From the supercomputers used to forecast weather and develop vaccines to multicore desktops, laptops, and mobile devices, most people utilize some form of high-performance computing (HPC) and/or parallel and distributed computing (PDC), whether directly or indirectly. As a result, companies and laboratories need people with the skills to efficiently utilize these modern computing systems. It is therefore important that all computer science (CS), computer engineering (CE), data science (DS), and similar compute-intensive programs equip their students with HPC/PDC skills. However, rapid changes in hardware platforms, programming languages, and development environments increasingly challenge educators to decide what to teach and how to teach it, in preparing students for careers in HPC/PDC.

The EduHPC workshop invites unpublished manuscripts from academia, industry, and national laboratories on topics in high-performance computing (HPC), parallel and distributed computing (PDC), distributed data science (DS), scalable AI and machine learning (AI/ML), and the Internet of Things and Edge computing (IoT/Edge) as they relate to undergraduate and graduate education, professional training and workforce development. Because of the growing importance of AI workloads on HPC systems, this year’s workshop will have a special focus on ML/DL training at scale. Past workshops have included papers from Computer Science, Computational Science and Engineering, Data Science, and computational courses in STEM and non-STEM disciplines.

The workshop brings together stakeholders from industry (developers, hardware and software vendors), national labs, and academia in the context of SC (Supercomputing), to hear the pedagogical challenges others are facing, share approaches to meeting such challenges, and generally exchange ideas related to HPC/PDC/DS/AI/ML/IoT education. In addition to paper presentations, this workshop will feature panels (e.g., sustainability and reproducibility in technical education), special sessions such as “Peachy Assignments,” and “Lightning Talks” and opportunities for collaboration, resource sharing, educator training, internships, and other means of increasing cross-fertilization between industry, government, and academia. At this year’s EDUHPC, Dr. Kathy Yelick from UC Berkeley will give an invited talk on HPC workforce development

Topics of interest include, but are not limited to:

• Pedagogical issues in incorporating HPC, PDC, DS, AI/ML and IoT/Edge into undergraduate and graduate education, especially in core courses.
• Evidence-based educational practices and novel ways of teaching HPC/PDC/DS/AI/ML/IoT/Edge topics.
• Professional training and workforce development initiatives in HPC/PDC/DS/AI/ML/IoT/Edge Teaching HPC, PDC and IoT/Edge in a data science context, especially data science degree programs.
• Experience reports on incorporating HPC and PDC topics into core CS/CE courses and into STEM fields and other areas.
• Pedagogical tools, materials, infrastructures, languages, and projects for HPC/PDC/DS/AI/ML/IoT/Edge education. Employers' experiences with and/or expectations of new graduates’ HPC/PDC/IoT/Edge proficiency.
• Education resources based on higher-level programming languages including but not limited to Chapel, Haskell, Scala, Python, Modern C++, and Julia and environments/frameworks such as Cilk, CUDA, SYCL, OpenCL, OpenACC, OpenMP, oneAPI, Hadoop, and Spark.
• Projects or units that introduce students to concepts relevant to distributed computing at scale, including IoT/Edge, networking, mobile devices, sensor networks, and similar topics.
• Efforts to evaluate generative AI (e.g. Chat-GPT) impacts on teaching HPC. Issues, experiences and best practices in addressing the gender gap in computing disciplines and broadening participation of underrepresented groups and enhancing the pipeline to research/development careers in HPC and related areas.

SC Reproducibility Initiative: In light of SC's focus on reproducibility, we also welcome manuscripts in which authors and/or students describe their attempts (or failures) to reproduce previously published PDC and HPC research results. We encourage all authors to provide an artifact description appendix (up to two pages) with their manuscript, describing the details of their software environments and computational experiments so that an independent study can replicate their results. In the context of educational research, the artifact description appendix may contain the detailed description of the tools or techniques, classroom evaluation settings, metrics, evaluation results, etc. More information on SC's reproducibility initiative can be found here.

EduHPC-23 Invited Speaker: Dr. Kathy Yelick (Vice Chancellor, Research and the Robert S. Pepper Professor of Electrical Engineering and Computer Sciences, University of California, Berkeley)

Submission Guidelines

We accept submissions for three categories: Full Papers (6-8 pages), Lightning Talks (2-page abstract) and Peachy Parallel Assignments (1-2-page abstract). Please see the preceding topic list, the descriptions below for additional details.

1. Full Papers: Authors should submit 6-8 page papers in PDF format through SC's Linklings submission site. The 8-page limit includes figures, tables, and references. An additional 2 pages can be used for the optional Artifact Description (AD) appendix. Submissions should be formatted using the ACM proceedings template (Note the change from the previously used IEEE format). Latex users, please use the “sigconf” option. Word authors can use the “Interim Layout”. All submitted papers will be peer-reviewed (single blind) and will be considered for a Best Paper award.
2. Lightning Talks: A Lightning Talk is a short (5-10 minute) presentation of a new or innovative idea, preliminary or intermediate research results, opportunity for partnership and collaboration, preliminary curriculum adoption experience of TCPP early adopters, or similar brief talk that will be of interest to HPC/PDC educators. To propose a Lightning Talk, submit a (maximum) 2-page PDF document via SC's Linklings submission site that includes the following elements: an abstract of 250 words, a description of the challenge/problem/educational need that your submission is addressing, a detailed description of the work and its significance or novelty, and lessons learned or potential insights to be drawn from your talk. The 2-page limit includes figures, tables, and references. An additional 2 pages can be used for the optional Artifact Description (AD) appendix. Submissions should be formatted using the ACM proceedings template. The accepted lightning talks will be further curated to create a summary paper co-authored by the Lightning Talk chair and the talk authors.
3. Peachy Parallel Assignments: Peachy Parallel Assignments are HPC/PDC educational assignments that cover any topics related to HPC or PDC education, and are:
   1. Tested - Assignments should have been used successfully in a class.
   2. Adoptable - Preference will be given to assignments that are widely applicable and easy to adopt. Traits of such assignments include coverage of widely taught concepts, use of common parallel languages and widely available hardware, having few prerequisites, and (with variations) being appropriate for different levels of students.
   3. Engaging and Inspirational - Peachy assignments should be fun and inspiring for students, encouraging students to spend time with the relevant concepts. Ideal assignments are those that students want to demonstrate to their roommate.
   Peachy assignments should be submitted through SC's Linklings submission site as a 1-2 page PDF that includes a 250-word abstract describing the assignment and the context in which it was used, plus a link to a public web page containing the complete set of files actually given to students (assignment description, scaffolding code, etc.). The rest of the document should include the following: What is the main idea of the assignment? What concepts are covered? Who are its targeted students? In what context was it used? What prerequisites does it assume of the students? What are its strengths and weaknesses? Are there variations that may be of interest? Between the submission and the online materials, it should be clear what an instructor needs to do in order to use the assignment. The 2-page limit includes figures, tables, and references. An additional 2 pages can be used for the optional Artifact Description (AD) appendix. Submissions should be formatted using the ACM proceedings template. The accepted peachy assignments will be further curated to create a summary paper co-authored by the Peachy Assignments chair and the assignments' authors. Authors will also be required to upload materials needed to give the assignment to a permanent repository (e.g. arXiv or figshare).

PUBLICATION

All accepted full papers, the Lightning Talks summary paper, and the Peachy Assignments summary paper will be published in the SC Workshop Proceedings and will be included in the IEEE Xplore digital library. The accepted peachy assignment abstracts and course materials will be published on the CDER website and archived with a permanent DOI at arXiv.org or figshare.

TRAVEL SUPPORT

Limited travel support will be available for TCPP Early Adopters and Trainees to attend EduHPC-23 and SC23.