pith. sign in

arxiv: 2606.31300 · v1 · pith:OGMQXFPQnew · submitted 2026-06-30 · 💻 cs.DC

An Empirical Analysis of High-Performance Computing Education in Germany

Anna-Lena Roth , Jonas Posner This is my paper

Pith reviewed 2026-07-01 03:32 UTC · model grok-4.3

classification 💻 cs.DC
keywords high-performance computingHPC educationcurriculum analysisGermanyempirical studycluster infrastructurepractical competenciestheoretical instruction
0
0 comments X

The pith

German universities teach HPC mostly as theory with limited hands-on cluster access for students.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper reviews module handbooks and infrastructure records from 102 German institutions to map HPC course offerings and their relation to available clusters. It finds that 67.6 percent of institutions offer at least one HPC course, but these are mainly master's-level electives with thin coverage in bachelor's programs. While 61.8 percent of institutions run HPC clusters, only 23 percent document them as available for teaching, and this restriction statistically links to weaker coverage of practical skills such as resource management, parallel debugging, and performance analysis. The central finding is therefore a structural tilt toward theoretical instruction over the development of applied competencies.

Core claim

An empirical review of 178 HPC-related courses shows that German higher education places HPC instruction predominantly at the master's level as electives, with limited bachelor's integration, and that restricted documented access to local HPC clusters for teaching purposes correlates with reduced curricular emphasis on practical competencies including cluster usage, resource management, parallel debugging, and performance analysis.

What carries the argument

The statistical association drawn between documented teaching access to institutional HPC clusters and the practical-competency coverage recorded in module handbooks.

If this is right

  • HPC courses remain mostly elective modules at the master's level with limited bachelor's integration.
  • Most institutions that operate clusters do not document them for educational use.
  • Restricted teaching access associates with measurably lower emphasis on practical competencies such as parallel debugging and performance analysis.
  • The overall pattern produces a documented imbalance between theoretical and applied HPC instruction.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Increasing documented teaching access to existing clusters could shift curricula toward more practical competency coverage.
  • The same method of handbook-plus-infrastructure review could be applied in other countries to test whether the imbalance is Germany-specific.
  • Institutional policies that reserve clusters exclusively for research may unintentionally limit student preparation for real HPC workloads.

Load-bearing premise

Module handbooks and course catalogs accurately reflect the competencies actually taught and practiced in the courses.

What would settle it

Direct observation or instructor surveys documenting substantial practical HPC activities in courses whose handbooks list few such components would undermine the reported imbalance.

Figures

Figures reproduced from arXiv: 2606.31300 by Anna-Lena Roth, Jonas Posner.

Figure 1
Figure 1. Figure 1: Percentage of HPC courses examined covering HPC topics (n = 178). [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Local HPC clusters at German academic institutions by number of [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of HPC segments in the Top500 list (left) and academic [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
read the original abstract

The growing importance of High-Performance Computing (HPC) requires the systematic integration of parallel programming and performance-oriented competencies into computational science curricula. Effective HPC education combines theoretical foundations with practical experience on real cluster infrastructures, enabling students to understand scalability, efficiency, and architectural differences between shared and distributed memory systems. However, cross-institutional evidence on how HPC education is implemented, and how curricula relate to locally available infrastructure, remains limited. We address this gap through a systematic empirical assessment of HPC education at 102 academic institutions in Germany. Based on module handbooks and course catalogs, we identified 178 HPC-related courses and evaluated their competency coverage and curricular placement. We additionally assessed local academic HPC cluster infrastructures with respect to availability, size, and documented accessibility for teaching. The results show that 67.6% of institutions offer at least one HPC-related course, but these offerings are predominantly elective modules at the master's level, with limited integration in bachelor's programs. Although 61.8% of institutions operate HPC clusters, only 23.0% explicitly document their availability for educational use, as infrastructures are mainly reserved for research. Statistical analysis indicates a significant association between restricted teaching access and reduced curricular emphasis on practical competencies such as resource management, cluster usage, parallel debugging, and performance analysis. Overall, the findings reveal a structural imbalance between theoretical instruction and the development of practical HPC competencies in German higher education.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper conducts a systematic document review of module handbooks and course catalogs from 102 German academic institutions, identifying 178 HPC-related courses. It evaluates competency coverage (theoretical foundations versus practical skills such as resource management, cluster usage, parallel debugging, and performance analysis), curricular placement (predominantly master's electives), and correlates these with local HPC cluster availability (61.8% of institutions) and documented teaching access (only 23.0%). Statistical tests show a significant association between restricted cluster access and lower emphasis on practical competencies, leading to the claim of a structural imbalance favoring theory over practice in German HPC education.

Significance. If the proxy validity of handbooks holds, the study supplies the first large-scale, cross-institutional empirical baseline on HPC education in Germany. The scale (102 institutions, 178 courses) and the infrastructure correlation are strengths that could inform national curriculum guidelines and funding priorities for computational science programs.

major comments (2)
  1. [Methods] Methods (document review and competency coding): competency coverage for practical skills (resource management, parallel debugging, performance analysis) is inferred exclusively from module handbook descriptions without any triangulation against actual syllabi, assignments, or instructor reports. This assumption is load-bearing for the central imbalance claim and the reported association with restricted cluster access; if handbooks systematically list aspirational rather than delivered content, both the imbalance statistic and the infrastructure correlation become unreliable.
  2. [Results] Results (statistical association): the claim of a 'significant association' between restricted teaching access and reduced practical competencies is presented without reporting the exact test statistic, p-value, effect size, or control for confounders such as institution size or discipline. This detail is required to assess whether the association supports the structural-imbalance conclusion.
minor comments (2)
  1. [Abstract] The abstract and methods should explicitly state the inter-rater reliability or coding protocol used for the 178 courses to allow reproducibility.
  2. [Results] Table or figure presenting the competency coverage breakdown by institution type or level (bachelor vs. master) would improve clarity of the placement findings.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, which helps clarify the methodological foundations and statistical reporting in our study. We respond to each major comment below and indicate planned revisions.

read point-by-point responses

  1. Referee: [Methods] Methods (document review and competency coding): competency coverage for practical skills (resource management, parallel debugging, performance analysis) is inferred exclusively from module handbook descriptions without any triangulation against actual syllabi, assignments, or instructor reports. This assumption is load-bearing for the central imbalance claim and the reported association with restricted cluster access; if handbooks systematically list aspirational rather than delivered content, both the imbalance statistic and the infrastructure correlation become unreliable.

    Authors: We acknowledge that module handbooks provide only a proxy for delivered content and may include aspirational elements. This is a recognized limitation of large-scale document-based curriculum studies, chosen here to enable systematic coverage of 102 institutions. In revision we will add an explicit subsection in Methods and a dedicated Limitations paragraph discussing proxy validity, citing comparable studies in higher-education research, and outlining the need for future triangulation via instructor surveys or syllabus collection. revision: yes

  2. Referee: [Results] Results (statistical association): the claim of a 'significant association' between restricted teaching access and reduced practical competencies is presented without reporting the exact test statistic, p-value, effect size, or control for confounders such as institution size or discipline. This detail is required to assess whether the association supports the structural-imbalance conclusion.

    Authors: We agree that the statistical details must be reported explicitly. The revised Results section will state the exact test (chi-square test of independence), the test statistic, p-value, effect size (Cramér’s V), and the outcome of stratified checks by institution size and type (university vs. university of applied sciences). These additions will allow readers to evaluate the robustness of the reported association. revision: yes

Circularity Check

0 steps flagged

Purely observational empirical study with no derivations or self-referential elements

full rationale

The paper is an empirical assessment that counts and classifies 178 HPC-related courses from module handbooks and catalogs across 102 German institutions, evaluates cluster infrastructures, and reports statistical associations. No equations, fitted parameters, predictions, or derivations appear in the provided text. Results are presented as direct observations from external institutional documents rather than constructed from internal definitions or self-citations. This matches the default expectation of no significant circularity for observational work.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that publicly available module handbooks and course catalogs accurately capture taught competencies and that the 102 institutions adequately represent German academic HPC education.

axioms (1)
  • domain assumption Module handbooks and course catalogs accurately reflect the curriculum and competencies covered in actual courses.
    The identification of 178 courses and evaluation of competency coverage depends on treating these documents as faithful representations of teaching practice.

pith-pipeline@v0.9.1-grok · 5777 in / 1321 out tokens · 64848 ms · 2026-07-01T03:32:53.089349+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

27 extracted references · 14 canonical work pages · 1 internal anchor

  1. [1]

    Sustainability13(2021).https://doi.org/10.3390/su132111782

    Al-Jody, T., Aagela, H., Holmes, V.: Inspiring the Next Generation of HPC En- gineers with Reconfigurable, Multi-Tenant Resources for Teaching and Research. Sustainability13(2021).https://doi.org/10.3390/su132111782

    work page doi:10.3390/su132111782 2021

  2. [2]

    In: Proceedings of EduHPC (2018).https://doi.org/10.1109/EduHPC.2018.00008

    Banchelli, F., Mantovani, F.: Filling the Gap Between Education and Industry: Evidence-Based Methods for Introducing Undergraduate Students to HPC. In: Proceedings of EduHPC (2018).https://doi.org/10.1109/EduHPC.2018.00008

    work page doi:10.1109/eduhpc.2018.00008 2018

  3. [3]

    Bischof, C.: Hessisches Kompetenzzentrum für Hochleistungsrechnen (HKHLR) (2026),https://www.hkhlr.de/de

    2026

  4. [4]

    Journal of Parallel and Distributed Computing201(2025)

    Bouvry, P., Brorsson, M., Canal, R., Eftekhari, A., Höfinger, S., Smets, D., Köstler, H., Kozubek, T., Krishnasamy, E., Llosa, J., Lukas-Rother, A., Martorell, X., Pleiter, D., Proykova, A., Sancho, M., Schenk, O., Silvano, C.: The European Master for HPC Curriculum. Journal of Parallel and Distributed Computing201(2025). https://doi.org/10.1016/j.jpdc.20...

    work page doi:10.1016/j.jpdc.2025.105081 2025

  5. [5]

    The Lustre Storage Architecture

    Braam, P.: The Lustre Storage Architecture (2019),http://arxiv.org/abs/1903. 01955, arXiv:1903.01955

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  6. [6]

    BSC Tools: Extrae (2026),https://tools.bsc.es/extrae

    2026

  7. [7]

    BSC Tools: Paraver (2026),https://tools.bsc.es/paraver

    2026

  8. [8]

    bwHPC: High Performance Computing, Data Intensive Computing and Large Scale Scientific Data Management in Baden-Württemberg (2025),https://www.bwhpc.de

    2025

  9. [9]

    In: Practice and Experience in Advanced Research Computing

    Callaghan, S., Filinger, W., Zhukov, I., Lederer, H., Urbanic, J.: Fifteen Years of International HPC Summer School. In: Practice and Experience in Advanced Research Computing. ACM (2025).https://doi.org/10.1145/3708035.3736011

    work page doi:10.1145/3708035.3736011 2025

  10. [10]

    io/sua4v_v1

    Diehl, P., Li, Y.W., Junghans, C., Holmen, J.K., MacCarthy, E., Parete-Koon, S., He, Y.H., Hartman-Baker, R., Lively, C., Gott, K., Gupta, L., Streu, K., Ghadar, Y., Kinsley, P., Herriman, J., Draeger, E.W., Eijkhout, V., Mehringer, S.: Shaping the Future Workforce: Challenges and Lessons Learned in HPC Education from National Labs and Computing Centers (...

    work page doi:10.35542/osf 2025

  11. [11]

    The Journal of Computational Science Education (2021).https://doi.org/10.22369/issn.2153-4136/12/3/2

    Döpper, T., Große-Wöhrmann, B., Lindner, D., Milakovic, D., Oexle, J., Resch, M.M., Scheel, O., Slotosch, S., Widmaier, L.: Expanding HLRS Academic HPC Sim- ulation Training Programs to More Target Groups. The Journal of Computational Science Education (2021).https://doi.org/10.22369/issn.2153-4136/12/3/2

    work page doi:10.22369/issn.2153-4136/12/3/2 2021

  12. [12]

    Education and Information Technologies29(2024)

    Hebert, J., Hratisch, R., Gomes, R., Kunkel, W., Marshall, D., Ghosh, A., Doss, I., Ma, Y., Stedman, D., Stinson, B., Varghese, A., Mohr, M., Rozario, P., Bhattacharyya, S.: High-Performance Computing in Undergraduate Edu- cation at Primarily Undergraduate Institutions in Wisconsin: Progress, Chal- lenges, and Opportunities. Education and Information Tech...

    work page doi:10.1007/s10639-024-12582-6 2024

  13. [13]

    Herold, F., Breuner, S., Heichler, J.: An Introduction to BeeGFS. Tech. rep., ThinkParQ, Kaiserslautern, Germany (2014), https://www.beegfs.io/docs/ whitepapers/Introduction_to_BeeGFS_by_ThinkParQ.pdf An Empirical Analysis of HPC Education in Germany 15

    2014

  14. [14]

    Hewlett Packard Enterprise: El Capitan (2025), https://www.hpe.com/us/en/ compute/hpc/lawrence-livermore-national-laboratory.html

    2025

  15. [15]

    Informationstechnische Gesellschaft im VDE and Gesellschaft für Infor- matik e.V.: Curriculum für Bachelor- und Masterstudiengänge Technische Informatik (March 2018), https://dl.gi.de/server/api/core/bitstreams/ 031111d5-9360-43e7-aed4-54bb34950308/content

    2018

  16. [16]

    Jülich Supercomputing Centre: JUPITER|The Arrival of Exascale in Europe (2025), https://www.fz-juelich.de/en/jsc/jupiter

    2025

  17. [17]

    In: Tools for High Performance Computing - Proceedings of the International Workshop on Parallel ToolsforHPC.Springer(2008)

    Knüpfer, A., Brunst, H., Doleschal, J., Jurenz, M., Lieber, M., Mickler, H., Müller, M.S., Nagel, W.E.: The Vampir Performance Analysis Tool-Set. In: Tools for High Performance Computing - Proceedings of the International Workshop on Parallel ToolsforHPC.Springer(2008). https://doi.org/10.1007/978-3-540-68564-7_9

    work page doi:10.1007/978-3-540-68564-7_9 2008

  18. [18]

    Leibniz Supercomputing Centre: SuperMUC-NG (2026),https://doku.lrz.de/ supermuc-ng-10745965.html

    2026

  19. [19]

    nhr-verein.de/

    NHR-Verein e.V.: Nationales Hochleistungsrechnen (2026), https://www. nhr-verein.de/

    2026

  20. [20]

    II - Big Data, Energy, and Distributed Computing

    Prasad, S.K., Weems, C.C., Sussman, A., Gupta, A., Estrada, T., Vaidyanathan, R., Ghafoor, S.K., Kant, K., Stunkel, C.B.: NSF/IEEE-TCPP Curriculum on Parallel and Distributed Computing for Undergraduates - V. II - Big Data, Energy, and Distributed Computing. In: Proceedings of the Technical Symposium on Computer Science Education. ACM (2023).https://doi.o...

    work page doi:10.1145/3545947.3569594 2023

  21. [21]

    In: Proceedings of the Working Group Reports on ITiCSE

    Raj, R.K., Romanowski, C.J., Impagliazzo, J., Aly, S.G., Becker, B.A., Chen, J., Ghafoor, S., Giacaman, N., Gordon, S.I., Izu, C., Rahimi, S., Robson, M.P., Thota, N.: High Performance Computing Education: Current Challenges and Future Directions. In: Proceedings of the Working Group Reports on ITiCSE. ACM (2020). https://doi.org/10.1145/3437800.3439203

    work page doi:10.1145/3437800.3439203 2020

  22. [22]

    The Journal of Computational Science Education (2024).https://doi.org/10

    Raoofy, A., Elis, B., Bode, V., Chung, M.T., Breiter, S., Schlemon, M., Herr, D.F., Fuerlinger, K., Schulz, M., Weidendorfer, J.: BEAST Lab: A Practical Course on Experimental Evaluation of Diverse Modern HPC Architectures and Accelerators. The Journal of Computational Science Education (2024).https://doi.org/10. 22369/issn.2153-4136/15/1/5

    2024

  23. [23]

    Roth, A.L.: HPC at German Universities - Data analysis references [Data set] (2025).https://doi.org/10.5281/zenodo.18722580

    work page doi:10.5281/zenodo.18722580 2025

  24. [24]

    Samuel, J., Brennan-Tonetta, M., Samuel, Y., Subedi, P., Smith, J.: Strategies for Democratization of Supercomputing: Availability, Accessibility and Usability of High Performance Computing for Education and Practice of Big Data Analytics (2021),https://arxiv.org/abs/2104.09091, arXiv:2104.09091

    work page arXiv 2021

  25. [25]

    Strohmaier, E., Dongarra, J., Simon, H., Meuer, H.: Top500 The List (2025), https://top500.org

    2025

  26. [26]

    University Paderborn: HPC.NRW (2026),https://hpc.dh.nrw/de/

    2026

  27. [27]

    The Journal of Computational Science Education (2020).https://doi.org/10.22369/issn.2153-4136/11/1/12

    Wofford, J.L., Lueninghoener, C.: The Supercomputer Institute: A Systems-Focused Approach to HPC Training and Education. The Journal of Computational Science Education (2020).https://doi.org/10.22369/issn.2153-4136/11/1/12

    work page doi:10.22369/issn.2153-4136/11/1/12 2020