Convolutions Predictable Offloading to an Accelerator: Formalization and Optimization
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914414391721984 |
|---|---|
| author | Husson, Benjamin Belcaïd, Mohammed Carle, Thomas Pagetti, Claire |
| author_facet | Husson, Benjamin Belcaïd, Mohammed Carle, Thomas Pagetti, Claire |
| contents | Convolutional neural networks (CNNs) require a large number of multiply-accumulate (MAC) operations. To meet real-time constraints, they often need to be executed on specialized accelerators composed of an on-chip memory and a processing unit. However, the on-chip memory is often insufficient to store all the data required to compute a CNN layer. Thus, the computation must be performed in several offloading steps. We formalise such sequences of steps and apply our formalism to a state of the art decomposition of convolutions. In order to find optimal strategies in terms of duration, we encode the problem with a set of constraints. A Python-based simulator allows to analyse in-depth computed strategies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_21792 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Convolutions Predictable Offloading to an Accelerator: Formalization and Optimization Husson, Benjamin Belcaïd, Mohammed Carle, Thomas Pagetti, Claire Hardware Architecture Convolutional neural networks (CNNs) require a large number of multiply-accumulate (MAC) operations. To meet real-time constraints, they often need to be executed on specialized accelerators composed of an on-chip memory and a processing unit. However, the on-chip memory is often insufficient to store all the data required to compute a CNN layer. Thus, the computation must be performed in several offloading steps. We formalise such sequences of steps and apply our formalism to a state of the art decomposition of convolutions. In order to find optimal strategies in terms of duration, we encode the problem with a set of constraints. A Python-based simulator allows to analyse in-depth computed strategies. |
| title | Convolutions Predictable Offloading to an Accelerator: Formalization and Optimization |
| topic | Hardware Architecture |
| url | https://arxiv.org/abs/2603.21792 |