arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.21792cs.AR
收藏

卷积可预测卸载至加速器:形式化与优化

Convolutions Predictable Offloading to an Accelerator: Formalization and Optimization

Université de Toulouse – IRIT(图卢兹大学 - IRIT);ONERA(法国航空与航天研究院);CS Group(CS集团)

Benjamin Husson, Mohammed Belcaïd, Thomas Carle, Claire Pagetti

AI总结:

研究通过形式化卷积卸载步骤,优化加速器执行效率,提升实时性要求下的卷积神经网络计算性能。

摘要:

卷积神经网络(CNNs)需要大量的乘加(MAC)操作。为了满足实时性要求,它们通常需要在由芯片内存和处理单元组成的专用加速器上执行。然而,芯片内存往往不足以存储计算CNN层所需的所有数据。因此,计算必须分步骤卸载。我们形式化了这些步骤序列,并将其应用于最先进的卷积分解。为了在持续时间内找到最优策略,我们将问题编码为一组约束。基于Python的模拟器允许深入分析计算策略。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑