|Central European Summer Time (CEST)|
|02:00 PM-2:10 PM CEST||Welcome SIGHPC Education Chapter
Nitin Sukhija, Slippery Rock University of Pennsylvania
|2:10 PM -3:00 PM||Session I (Chair: Susan Mehringer)|
|2:10 PM -2:20 PM|| "What Influences Students Understanding of Scalability Issues in Parallel Computing?". Juan Chen, Brett Becker, Youwen Ouyang and Li Shen
Graduates with high performance computing (HPC) skills are more in demand than ever before, most recently fueled by the rise of artificial intelligence and big data technologies. However, students often find it challenging to grasp key HPC issues such as parallel scalability. The increased demand for processing large-scale scientific computing data makes more essential the importance of mastering parallelism, with scalability often being a crucial factor. This is even more challenging when non-computing majors require HPC skills. This paper presents the design of a parallel computing course offered to atmospheric science majors. It discusses how the design addressed challenges presented by non-computer science majors who lack a background in fundamental computer architecture, systems, and algorithms. The content of the course focuses on the concepts and methods of parallelization, testing, and the analysis of scalability. Considering all students have to confront many (non-HPC) scalability issues in the real world and there may be similarities between real-world scalability and parallel computing scalability, the course design explores this similarity in an effort to improve student's understanding of scalability issues in parallel computing. The authors present a set of assignments and projects that leverage the Tianhe-2A supercomputer, ranked #4 in the TOP500 list of supercomputers for testing. They present preand post-questionnaires to explore the effectiveness of the class design and find an 11.7% improvement in correct answers, and a decrease of 36.8% in obvious but wrong answers. The authors also find that students are in favor of this approach.
|2:20 PM -2:30 PM||"HPC at Curtin University as an Honours course". Chris Harris, Maciej Cytowski, Nigel Marks, Andrew Rohl, Paul Hancock, Paolo Raiteri and Igor Bray
Applications of high-performance computing are now ubiquitous across almost all data-driven academic disciplines. The present-day challenge is to educate the next generation of scientists in the usage of HPC systems. At Curtin University, in partnership with staff and infrastructure of the Pawsey Supercomputer Centre, an Honours level HPC course has been put together by a team of computational physicists and chemists. The topics considered include Serial Optimisation, parallel programming with MPI and OpenMP, programming with GPUs using CUDA and OpenAcc, large-scale I/O, workflows and containers. In its first year the course has been taken by students from Mathematics, Computing and Physics. We expect this to grow to broader areas once more students take the prerequisite courses in computing such as Unix and C programming.
|2:30 PM -2:40 PM||"Partnered Active Learning Increases Engagement and Application for Learner". Ann Backhaus, Marco De La Pierre,Alexis Espinosa Gayosso, Christina Hall, Evan Floden, Paolo Di Tommaso and Jordi Blasco
"..70% of employees say they haven't even mastered the skills they need for their jobs today."
If we haven't mastered the skills for today, how will we prepare ourselves - and our learners - for tomorrow, a highly technological world evolving at accelerated rates?
Active Learning. "[G]reat training programs encourage participants to practice new skills in the context of real-life situations."
How do we encourage active learning in our training? In brief, we look for ways to make the learning directly relevant and meaningful to learners. This requires the coupling of general learning and bespoke application. We're weaving together outside learning (pre-requisites), in-class knowledge building, application and consolidation, and social learning. Pre-COVID, we trialed this approach in a class-plus-hackathon for Nextflow, a tool that aids researchers in the writing of computational workflows. Despite the intensity of the four-day event, attendees were highly engaged. Impressive longitudinal metrics encouraged us to try this approach online, post-COVID. Online, we're experimenting with our Containers in High Performance Computing (HPC) training. We've targeted two distinct scientific domains: Computational Fluid Dynamics and Bioinformatics. The online approach is similar to the classroom approach, consolidating learning with hands-on application. We're trialing different hackathon techniques best fitting each domain.
What makes it possible for us to step out and experiment?
Partnership. We've joined forces. Pawsey specialists are collaborating closely with domain experts to develop and deliver training. The result is training that is highly relevant and applicable - something only true collaboration can make happen.
|2:40 PM -2:50 PM||"Exploring Remote Learning Methods for User Training in Research Computing".Dhruva K. Chakravorty, Lisa Perez, Shaina D. Le, Marinus Pennings, Dylan Rodriguez and Levi Jorda
The COVID-19 national health crisis has forced a sudden and drastic move to online delivery of instruction across the nation. This almost instantaneous transition from a predominantly traditional "in-person" instruction model to a predominantly online model has forced programs to rethink instructional approaches. Prior to the changes brought by COVID 19 related national social distancing, online training in research computing (RC) was typically limited to "live-streaming" informal in-person training sessions. These sessions were augmented with hands-on exercises on live notebooks for remote participants, with almost no assessments of student learning. Unlike select instances such as the PetaScale institute, training curricula were designed with the in-person attendee in mind. Sustained training for RC is particularly important since research computing activities continued to thrive while several other avenues of research were diminished. Here we report on two educational approaches that were implemented at Texas A & M's program in Spring 2020. These sessions were offered over Zoom with the instructor assisted by moderators using the chat features. The first approach duplicated our traditional in-person sessions in an online setting. These sessions were taught by staff and followed a half day approach. A second approach involved the use of shorter pop-up courses in which participants chose the topic matter. This approach implemented a peer-learning environment in which students taught and moderated the training sessions. These sessions were supplemented with YouTube videos. Both programs steadily evolved as we experimented with educational approaches to strengthen our curricula
|2:50 PM -3:00 PM||"The Performance Optimisation and Productivity Centre of Excellence: From Improving Parallel Codes to HPC training and education". Fouzhan Hosseini
HPC applications are often very complex and their behavior depends on a wide range of factors from algorithms, to programming models, library and language implementations and hardware. The task of understanding performance bottlenecks of a parallel code and making improvements often ends up being a daunting trial and error process. To make the problem even more complicated, many HPC applications inherit different layers of legacy code, written and optimized for a different computing era. To optimize the performance of a parallel application, the first step is to understand the behavior of the application. However, there is often a lack of quantitative understanding of the actual behavior of HPC applications. The Performance Optimisation and Productivity (POP) Centre of Excellence, funded by the EU under the Horizon 2020 Research and Innovation Programme, attempts to establish a quantitative methodology for the assessment of parallel codes. This methodology uses a set of hierarchical metrics, where each metric represents relative impact of one cause of inefficiency in parallel codes. These metrics provide a standard, objective way to characterise different aspects of the performance of parallel codes. In this paper, we will describe how the POP methodology allowed us to quickly identify the performance bottlenecks of applications from different domains of science and engineering. In addition, the POP methodology facilitates training HPC experts and performance analysts by defining a common and systematic approach to assessing and improving parallel codes.
|3:00 PM -3:10 PM||Q/A session|
|3:10 PM -3:20 PM||Break|
|3:20 AM -4:00 PM||Session II(Chair: Scott Lathrop)|
|3:20 PM -3:30 PM|| "Learning Portals - Call to Action". Scott Lathrop, Robert Panoff, Julia Mullen, Weronika Filinger, Susan Mehringer and Alan O'Cais
This session will address the lessons learned in providing access to portals. The presenters will describe their experiences working with the community to identify the most important functions including the breadth of content, classifications and organization of materials, reviews of materials, tools for keeping repositories up-to-date, and long-term preservation strategies. The presenters will discuss the positive benefits, the shortcomings, and the opportunities for meeting the needs of the community. We will provide participants with information about a range of portals from around the world. These include: the Computational Science Education Reference Desk (http://www.shodor.org/refdesk/); the National Science Digital Library (https://nsdl.oercommons.org/); the HPC University (http://www.hpcuniversity.org/) portal; the XSEDE training course catalogue (https://portal.xsede.org/training/course-catalog); the ACM SIGHPC Education Chapter High Performance Resources (https://sighpceducation.acm.org/resources.html); the PRACE Training Portal (http://www.training.prace-ri.eu); the Supercomputing Training Portal (http://supercomputing.cyi.ac.cy); the PRACE Best Practice Guides (http://www.prace-ri.eu/best-practice-guides/); the PRACE Code Vault (https://gitlab.com/PRACE-4IP/CodeVault); POP's (Performance Optimisation and Productivity) training materials (https://pop-coe.eu/further-information/learning-material); BioExcel's repository (http://krc.bioexcel.eu/training); CompBioMed (https://www.compbiomed.eu/training/training-repository/); ARCHER's training resources (http://www.archer.ac.uk/training/online/); and the E-CAM Training Portal (https://training.e-cam2020.eu/datasets?size=96). We will continue to seek out other collections and the lessons learned from their developers to share with the participants. The session will emphasize an open discussion with the attendees to identify additional capabilities that should be considered to ensure that portals best serve the international community of computational science and HPC researchers, educators, learners, and practitioners. We will use the community forum to develop recommendations for developers of portals to lay the groundwork for ensuring that portals and collections continue to evolve and improve to best serve their constituents. The session will end with a “Call to Action!” inviting the international community to participate in ongoing discussions with the ACM SIGHPC Education Chapter Education Content Committee. The objective is to develop a collaborative, international community aimed at providing guidance to portal developers for deploying high quality learning materials and resources.
|3:30 PM -3:40 PM||"Booting Up HPC Training Programs @ ODU". Wirawan Purwanto, John Pratt and Terry Stilwell
In this brief talk, we describe our journey to "boot up" the computational training programs at the Old Dominion University (ODU). We began in 2016 by participating as a satellite site of the XSEDE HPC Monthly Workshops. ODU's participation in the DataUp program, sponsored by South Big Data Hub, has enabled us to offer our own training. After several staff members became certified Carpentries instructors in 2019, we began regularly offering Software Carpentry as well as "Introduction to HPC" workshops to our campus community. With the funding from the National Science Foundation, we are also developing a computational-skill training program named "DeapSECURE" (https://deapsecure.gitlab.io/), targeted for students with an interest in the cybersecurity domain. Overall, our training program has increased the student's baseline computational skills. Our experience and lessons learned may be of interest for other universities who are also endeavoring to start their own training program in computing with limited resources.
|3:40 PM -3:50 PM||"Virtual GPU Hackathon: Lessons Learned". Mozhgan Kabiri Chimeh
COVID-19 has changed the landscape across the world as our scientific, medical and technology communities unite in an effort to slow the spread and develop ways to combat the infection. Conferences have started incorporating new platforms to facilitate interaction and engagement among attendees with hundreds of the most talented, experienced scientists and researchers in the world and they are bringing this content to the global community. To ensure that the developer and research communities can still participate in the GPU Hackathons and Bootcamps--enabling them to develop the hands-on skills needed to leverage GPUs to accelerate their scientific codes, we diligently investigated the best online tools available to bring our events into a virtual environment. In this talk, I would like to share lessons learned from our first completely digital GPU Hackathon concluded in May 2020.
|3:50 PM -4:00 PM||"Best Practices for Online Training".Weronika Filinger, David Henty
EPCC has had substantial experience of delivering online training over several years. This includes HPC training for the UK national ARCHER service, assessed courses for Edinburgh University’s online MSc programmes, and a large-scale introductory MOOC for the PRACE project. In this presentation we will briefly discuss our experiences of delivering online training, including methods for supporting remote practical exercises and ways to keep students engaged.
|4:00 PM -4:10 PM||"Incorporating Hands-On Games to Build HPC Intuition". Julia Mullen, Lauren Milechin and Sid Samsi
Serious Games provide pathways for learners of all ages to develop intuition about concepts that are new to them. In this talk we describe our experiences and lessons learned from incorporating serious games into classroom presentations about supercomputing. Two of the games, the Puzzle Game and the Scheduler Game, were integrated into a university workshop on High Performance Computing and the third, the Sorting Game1, was used in a primary school classroom. For each game, we will describe our learning objectives and implementation. We close with a summary of lessons learned and ideas for transitioning to physically distanced classrooms.
|4:10 PM -4:20 PM||"The HPC Certification Forum - Activity Update 2020". Julian Kunkel, Kai Himstedt, Lev Lafayette, Anja Gerbes, Waseem
Kamleh, Christian Meesters, Weronika Filinger
The goal of the HPC Certification Forum is to clearly categorize, define, and examine competencies expected from proficient HPC practitioners. The ultimate goal of the forum is to offer a free globally acknowledged certification program that would make the HPC education and training more transparent and quantifiable for training providers, and easier to navigate for practitioners. The talk gives an overview of the performed activities, the lessons learned and prospects.
|4:20 PM -4:30 PM||Q/A session|
|4:30 PM -4:40 PM||Break|
|4:40 PM -5:20 PM||Panel: "Best Practices for fostering HPC Workforce and Training"
Elizabeth Bautista(Lawrence Berkeley National Laboratory), Jennifer Halstead (Petit Financial Services), LaWana Richmond (University of California, San Diego), Aaron Weeden (Shodor Inc.) and Maciej Cytowski (Pawsey Supercomputing Centre)
|5:20 PM -5:30 PM||Q/A|
|5:30 PM -5:55 PM|| Break-out Sessions:
BREAKOUT ROOM 1 Topic - Learning Portals
BREAKOUT ROOM 2 Topic - Best practices for Online learning
BREAKOUT ROOM 3 Topic - Serious games
BREAKOUT ROOM 4 Topic - HPC Workforce
|5:55 PM -6:00 PM||Review Next Steps and Action Plans|