pith. sign in

arxiv: 2503.03722 · v4 · pith:IJHN6DABnew · submitted 2025-03-05 · 💻 cs.OS · cs.AR

Where Linux Breaks Under Radiation: A Cross-Architecture Kernel-Level Characterization of Proton-Induced Failures in COTS SoCs

Saad Memon , Rafal Graczyk , Tomasz Rajkowski , Jan Swakon , Damian Wrobel , Sebastian Kusyk , Seth Roffe , Mike Papadakis This is my paper

Pith reviewed 2026-05-23 01:09 UTC · model grok-4.3

classification 💻 cs.OS cs.AR
keywords Linux kernelradiation effectsSEFIproton irradiationCOTS SoCkernel failureseMMCprocess node
0
0 comments X

The pith

On the 14 nm SoC, 90% of proton-induced Linux failures funnel through a single eMMC storage path.

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

The paper subjects three Linux platforms on different process nodes to proton irradiation between 20 and 58 MeV and traces every observed failure back through kernel logs. It finds that failure locations vary sharply by node: memory management and drivers account for most events on the two 40 nm platforms, while the 14 nm device routes nearly all failures through the eMMC subsystem. This concentration matters because it identifies a narrow point where a single peripheral upset can dominate system-level reliability. The work also reconstructs how individual faults can propagate across up to six kernel subsystems before producing a terminal crash.

Core claim

Through kernel log forensics on three irradiated platforms, all 133 observed Linux failures were attributed to specific handlers. On the 14 nm FinFET ARM Cortex A53 SoC, 90% of failures concentrate in one eMMC storage path (56% filesystem failures and 34% driver failures). On the 40 nm platforms, memory management and driver handlers together produce 67 to 78% of events. The 14 nm SoC exhibits roughly an order of magnitude lower Linux SEFI cross section. Reconstructed chains show faults cascading through as many as six kernel subsystems before terminal failure.

What carries the argument

Kernel log forensics that traces each radiation-induced failure to its originating handler in the Linux kernel.

If this is right

  • A SEFI-susceptible peripheral can dictate overall system reliability more than the rest of the kernel combined.
  • Targeted hardening at identified subsystem boundaries can replace blanket redundancy approaches.
  • Failure profiles differ enough by process node that architecture-specific mitigations are required.
  • Propagation chains spanning up to six subsystems indicate the need to monitor handoff points between kernel components.

Where Pith is reading between the lines

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

  • Designers of space systems should first audit storage peripherals on advanced-node COTS parts for radiation vulnerability.
  • The same log-tracing method could be used on other operating systems to locate analogous single-point failure paths.
  • Controlled variation of beam angle in follow-on tests could help separate process-node effects from irradiation geometry.

Load-bearing premise

Kernel logs accurately and completely record every radiation-induced failure and correctly identify its originating handler.

What would settle it

An observed Linux crash under proton irradiation whose cause cannot be matched to any logged handler or for which the logs contain no entry at all.

Figures

Figures reproduced from arXiv: 2503.03722 by Damian Wrobel, Jan Swakon, Mike Papadakis, Rafal Graczyk, Saad Memon, Sebastian Kusyk, Seth Roffe, Tomasz Rajkowski.

Figure 1
Figure 1. Figure 1: Unmitigated radiation-induced critical soft error on [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 1
Figure 1. Figure 1: Likewise, the strong inter-dependencies among Linux [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Block diagram of ARM Cortex-A53 SoC (left) and [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: 2D beam profile, normalised intensity We conducted proton radiation experiments at the IFJ PAN in Krakow, Poland, using a 60 MeV proton beam produced by the AIC-144 isochronous cyclotron [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Linux Kernel Performance Metrics: Idle vs. Stressed ( [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: DUTs in Proton Beam Testing: Raspberry Pi Zero 2W (Left), NXP i.MX 8M Plus (Center), OrangeCrab FPGA (Right) [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: RP3A0 utilizing a System-in-Package (SiP) [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Cross-sections of Linux crashes across all DUTs at [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The pie chart highlights vulnerable components based [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 8
Figure 8. Figure 8: Vulnerability shares for Linux kernel components across different DUTs: Memory management (mm) dominates in [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
read the original abstract

Linux is increasingly deployed in Low Earth Orbit on commercial off the shelf systems on chip that were not designed for space radiation. Ionizing particles can trigger single event functional interrupts that crash the kernel without warning. Prior work mainly measured board level cross sections, leaving unclear which Linux subsystems fail and how a single upset propagates into an operating system wide failure across architectures, stress conditions, and irradiation conditions. We address this gap by subjecting three Linux platforms to proton irradiation in the 20 to 58 MeV range: a Raspberry Pi Zero 2W with a 40 nm planar ARM Cortex A53, an NXP i MX 8M Plus with a 14 nm FinFET ARM Cortex A53, and an OrangeCrab ECP5 FPGA hosting a VexRiscV RV32I soft core at 40 nm. Through kernel log forensics, we trace all 133 observed Linux failures, most of which have not been previously reported, to their originating kernel handlers. Failure profiles differ sharply across nodes. On the two 40 nm platforms, memory management and driver handlers account for 67 to 78% of events, while on the 14 nm SoC approximately 90% of failures funnel through a single eMMC storage path, comprising 56% filesystem failures and 34% driver failures. This shows that a SEFI susceptible peripheral can strongly dictate system reliability. The 14 nm SoC also shows roughly an order of magnitude lower Linux SEFI cross section, although irradiation geometry and DRAM exposure differences preclude isolating the contribution of process scaling. Reconstructed propagation chains show that faults can cascade through up to six kernel subsystems before terminal failure in severe events. Rather than motivating blanket redundancy, these results identify the kernel subsystem boundaries where radiation induced faults originate, enabling targeted mitigations for hardening COTS Linux systems for orbit.

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

3 major / 2 minor

Summary. The paper reports proton irradiation experiments (20-58 MeV) on three COTS Linux platforms (RPi Zero 2W 40 nm ARM, i.MX 8M Plus 14 nm ARM, OrangeCrab ECP5 40 nm RISC-V), tracing 133 kernel failures via log forensics to originating handlers. It finds sharp architecture differences: memory/driver handlers dominate on the 40 nm nodes (67-78%), while ~90% of failures on the 14 nm SoC route through a single eMMC path (56% filesystem, 34% driver). The 14 nm platform also exhibits an order-of-magnitude lower SEFI cross section. The work concludes that a single SEFI-susceptible peripheral can dictate system reliability and that targeted mitigations at subsystem boundaries are preferable to blanket redundancy.

Significance. If the kernel-log tracing is shown to be complete and unbiased, the results supply concrete, architecture-specific failure-mode data that directly informs hardening strategies for COTS Linux in LEO. The identification of a dominant eMMC funnel on the 14 nm node and the reconstruction of multi-subsystem propagation chains are the most actionable contributions; the cross-section comparison is secondary because geometry and DRAM exposure differences are acknowledged.

major comments (3)
  1. [Abstract / §3] Abstract and §3 (methods): the claim that 'we trace all 133 observed Linux failures' to originating handlers is load-bearing for the 90% eMMC funnel result on the 14 nm SoC. The manuscript must demonstrate that silent crashes, log-buffer corruption, or unlogged paths were excluded (e.g., via watchdog timestamps, post-irradiation memory dumps, or controlled injection tests); without such verification the 56%/34% split and the 'single peripheral dictates reliability' conclusion cannot be considered robust.
  2. [Abstract / Results] Abstract and results section: percentages (90%, 56%, 34%, 67-78%) are reported without error bars, binomial confidence intervals, or a statement of the total number of proton events per platform. Because the central claim is a quantitative comparison across nodes, the absence of statistical controls makes it impossible to judge whether the observed differences exceed sampling variation.
  3. [§2] Irradiation setup (presumably §2): full beam parameters (flux, fluence, angle, shielding of DRAM vs. SoC) are not summarized in the abstract and must be tabulated with run-by-run values; the order-of-magnitude cross-section difference cannot be interpreted without these data.
minor comments (2)
  1. [Figures] Figure captions should explicitly state the number of events per platform and the irradiation conditions used for each bar or pie chart.
  2. [Introduction] The term 'SEFI' is used for both board-level and kernel-level events; a short definition distinguishing the two would improve clarity for readers outside radiation-effects communities.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive feedback, which highlights important aspects of methodological transparency and statistical presentation. We respond to each major comment below, indicating planned revisions where appropriate.

read point-by-point responses

  1. Referee: [Abstract / §3] Abstract and §3 (methods): the claim that 'we trace all 133 observed Linux failures' to originating handlers is load-bearing for the 90% eMMC funnel result on the 14 nm SoC. The manuscript must demonstrate that silent crashes, log-buffer corruption, or unlogged paths were excluded (e.g., via watchdog timestamps, post-irradiation memory dumps, or controlled injection tests); without such verification the 56%/34% split and the 'single peripheral dictates reliability' conclusion cannot be considered robust.

    Authors: The 133 failures are defined as those that produced recoverable kernel log entries; the tracing method relies on these logs and handler call stacks. The experimental protocol used a hardware watchdog to detect and log hangs, with timestamps cross-checked against log buffers to rule out corruption in the captured events. Post-irradiation system state was inspected for persistent anomalies not reflected in logs. However, we did not conduct exhaustive memory dumps or fault-injection campaigns across all runs. We will revise §3 to explicitly delimit the scope of 'observed' failures, detail the watchdog and buffer-validation steps, and qualify the eMMC-funnel conclusion accordingly. This constitutes a partial revision. revision: partial

  2. Referee: [Abstract / Results] Abstract and results section: percentages (90%, 56%, 34%, 67-78%) are reported without error bars, binomial confidence intervals, or a statement of the total number of proton events per platform. Because the central claim is a quantitative comparison across nodes, the absence of statistical controls makes it impossible to judge whether the observed differences exceed sampling variation.

    Authors: We agree that binomial confidence intervals and explicit event counts are necessary to support the cross-node comparisons. We will add 95% Wilson-score intervals for all reported percentages, state the total proton events and observed failures per platform, and include these values in both the abstract and results tables. This revision will allow readers to evaluate whether architecture-specific differences exceed sampling variation. revision: yes

  3. Referee: [§2] Irradiation setup (presumably §2): full beam parameters (flux, fluence, angle, shielding of DRAM vs. SoC) are not summarized in the abstract and must be tabulated with run-by-run values; the order-of-magnitude cross-section difference cannot be interpreted without these data.

    Authors: We will add a dedicated table in §2 that compiles run-by-run beam parameters (flux, fluence, energy, angle, and shielding configuration for DRAM versus SoC die). Although some parameters appear in the detailed methods, the table will make them immediately accessible and will be referenced from the abstract and results when discussing cross-section values. This addresses the interpretability concern directly. revision: yes

Circularity Check

0 steps flagged

No circularity: pure experimental reporting of observed failure traces

full rationale

The paper performs proton irradiation experiments on three COTS Linux platforms and reports direct measurements of 133 kernel failures traced via log forensics. No equations, fitted parameters, predictions, or derivations appear in the provided text. The central claim (90% funneling through eMMC on 14 nm SoC) is a direct count from observed events, not a reduction to any self-referential input or self-citation chain. Kernel log tracing is presented as an empirical method without mathematical modeling that could introduce circularity. This is self-contained experimental reporting.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper is an empirical measurement study; it rests on standard domain assumptions about radiation testing rather than new mathematical constructs or fitted parameters.

axioms (1)
  • domain assumption Proton beams in the 20-58 MeV range produce single-event functional interrupts representative of low-Earth-orbit radiation effects on COTS SoCs.
    Invoked by the choice of irradiation conditions and the interpretation of observed SEFIs as space-relevant.

pith-pipeline@v0.9.0 · 5901 in / 1330 out tokens · 71789 ms · 2026-05-23T01:09:29.645771+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

71 extracted references · 71 canonical work pages

  1. [1]

    Linux statistics – 2024

    Branka, “Linux statistics – 2024.” Available: https://truelist.co/blog/ linux-statistics/, 2024. Accessed: Feb. 23, 2025

    work page 2024

  2. [2]

    Android’s latest statistics 2025: How many people have androids?

    Sam, “Android’s latest statistics 2025: How many people have androids?.” Available: https://avada.io/blog/how-many-people-have-androids/, 2025. Accessed: 2025-02-20

    work page 2025

  3. [3]

    Space-operating linux: An operating system for computer vision on commercial-grade equipment in leo,

    E. Miller, C. Heistand, and D. Mishra, “Space-operating linux: An operating system for computer vision on commercial-grade equipment in leo,” in 2023 IEEE Aerospace Conference , pp. 1–12, IEEE, 2023

    work page 2023

  4. [4]

    Current use of linux in spacecraft flight software,

    H. Leppinen, “Current use of linux in spacecraft flight software,” IEEE Aerospace and Electronic Systems Magazine , vol. 32, no. 10, pp. 4–13, 2017

    work page 2017

  5. [5]

    Beck, Evaluation and analysis of Linux-based applications on mul- ticore processor in a space environment

    T. Beck, Evaluation and analysis of Linux-based applications on mul- ticore processor in a space environment . Ph.d. thesis, ISAE-Institut Supérieur de l’Aéronautique et de l’Espace, 2024

    work page 2024

  6. [6]

    Embedded linux plat- form for a fault tolerant space based parallel computer,

    B. Ramesh, T. Bretschneider, and I. McLoughlin, “Embedded linux plat- form for a fault tolerant space based parallel computer,” in Proceedings of the Real-Time Linux Workshop , pp. 39–46, 2004

    work page 2004

  7. [7]

    Is linux ready for space applications,

    S. S. Prieto, I. G. Tejedor, D. Meziat, and A. V . Sánchez, “Is linux ready for space applications,” in Open Source World Conference; Junta de Andalucıa: Mlaga, Spain, 2004

    work page 2004

  8. [8]

    The future of space 2060 and implications for us strategy: Report on the space futures workshop (colorado springs, co: Afspc, september 5, 2019), 6–9

    A. F. S. Command, “The future of space 2060 and implications for us strategy: Report on the space futures workshop (colorado springs, co: Afspc, september 5, 2019), 6–9.”

    work page 2060

  9. [9]

    Modern small satellites-changing the economics of space,

    M. N. Sweeting, “Modern small satellites-changing the economics of space,” Proceedings of the IEEE , vol. 106, no. 3, pp. 343–361, 2018

    work page 2018

  10. [10]

    The paradigm shift of newspace: New business models and growth of the space economy,

    W. Peeters, “The paradigm shift of newspace: New business models and growth of the space economy,” New Space, vol. 12, no. 3, pp. 202–213, 2024

    work page 2024

  11. [11]

    The newspace race: the private space industry and its effect on public support for nasa funding,

    I. Sehovic, “The newspace race: the private space industry and its effect on public support for nasa funding,” tech. rep., working paper, March 2018, 52 p, available at https://www. academia. edu . . . , 2018

    work page 2018

  12. [12]

    Mars helicopter technology demon- strator,

    B. Balaram, T. Canham, C. Duncan, H. F. Grip, W. Johnson, J. Maki, A. Quon, R. Stern, and D. Zhu, “Mars helicopter technology demon- strator,” in 2018 AIAA atmospheric flight mechanics conference , p. 0023, 2018

    work page 2018

  13. [13]

    Radiation-induced soft errors in advanced semiconduc- tor technologies,

    R. C. Baumann, “Radiation-induced soft errors in advanced semiconduc- tor technologies,” IEEE Transactions on Device and materials reliability , vol. 5, no. 3, pp. 305–316, 2005

    work page 2005

  14. [14]

    Risk assessment for the use of cots devices in space systems under consideration of radiation effects,

    J. Budroweit and H. Patscheider, “Risk assessment for the use of cots devices in space systems under consideration of radiation effects,” Elec- tronics, vol. 10, no. 9, p. 1008, 2021

    work page 2021

  15. [15]

    Cots and the newspace,

    J. Estela, “Cots and the newspace,” Radiation Effects on Integrated Circuits and Systems for Space Applications , pp. 329–346, 2019

    work page 2019

  16. [16]

    Single event upsets caused by solar energetic heavy ions,

    A. Tylka, W. Dietrich, P. Boberg, E. Smith, and J. Adams, “Single event upsets caused by solar energetic heavy ions,” IEEE Transactions on Nuclear Science , vol. 43, no. 6, pp. 2758–2766, 1996

    work page 1996

  17. [17]

    Swif-it: A tool for memory fault injection and protection,

    J. Freschl and D. Xue, “Swif-it: A tool for memory fault injection and protection,” 2005

    work page 2005

  18. [18]

    Starlink satellite constellation

    SpaceX, “Starlink satellite constellation.” Online. Available: https://www. starlink.com/, 2025

    work page 2025

  19. [19]

    Starlink statistics,

    J. McDowell, “Starlink statistics,” 2025. Accessed: 2025-02-12

    work page 2025

  20. [20]

    Starlink surpasses 4 million users,

    T. Verge, “Starlink surpasses 4 million users,” September 2024

    work page 2024

  21. [21]

    Spacex: We’ve launched 32,000 linux computers into space for starlink internet,

    Z. Staff, “Spacex: We’ve launched 32,000 linux computers into space for starlink internet,” 2024. Accessed: 2024-12-01

    work page 2024

  22. [22]

    Spacex uses linux and x86 processors in the falcon 9,

    SpaceX, “Spacex uses linux and x86 processors in the falcon 9,” Aerospace Technology Review, 2025. Accessed: 2025-02-19

    work page 2025

  23. [23]

    The connection between space weather and single event upsets in polar low earth orbit satellites,

    S. Katz, U. Goldvais, and C. Price, “The connection between space weather and single event upsets in polar low earth orbit satellites,” Advances in Space Research , vol. 67, no. 10, pp. 3237–3249, 2021

    work page 2021

  24. [24]

    Elc: Spacex lessons learned,

    J. Edge, “Elc: Spacex lessons learned,” Linux Weekly News

    work page

  25. [25]

    Krol, A generalization of fault-tolerance based on masking

    T. Krol, A generalization of fault-tolerance based on masking. PhD thesis, Technische Universiteit Eindhoven, 1991. PhD Thesis 1 (Research TU/e / Graduation TU/e), Electrical Engineering

    work page 1991

  26. [26]

    FCC Order for OneWeb Phase 2 Expansion to 6,372 Satellites

    Federal Communications Commission, “FCC Order for OneWeb Phase 2 Expansion to 6,372 Satellites.” Federal Communications Commission, DA-22-970A1, 2022. [Online]. Available: urlhttps://docs.fcc.gov/public/attachments/DA-22-970A1.pdf

    work page 2022

  27. [27]

    FCC Authorization for Project Kuiper 3,236 Satellites

    Federal Communications Commission, “FCC Authorization for Project Kuiper 3,236 Satellites.” Federal Communications Commission, FCC-20- 102A1, 2020. [Online]. Available: urlhttps://docs.fcc.gov/public/attachments/FCC-20-102A1.pdf

    work page 2020

  28. [28]

    Overview of in-orbit radiation induced spacecraft anomalies,

    R. Ecoffet, “Overview of in-orbit radiation induced spacecraft anomalies,” IEEE Transactions on Nuclear Science , vol. 60, no. 3, pp. 1791–1815, 2013

    work page 2013

  29. [29]

    On-orbit anomalies: Investigations and root cause determination,

    R. Ecoffet et al. , “On-orbit anomalies: Investigations and root cause determination,” in IEEE NSREC, vol. 1198, 2011

    work page 2011

  30. [30]

    Baumann and K

    R. Baumann and K. Kruckmeyer, Radiation handbook for electronics, available on TI.com (https://www.ti.com/applications/industrial/aerospace-defense/technical- documents.html). 01 2019

    work page 2019

  31. [31]

    Scaling and technology issues for soft error rates,

    A. H. Johnston et al., “Scaling and technology issues for soft error rates,” in Proc. of the 4th Annual Conf. on Reliability , Citeseer, 2000

    work page 2000

  32. [32]

    Single-event effects in advanced cmos technologies,

    V . Pouget, “Single-event effects in advanced cmos technologies,” in European Conference on Radiation and its Effects on Components and Systems (RADECS) Short Course , 2023

    work page 2023

  33. [33]

    Linux kernel components have strong inter- dependencies, making it difficult to remove or replace them (Figure 1)

    Unikraft Project, “Linux kernel components have strong inter- dependencies, making it difficult to remove or replace them (Figure 1).” https://unikraft.org/docs/concepts/design-principles. Accessed: 2025-01- 17

    work page 2025

  34. [34]

    Accelerated nuclear radiation effects on the raspberry pi 3b+,

    C. J. Corley, H. M. Quinn, and E. E. Swartzlander, “Accelerated nuclear radiation effects on the raspberry pi 3b+,” in 2022 IEEE Radiation Effects Data Workshop (REDW)(in conjunction with 2022 NSREC) , pp. 1–8, IEEE, 2022

    work page 2022

  35. [35]

    Analyzing the impact of fault-tolerance methods in arm processors under soft errors running linux and parallelization apis,

    G. S. Rodrigues, F. Rosa, Á. B. de Oliveira, F. L. Kastensmidt, L. Ost, and R. Reis, “Analyzing the impact of fault-tolerance methods in arm processors under soft errors running linux and parallelization apis,” IEEE Transactions on Nuclear Science , vol. 64, no. 8, pp. 2196–2203, 2017

    work page 2017

  36. [36]

    Survey of processors for space,

    R. Ginosar, “Survey of processors for space,” Data Systems in Aerospace (DASIA). Eurospace, pp. 1–5, 2012

    work page 2012

  37. [37]

    A detailed methodology to compute soft error rates in advanced technologies,

    M. Riera, R. Canal, J. Abella, and A. Gonzalez, “A detailed methodology to compute soft error rates in advanced technologies,” in 2016 Design, Automation & Test in Europe Conference & Exhibition (DATE) , pp. 217– 222, IEEE, 2016

    work page 2016

  38. [38]

    Accelerated nuclear radiation effects on the raspberry pi 3b+ and pi 4,

    C. J. Corley, H. M. Quinn, and E. E. Swartzlander, “Accelerated nuclear radiation effects on the raspberry pi 3b+ and pi 4,” in2023 IEEE Radiation Effects Data Workshop (REDW)(in conjunction with 2023 NSREC), pp. 1– 9, IEEE, 2023

    work page 2023

  39. [39]

    Raspberry pi zero and 3b+ see and tid test results,

    S. M. Guertin, S. Vartanian, and A. C. Daniel, “Raspberry pi zero and 3b+ see and tid test results,” in 2022 IEEE Radiation Effects Data Workshop (REDW)(in conjunction with 2022 NSREC) , pp. 1–5, IEEE, 2022

    work page 2022

  40. [40]

    Raspberry pis for space guideline

    S. M. Guertin, “Raspberry pis for space guideline.”

    work page

  41. [41]

    Board level proton testing book of knowledge for nasa electronic parts and packaging program,

    S. M. Guertin, “Board level proton testing book of knowledge for nasa electronic parts and packaging program,” 2017

    work page 2017

  42. [42]

    Proton testing of nvidia jetson tx1,

    E. J. Wyrwas, “Proton testing of nvidia jetson tx1,” tech. rep., 2017

    work page 2017

  43. [43]

    Proton testing of nvidia jetson tx2,

    E. J. Wyrwas, “Proton testing of nvidia jetson tx2,” tech. rep., 2019

    work page 2019

  44. [44]

    Analyzing the impact of the operating system on the reliability of a risc-v fpga imple- mentation,

    I. Wali, A. Sánchez-Macián, A. Ramos, and J. A. Maestro, “Analyzing the impact of the operating system on the reliability of a risc-v fpga imple- mentation,” in 2020 27th IEEE International Conference on Electronics, Circuits and Systems (ICECS) , pp. 1–4, IEEE, 2020

    work page 2020

  45. [45]

    it was designed to exercise various physical subsystems of a computer as well as the various operating system kernel interfaces..” https://github.com/ColinIanKing/stress-ng

    “Colinianking/stress-ng: Stress-ng will stress test a computer system in various selectable ways. it was designed to exercise various physical subsystems of a computer as well as the various operating system kernel interfaces..” https://github.com/ColinIanKing/stress-ng. (Accessed on 08/26/2023)

    work page 2023

  46. [46]

    Proton single event effects (see) guideline,

    K. A. LaBel, “Proton single event effects (see) guideline,” Submitted for publication on the NASA Electronic Parts and Packaging (NEPP) Program web site , 2009

    work page 2009

  47. [47]

    Absorbed Dose Determination in External Beam Radiotherapy . No. 398 in Technical Reports Series, Vienna: INTERNATIONAL ATOMIC EN- ERGY AGENCY , 2001

    work page 2001

  48. [48]

    The stopping and range of ions in matter,

    J. F. Ziegler and J. P. Biersack, “The stopping and range of ions in matter,” in Treatise on heavy-ion science: volume 6: astrophysics, chemistry, and condensed matter, pp. 93–129, Springer, 1985

    work page 1985

  49. [49]

    Raspberry Pi Zero 2 W

    “Raspberry Pi Zero 2 W.” https://www.raspberrypi.com/products/ raspberry-pi-zero-2-w/. (Accessed on 08/03/2023)

    work page 2023

  50. [50]

    Radiation damage threshold of satellite cots components: Raspberry pi zero for opal cubesat,

    J. Mojica Decena, B. Wood, R. J. Martineau, M. Tayor, and J. Dennison, “Radiation damage threshold of satellite cots components: Raspberry pi zero for opal cubesat,” in Utah State University Student Research Symposium 2018, 2018

    work page 2018

  51. [51]

    NXP i.MX 8M Plus - verdin system on module

    “NXP i.MX 8M Plus - verdin system on module.” https://www.toradex. com/computer-on-modules/verdin-arm-family/nxp-imx-8m-plus. (Ac- cessed on 08/03/2023)

    work page 2023

  52. [52]

    Portainer: Container management software for kubernetes and docker

    “Portainer: Container management software for kubernetes and docker.” https://www.portainer.io/. (Accessed on 02/01/2024)

    work page 2024

  53. [53]

    Raspberry pi documentation - processors

    “Raspberry pi documentation - processors.” https://www.raspberrypi.com/ documentation/computers/processors.html. (Accessed on 08/26/2023)

    work page 2023

  54. [54]

    An empirical study of operating systems errors,

    A. Chou, J. Yang, B. Chelf, S. Hallem, and D. Engler, “An empirical study of operating systems errors,” in *Proc. 18th ACM Symp. on Operating Systems Principles*, pp. 73–88, 2001

    work page 2001

  55. [55]

    Orangecrab

    “Orangecrab.” https://orangecrab-fpga.github.io/orangecrab-hardware/r0. 2/. (Accessed on 08/03/2023)

    work page 2023

  56. [56]

    Github - spinalhdl/vexriscv: A fpga friendly 32 bit risc-v cpu im- plementation

    “Github - spinalhdl/vexriscv: A fpga friendly 32 bit risc-v cpu im- plementation.” https://github.com/SpinalHDL/VexRiscv. (Accessed on 08/03/2023)

    work page 2023

  57. [57]

    Github - enjoy-digital/litex: Build your hardware, easily!

    “Github - enjoy-digital/litex: Build your hardware, easily!.” https://github. com/enjoy-digital/litex. (Accessed on 08/03/2023)

    work page 2023

  58. [58]

    A catalog and in-hardware evaluation of open-source drop-in compatible risc-v softcore processors,

    C. Heinz, Y . Lavan, J. Hofmann, and A. Koch, “A catalog and in-hardware evaluation of open-source drop-in compatible risc-v softcore processors,” in 2019 International Conference on ReConFigurable Computing and FPGAs (ReConFig), pp. 1–8, IEEE, 2019

    work page 2019

  59. [59]

    Buildroot - making embedded linux easy

    “Buildroot - making embedded linux easy.” https://buildroot.org/. (Ac- cessed on 08/03/2023)

    work page 2023

  60. [60]

    Single Event Functional Interrupt (SEFI),

    JEDEC Solid State Technology Association, “Single Event Functional Interrupt (SEFI),” 2025. Accessed: 2025-01-14

    work page 2025

  61. [61]

    Fysv445 (ph1) radiation interactions in electronic materi- als, jss28, jyu, finland (2018)

    A. Javanainen, “Fysv445 (ph1) radiation interactions in electronic materi- als, jss28, jyu, finland (2018).” http://research.jyu.fi/radef/JSS28/PH1/#1,

    work page 2018

  62. [62]

    (Accessed on 02/05/2024). 14

    work page 2024

  63. [63]

    The concept of confidence or fiducial limits applied to the poisson frequency distribution,

    W. E. Ricker, “The concept of confidence or fiducial limits applied to the poisson frequency distribution,” Journal of the American Statistical Association, vol. 32, no. 198, pp. 349–356, 1937

    work page 1937

  64. [64]

    Nxp debuts i.mx neural process- ing

    NXP Semiconductors, “Nxp debuts i.mx neural process- ing.” https://www.nxp.com/company/about-nxp/newsroom/ NW-DEBUT-IMX-NEURAL-PROCESSING. Accessed: 2025-01- 20

    work page 2025

  65. [65]

    Creme96: A revision of the cosmic ray effects on micro-electronics code,

    A. J. Tylka, J. H. Adams, P. R. Boberg, B. Brownstein, W. F. Dietrich, E. O. Flueckiger, E. L. Petersen, M. A. Shea, D. F. Smart, and E. C. Smith, “Creme96: A revision of the cosmic ray effects on micro-electronics code,” IEEE Transactions on Nuclear Science , vol. 44, no. 6, pp. 2150– 2160, 1997

    work page 1997

  66. [66]

    Creme-mc weibull distribution help,

    Vanderbilt University, ISDE, “Creme-mc weibull distribution help,” 2025. Accessed: 2025-02-21

    work page 2025

  67. [67]

    Performance Impact of Docker Contain- ers on Linux

    L. Ferraz, “Performance Impact of Docker Contain- ers on Linux.” Available: https://www.torizon.io/blog/ containers-in-linux-and-performance-impact, Dec. 13, 2022. Accessed: Apr. 3, 2025

    work page 2022

  68. [68]

    Btrfs - a modern copy-on-write filesystem for linux

    C. Mason and the Btrfs Community, “Btrfs - a modern copy-on-write filesystem for linux.” https://btrfs.wiki.kernel.org, 2007. Accessed: 2025- 03-02

    work page 2007

  69. [69]

    Litedram ecc module

    EnjoyDigital, “Litedram ecc module.” https://github.com/enjoy-digital/ litedram/blob/master/litedram/frontend/ecc.py, 2024. Accessed: 2024-10- 23

    work page 2024

  70. [70]

    Fault tolerant system design patterns for spacex falcon 9 flight software,

    N. Rajadurai, “Fault tolerant system design patterns for spacex falcon 9 flight software,” EmbeddedRelated, 2020. Accessed: 2024-10-23. Saad Memon received the Erasmus Mundus Eu- ropean Double M.Sc. degree in Dependable Soft- ware Systems from the University of St. Andrews, UK, in 2019, with a specialization in Artificial In- telligence. He is currently p...

    work page 2020

  71. [71]

    He is best known for his work on Mutation Testing, for which he has been awarded the IEEE TCSE Rising Star Award 2020

    His research interests include software testing, static analysis, prediction modeling, and search-based software engineering. He is best known for his work on Mutation Testing, for which he has been awarded the IEEE TCSE Rising Star Award 2020. He has been awarded several ACM SIGSOFT Distinguished Paper and Artifact Awards and a Facebook Research Award (2...

    work page 2020