EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Measurement Boundaries and Sequence-Dependent Hardware-State Realization in Edge-AI Energy Benchmarking: A Blocked Factorial Study on the NVIDIA Jetson Orin Nano

Adem Tek, Lucas Weißbeck, Alexander Rachmann, Hendrik Poschmann

Electronics · 31.07.2026 · 15 (15) · 3372

DOI: 10.3390/electronics15153372 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 31.07.2026. Verfasst von Adem Tek und 3 weiteren Autor:innen. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Edge artificial-intelligence (AI) inference energy comparisons can mislead when timing, system-energy integration, and requested hardware states are conflated. We audited 558 Jetson Orin Nano runs; the balanced primary design comprised 378 runs on three distinct physical boards, spanning two convolutional neural networks (CNNs), FP16/FP32/FP64 tensor dtypes, seven batch sizes, and three requested profile branches. Forward performance was separated from onboard input-rail (VDD_IN) energy over the logged window. In the full grid, precision accounted for 71.5% of log-scale forward-throughput variation and 66.4% of logged-window energy variation, largely reflecting FP64 stress. In the FP16/FP32 slow + medium sensitivity, batch instead accounted for 54.1% of forward-throughput variation. Across these cross-board-stable branches, FP32 retained 68.4% and 54.8% of FP16 throughput for MobileNetV2 and ResNet-50, while using 1.212 and 1.612 times the logged energy. At batch one, FP16 retained only 85.4% and 84.7% of FP32 throughput. All 126 requested-fast runs exhibited an exact observed association: the realized low/high clock regime matched whether the last non-fast predecessor was slow (57/57) or medium (69/69). Thus, requested profile labels did not define independently realized treatments. Reliable claims require compatible measurement boundaries, explicit protocol overheads, and verified state reset/read-back.

Zitieren

APA

Tek, A., Weißbeck, L., Rachmann, A., & Poschmann, H. (2026). Measurement Boundaries and Sequence-Dependent Hardware-State Realization in Edge-AI Energy Benchmarking: A Blocked Factorial Study on the NVIDIA Jetson Orin Nano. Electronics, 15(15), 3372. https://doi.org/10.3390/electronics15153372

Vancouver

Tek A, Weißbeck L, Rachmann A, Poschmann H. Measurement Boundaries and Sequence-Dependent Hardware-State Realization in Edge-AI Energy Benchmarking: A Blocked Factorial Study on the NVIDIA Jetson Orin Nano. Electronics. 2026;15(15):3372. doi:10.3390/electronics15153372

BibTeX
@article{Tek2026,
  author = {Tek, Adem and Weißbeck, Lucas and Rachmann, Alexander and Poschmann, Hendrik},
  title = {Measurement Boundaries and Sequence-Dependent Hardware-State Realization in Edge-AI Energy Benchmarking: A Blocked Factorial Study on the NVIDIA Jetson Orin Nano},
  journal = {Electronics},
  year = {2026},
  volume = {15},
  number = {15},
  pages = {3372},
  doi = {10.3390/electronics15153372}
}
Kostenlos starten →