Influence of Parallelism in Vector-Multiplication Units on Correlation Power Analysis
Manuel Brosch, Matthias Probst, Stefan Kögler, Georg Sigl · 2026
Read Influence of Parallelism in Vector-Multiplication Units on Correlation Power Analysis by Manuel Brosch, Matthias Probst, Stefan Kögler, Georg Sigl (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-05828.txt.
Description
Manuel Brosch and 3 co-authors analyze in 2026 how parallel computation inside neural network accelerators affects side-channel attacks that read power consumption. Edge devices running neural networks are physically accessible, so attackers can measure power draw to steal model secrets. Accelerators speed inference by running many multiply-and-accumulate operations at once.
The study focuses on neurons in one fully connected layer processing the same input in parallel. It evaluates theoretically how concurrent work changes total power draw and attack success. Equations derived from that behavior describe how the match falls as parallelism rises. A vector-multiplication unit built on an FPGA validated the rules, giving hardware designers a quantitative handle on a real confidentiality trade-off. More work done at the same time blurs the signal an attacker needs, which is good news for device makers.
Catalog record
| CREATOR | Manuel Brosch, Matthias Probst, Stefan Kögler, Georg Sigl |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-05828