Metraipolis › archive › dc_a5280af99eee
TinyML Hits the Memory Wall: A Benchmark of Ten Sub-Watt Edge Devices — arXiv:2611.03392
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.03392
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T11:13:29.956Z
- SHA-256 of the archived text
253ba34383931ab7d25e574550f673eb28ff671201e81b9677e0dfc0578a1c1f- Stable URL of this page
https://metraipolis.com/documento/dc_a5280af99eee
How to cite
TinyML Hits the Memory Wall: A Benchmark of Ten Sub-Watt Edge Devices — arXiv:2611.03392 — arXiv (en), https://arxiv.org/abs/2611.03392 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a5280af99eee (Metraipolis archive dc_a5280af99eee, sha256 253ba34383931ab7d25e574550f673eb28ff671201e81b9677e0dfc0578a1c1f)
BibTeX
@misc{dca5280af99eee,
title = {TinyML Hits the Memory Wall: A Benchmark of Ten Sub-Watt Edge Devices — arXiv:2611.03392},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.03392},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a5280af99eee; sha256 253ba34383931ab7d25e574550f673eb28ff671201e81b9677e0dfc0578a1c1f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03392. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Neural network inference at the edge must fit tight energy budgets. Existing benchmarks compare few devices, do not report energy on a common basis, and omit the newest microcontrollers with an NPU or the 1 GHz Cortex-M85. Hardware-aware design also ranks networks by MAC count, yet no study has tested whether this proxy holds across accelerators. We benchmark sub-watt edge AI hardware by evaluating four CNNs (MobileNetV3, MCUNetV1, EfficientNet, SqueezeNet) and two tiny transformers (MobileViT-XXS, DeiT-tiny) on ten commercial devices: an in-sensor DSP (Sony IMX500), four MCUs with an NPU (STM32N6, Alif Ensemble E7/E8, GreenWaves GAP9), and five MCUs without an NPU (STM32U5, Apollo510B, ESP32-S3 Xtensa LX7, RA8M2, STM32V8). Four devices are novel (Alif E8, Apollo510B, RA8M2, STM32V8). Measured as energy above idle, the IMX500 (0.13-0.35 mJ) uses 1.3-3x less energy than the STM32N6 (0.31-1.05 mJ) on every model; the GAP9 uses the least on MobileNetV3 and the Alif E7 the least on MCUNetV1 and SqueezeNet. All four use 6-34x less energy than the STM32V8 (1.85-3.98 mJ), and the Alif E7 and IMX500 form the power-performance Pareto front. Latency on all four accelerators scales with memory traffic (bytes of weights and activations, r >= 0.98) and not with MACs, so they are memory-bound, while no MCU without an NPU is. The NPU advantage grows with intensity from 1.4x over a 1 GHz Cortex-M85 on MCUNetV1 to 8.2x on SqueezeNet. MAC count is therefore a misleading cost proxy for sub-watt accelerators. Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.AR, nuovi lavori del giorno). Categorie arXiv: Hardware Architecture (cs.AR). Pagina del lavoro: https://arxiv.org/abs/2611.03392 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T11:00:22.175Z dalla pagina https://arxiv.org/list/cs.AR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).