboration between the instructors and the learners, where decisions related to planning aim to maximise the exchange of knowledge in both directions: from the training materials and training team to the learners (i.e. what will be taught), and from the learners in terms of feedback to the training team (and by extension, the training materials).

Based on this aim, the goals of a training pilot are typically twofold:

• For the learner: deliver a positive learning experience for attendees that is measurable against the learning objectives for the training.
• For the training team: gather actionable feedback to improve the organisation, delivery, and training materials for future events (and sometimes, the same event whilst it's running).

This post will introduce the key areas in which decisions need to be made during the early planning stages. During planning it's a good idea to capture decisions within a pilot brief, which is a short, living document that evolves during the pre-workshop stages and forms a single agreed point of reference for those involved. Future posts in this series will delve deeper into these areas, as well as provide guidance on the delivery of pilots and handling post-training feedback. Another important aspect to consider early is the preparation of training materials themselves, so we'll also cover the development of training materials in this post.

What will be Piloted?

When pilots are being planned there is often an understandable focus on the learning content and training materials that are being piloted, but it's a good idea to ensure that how it will be organised, delivered, and assessed receives equal attention.

For a group developing an entirely new course for the first time, this would obviously involve piloting many of these aspects for the first time. However for more established groups, or projects that aim to deliver multiple training events, some aspects may be reused alongside established ones, so it's a good idea to discuss and explicitly record in the pilot brief how each aspect will be handled. Examples may include some or all of the following:

• For content, it could be a whole course of new modules, new versions of existing modules, or perhaps piloting a selection of new modules that will form part of a larger course in the future. If it's the latter, it makes sense to pilot these in logical groups and orderings as much as possible, so there is a consistent learning narrative throughout the pilot.
• You might be trying out new ways of delivering an existing course, perhaps piloting an online version of one that was previously conducted in person.
• Perhaps you're aiming to increase course attendance, so you're piloting new ways of promoting events to your target audience.

Alongside any new aspect to be piloted, ensure there are ways to assess to what extent it has been successful and to capture how to improve it for next time. We'll cover some approaches to this throughout the series.

Method and Medium of Delivery

There are two delivery models, both organised around a scheduled event, that we'll consider in this guide series:

• Instructor-led (pedagogical), where learners follow the lead of an instructor in a shared learning environment, using techniques such as presentations or live coding.
• Supported self-learning (andragogical), where the responsibility and pace of learning are passed to the learner, although within an environment shared with other learners.

In both cases, helpers (or other instructors) are on hand to answer questions and help solve problems.

At the far end of the self-learning spectrum, there is also asynchronous self-learning, where learning is conducted purely in the learners' own environments and typically without direct support if problems are encountered.

The rise of teleconference platforms, together with the pandemic, has also seen online training events becoming increasingly popular. How does this compare with in-person training, for both instructor-led and supported self-learning?

As we'll see in the next post, there are many benefits and drawbacks to these approaches.

How to Develop Training Materials?

The development of training materials - and the pedagogy behind it - is a volume of guides in itself, but there are key aspects to note that help with running pilots. An established approach to lesson development and one recommended by the Carpentries, is backwards design (also known as reverse instructional design). Using this approach, instead of starting with the topics to teach and writing content, the process starts by considering the intended learning outcomes as a set of goals (and in the case of pilots, these goals should stem from the overall goals of the pilot). The three fundamental steps to backwards design are:

• Identify what learners should be able to do by the end of the pilot as a set of intended learning outcomes - what should the learners be able to do at the end of the pilot?
• Develop assessments to determine whether learning has taken place, derived directly from the intended learning outcomes, instead of solely as part of the organisation of a pilot which is fairly typical.
• Plan lesson content that prepares learners to complete the assessment successfully, which keeps lesson content relevant, since it's directly derived from what will be assessed and by extension from the learning outcomes.

Another advantage of this approach is that the learning outcomes are usable directly to set clear course expectations with learners, in terms of what will be taught, when advertising the workshop and introducing the course on the day.

Another key aspect is the types of learning material to use, with many workshops making use of multiple types:

Once the materials are in a functional state, and ideally long before their pilot, invite colleagues external to the training materials development team to review them. A fresh perspective on the material will help spot potentially fundamental errors and obstacles to learning that are missed during development due to expert awareness gaps or being "too close" to the material. Where possible, invite reviews from two different types of reviewer:

• Those with similar prerequisite skills to the target learners, i.e. what are the learning obstacles, what needs clarification, what works well and what needs improvement from a learner perspective
• Those with knowledge and experience of the training topics, i.e. suggest ways to simplify the taught techniques, identify missed and valuable learning opportunities that should be included, and prioritise parts to be removed based on their practical value (which is useful if the material is known to be too long)

What's Next?

Over the next few months, the three remaining posts in this series will look into the following:

• Building on this post, the decisions around where and how the pilot will be delivered
• How to collect and use feedback to improve your training pilot, from the perspectives of both the learner and training team
• Ways to promote your event and how to actually deliver the pilot, in terms of staffing, scheduling and running it

Alternatively, if you would prefer to have all the guidance at once, take a look at the full guide on Zenodo.

The UNIVERSE-HPC project is excited to announce the publication of a new guide aimed at helping educators and trainers deliver successful pilot workshops. This valuable resource provides insights into what must be done to deliver a successful pilot, and what differs from running an established workshop.

The guide will help you navigate the many decisions related to planning, organising and delivering a pilot workshop, so that both the learners and the training team get the most out of it. Whilst this guide is aimed at those involved in training pilots, many aspects also apply to training in general, so may also be helpful for those running established training events.

The UNIVERSE-HPC project is running a free online pilot training course on OpenMP and the Message Passing Interface (MPI) which is open to all researchers across the UK. This course will run in the afternoons from Monday 12 - Friday 16 August, and will be held online via Microsoft Teams. You can elect to attend the sessions for either OpenMP, MPI, or both.

The course will cover conceptual and hands-on introductions to writing C programs that use these parallelisation technologies. This event will be delivered using a mix of presentations, instructor-led live coding (where learners code along with the instructor), and supported self-learning, with training materials including exercises for you to progress through at your own pace, but with instructors on hand to assist with any questions or challenges you encounter.

Sessions will run 12:30-16:00 BST (GMT+1) on each day, with OpenMP sessions on 12-13 August, and MPI sessions on 14-16 August.

You'll need some experience in the following to attend this course:

• The Bash shell command line, e.g. navigating a file system, using Bash commands
• The C programming language, e.g. writing and compiling basic C programs (including variables, loops, functions and C pointers)

This training course is limited to 20 participants on a first come, first served basis. As a pilot training course, you will be asked for feedback on the materials, the format, and other aspects of the course, and thus this is also an opportunity to contribute to the future development of OpenMP and MPI training courses.

The UNIVERSE-HPC or ‘Understanding and Nurturing an Integrated Vision for Education in RSE and HPC’ project will define a training curriculum framework – spanning from undergraduate to continuing professional development level - for Research Software Engineers (RSEs) specialising in high performance computing (HPC).

Universe-HPC aims to define a training curriculum framework – spanning from undergraduate to continuing professional development level – for Research Software Engineers (RSEs) specialising in high performance computing (HPC). As part of this project, we explore ways to make training more accessible, for audiences at various skill levels. Together with HPC resources kindly provided by ARCHER2, in March we piloted an online foundational course for researchers in Southampton who were interested in learning fundamental high-performance computing (HPC) skills.

One of the outputs of UNIVERSE-HPC is to make existing training materials available, particularly in the areas of Research Software Engineering and HPC. These materials are written in an easily reusable Markdown format in a GitHub repository under a Creative Commons Licence, for anyone to incorporate into their own courses or reuse however they wish. So far, the project has made 28 different modules available, from introductory aspects of mathematics and Python to topics such as advanced version control, unit testing, and scientific computing. One of the more recent ones is an introductory HPC course originally developed by EPCC, which is currently being ported into this format and was piloted in March this year, using ARCHER2 as the learning platform. The topics included introductions to Supercomputing, parallel and distributed computation, and computer simulations. These gave the attendees the opportunity to gain experience with compiling example HPC applications written using OpenMP and the Message Passing Interface (MPI). Both of these technologies take different approaches within programs to conduct computational processing in parallel, and hence complete much faster.

Delivering any form of computational training can be a challenge at the best of times, particularly with the breadth of software required on attendees' machines, and navigating the technical complexities and problems typically encountered during a workshop. Providing training for high performance computing presents a new set of challenges, particularly around navigating the new principles, paradigms, technologies and architectures within a learning narrative that doesn't overwhelm the learners.

The Real Thing™ as a Learning Platform

Our foundational course aims to provide a broad introduction to HPC, both conceptually and practically, covering the key topics in introductory detail but also allowing the learners hands-on experience with the technologies typically associated with HPC, such as batch job schedulers like Slurm, and parallelisation implementations like OpenMP and MPI. OpenMP and MPI are very useful if computations would otherwise take a long time. OpenMP allows a program to parallelise computation across multiple threads by specifying parts of the program that can be computed in parallel. On the other side, MPI allows parallel computation at a larger scale, across many copies of the same program running simultaneously, and is able to communicate data via messages.

However, in addition to providing a wealth of training courses themselves, ARCHER2 also allows access to HPC resources for external training providers on request. The huge benefit of hosting training on such a platform is that as well as gaining practical experience on a real HPC infrastructure - the UK's National Supercomputing Service - learners need only follow an ARCHER2 registration procedure and install an SSH client program on their machines. An SSH client allows attendees to connect remotely to other machines and infrastructures like ARCHER2, 

which has these technologies and tools (and many others) already installed for use. This greatly simplifies this aspect for novices and helps avoid dedicating a lot of time at the start of the workshop to resolving installation problems. In the case of Slurm, it wouldn't make a lot of sense, given HPC users do not need to install their own batch schedulers on their machines.

The other huge benefit is that ARCHER2 allowed the learner to continue to access ARCHER for a week after the pilot, which meant they had the option to continue to explore the infrastructure and apply what they had learned.

Supported Self-Learning

Our previous pilot last August used a hybrid format, with introductory lessons delivered using a live coding, instructor-led approach, and later lessons provided in self-learning sessions. From the post-workshop survey, overall, learners found the pilot an effective learning experience (8.2/10, n=9) that compared favourably with previous online training courses learners had attended (7.7/10, n=9).

The March pilot was entirely self-learning, with instructors on hand to help out with any difficulties, and learners able to proceed at entirely their own pace. This approach was well received (7.2/10, n=9), with some lauding it ("Loved that it was self paced, a good range of topics but were all well detailed"), although others would have liked more practical exercises. It was also noted that a hybrid approach would have been of benefit here too.

We also made use of Oxford's Gutenberg training platform again, deployed at Southampton, which hosted the training materials and was very well received (8.7/10, n=9). In particular, we used its more advanced training tools, which were really useful. Firstly, the ability for instructors to monitor progress through the material in real time as learners ticked-off exercises as they finished them. We found this very useful to measure progress, although some learners mentioned this was sometimes they would forget to do, so we're working on incorporating reminders into the system. Secondly, we also made use of its material annotations feature, which allows learners and instructors alike to add comments to passages of the material.  These can be monitored in real time - a very useful way to rapidly record and monitor issues as they arise.

We're currently improving the course based on the feedback we've received, prior to making these modules available as part of the UNIVERSE-HPC course materials.

The UNIVERSE-HPC project is working to identify and promote best practices in the training of Research Software Engineers (RSEs) to work in High Performance Computing, with a particular focus on nurturing diversity. In addition to writing, piloting, and delivering training material (e.g. Byte-sized RSE), UNIVERSE-HPC is running a series of seminars to highlight best practices in training.  

Widening Participation in Education - What Can We Learn Post-Pandemic?

In March we were joined by Mishka Nemes (Skills and Training Manager, The Alan Turing Institute), who delivered the second seminar in our UNIVERSE-HPC seminar series. Mishka discussed how the disruption of traditional learning as a result of the COVID-19 pandemic can widen participation by outlining new schemes and suggested improvements to existing ones. 

An excellent example of this is the Neuromatch Academy. The Neuromatch Academy was designed as an alternative to the conventional model of intensive training through summer schools. These schools are often less inclusive due to their high cost and the barriers they pose for those unable to commit to extensive travel or time requirements. Instead, the Neuromatch Academy delivers a mix of live online courses (across multiple time zones) and self-paced learning.  It also encourages mentorship and peer learning. 

For existing programs, Mishka discussed a few considerations which could widen participation, including:

1. Time: Make sure that the time window for application is long enough that it allows people to balance it with other commitments.
2. Representation: Aim to increase diversity within the program (e.g. of the speakers) to ensure that potential participants feel welcome and represented.
3. Information: Run webinars and FAQ sessions aimed at addressing potential questions and concerns and encouraging participation.  
4. Selection: If there are more applicants than places, it can be tricky to choose fairly between candidates; once basic criteria are met, there is often little to separate qualified candidates. In these cases, it may be more inclusive to allocate places through a lottery.   
5. Funding: Ensure that there is access to funding to help with expenses so that everything is as cost-neutral as possible to remove barriers to participation.

Since the COVID-19 pandemic, there has, of course, been a rise in online and hybrid learning. Hybrid teaching, where in-person and online students attend the same session, can be very tricky to get right, but parallel in-person and online sessions can work really well! An important component of this is that educators need to be proactive about building interaction into their online learning materials (e.g. through polls or discussions). Training can be used to build communities, which helps to nurture peer support, running both pre- and post-engagement activities can help build these communities. In addition, peer learning and peer mentoring can be really valuable for learning. Peer mentoring is particularly useful as it is less intimidating than meeting with a more senior mentor, and can help build relationships.

There’s much more in the seminar which is available in the video below.  If the community can adopt these and other good practices, it could help to build a more inclusive RSE community. 

UNIVERSE-HPC will run their first Hackathon on Friday 26 April from 10:00 to 16:30 BST. The event will take place in person in Oxford and online. This one-day Hackathon aims to enhance the course materials that UNIVERSE-HPC have been developing over the past couple of years. This will be a great opportunity to network with people interested in training for RSEs, contribute to training materials, and learn about Gutenberg, a new training platform! 

There will be three strands that you can get involved in and contribute to: Course content, from beginner to advanced topics; Infrastructure, enhancing/extending/contributing to the Gutenberg training platform; and Pedagogy, creating learning objectives for lessons. You can already have a look at the list of open issues if you’d like, though you’re by no means restricted to this list. 

This will also be a fantastic opportunity to learn about using and deploying Gutenberg, an open-source teaching platform developed at the University of Oxford, that you could use in your own teaching. Gutenberg uses course materials written in Markdown and offers the features of a learning management system, for example, giving access to a specific group of enrolled students and tracking their progress through a course. Have a look at an example deployment from Oxford. 

Don't miss out on this exciting opportunity to enhance your skills, contribute to training materials, and network with people interested in training for RSEs. Register now to secure your spot!