Metraipolis › archive › dc_069be2cb0012
Right In-Place (RiP) Convolution: A Simple, General, and Near-Optimal Strategy for Memory-Efficient CNN Inference — arXiv:2610.00586
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.00586
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T19:16:24.194Z
- SHA-256 of the archived text
5e2747cbe7e4e8073b0ab01f0a6939c60fbde9c24afd1ac880b9aec6f1a50917- Stable URL of this page
https://metraipolis.com/documento/dc_069be2cb0012
How to cite
Right In-Place (RiP) Convolution: A Simple, General, and Near-Optimal Strategy for Memory-Efficient CNN Inference — arXiv:2610.00586 — arXiv (en), https://arxiv.org/abs/2610.00586 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_069be2cb0012 (Metraipolis archive dc_069be2cb0012, sha256 5e2747cbe7e4e8073b0ab01f0a6939c60fbde9c24afd1ac880b9aec6f1a50917)
BibTeX
@misc{dc069be2cb0012,
title = {Right In-Place (RiP) Convolution: A Simple, General, and Near-Optimal Strategy for Memory-Efficient CNN Inference — arXiv:2610.00586},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.00586},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_069be2cb0012; sha256 5e2747cbe7e4e8073b0ab01f0a6939c60fbde9c24afd1ac880b9aec6f1a50917},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.00586. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Activation memory, not compute, limits CNN inference on constrained hardware such as microcontrollers. Direct in-place convolution removes the dual-buffer cost, but the memory-optimal formulation of Gural and Murmann assumes valid padding, unit stride, unit dilation, and odd square kernels, and needs a non-sequential traversal costing $2\times$ inference time in transposes. We identify two regimes in which their published closed form does not hold: (1) an under-allocation of exactly $(k-1)C_{in} \bmod (C_{out}-C_{in})$ scalars, active on every convolutional layer of their own deployed network and manifesting as a silent corruption of still-live input; (2) an unbounded overestimate, up to $2{,}432\times$, once the critical leg leaves the output grid. We correct both and generalize to arbitrary stride, dilation, padding, and rectangular kernels. We then propose Right In-Place (RiP) convolution, a bit-identical operation in which every layer reads its input right-aligned in a shared workspace and writes its output left-aligned from index zero. The debt is piecewise affine in the output pixel index, so evaluating its breakpoints in $O(1)$ yields the minimum safe gap without enumerating the output grid, with row-major access preserved. Across $10{,}000$ random layers RiP produced no corruption, and across 84 convolutional layers from 25 architectures it matches the herringbone workspace exactly on 58 and within 5% on 81, using 24.8% less memory than dual buffering on average. Written into TinyEngine's kernels and deployed to a Raspberry Pi Pico 1 and Pico 2, it cuts peak activation memory across eleven MCUNet models by 12.5 to 33.3% at unchanged cycle counts and bit-identical outputs, raising the number of models that fit the Pico 1's 256 KB SRAM from six to nine.
Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-10-02 in cs.CV).
Categorie arXiv: Machine Learning (cs.LG) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.00586 — 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 2026-10-06T17:37:36.925Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).