Metraipolis › archive › dc_71d6603517e7

DEX: Digit-Level Early Exit for Energy-Efficient MSDF Neural Network Inference — arXiv:2610.11748

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.11748
License (as archived): CC0-1.0
Archived: 2026-10-09T06:00:34.626Z
SHA-256 of the archived text: 5cd6fb4cad92e1ec967420f25782a2424b7d1b4569534cbe91d49ccb5c20d684
Stable URL of this page: https://metraipolis.com/documento/dc_71d6603517e7

How to cite

DEX: Digit-Level Early Exit for Energy-Efficient MSDF Neural Network Inference — arXiv:2610.11748 — arXiv (en), https://arxiv.org/abs/2610.11748 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_71d6603517e7 (Metraipolis archive dc_71d6603517e7, sha256 5cd6fb4cad92e1ec967420f25782a2424b7d1b4569534cbe91d49ccb5c20d684)

BibTeX

@misc{dc71d6603517e7,
  title = {DEX: Digit-Level Early Exit for Energy-Efficient MSDF Neural Network Inference — arXiv:2610.11748},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.11748},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_71d6603517e7; sha256 5cd6fb4cad92e1ec967420f25782a2424b7d1b4569534cbe91d49ccb5c20d684},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11748. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

U-Net inference for brain-tumor segmentation requires billions of multiply-accumulate operations, motivating hardware that can reduce computation dynamically rather than relying only on fixed precision or static model compression. Most-significant-digit-first (MSDF) arithmetic exposes the leading digits of a result during computation, enabling output-dependent decisions before the full value is generated. This paper presents an MSDF accelerator for quantized U-Net segmentation with a two-stage grouped processing element supporting signed INT8 operands and in-stream bias accumulation. Four runtime mechanisms operate directly on the output digit stream: exact early negative detection (END) in ReLU layers, exact sign-only decision making in the segmentation head, calibrated low-order-digit skipping, and calibrated pruning. The two approximate mechanisms are selected offline under an accuracy constraint, while execution requires only lightweight control and does not modify the stored weights. On a residual U-Net trained with nnU-Net for BraTS, the proposed mechanisms reduce digit cycles by 38.38\% while achieving a mean Dice score of 80.58\% on 73 held-out cases, compared with 81.20\% for the floating-point model; the exact mechanisms alone reduce cycles by 18.79\% without altering the quantized output. Synthesized in 45~nm, the processing element operates at 500~MHz, occupies 0.858~mm$^2$, and consumes 0.726~mJ per $192\times192$ patch under switching-activity-annotated power analysis. A projected eight-output accelerator with shared activation delivery achieves 16.6~ms latency and 1.67~mJ per patch.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Hardware Architecture (cs.AR) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.11748 — 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 2026-10-09T05:47:01.895Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).