Publications

Reformulating the direct convolution for high-performance deep learning inference on ARM processors

TYPE OF PUBLICATION

AUTHORS

Sergio Barrachina, Adrián Castelló, Manuel F. Dolz, Tze Meng Low, Héctor Martínez, Enrique S. Quintana-Ortí, Upasana Sridhar, Andrés E. Tomás,

PUBLISHER

Journal of Systems Architecture

YEAR OF PUBLICATION

2023

PLACE OF PUBLICATION

ISSN

ISSN 1383-7621

DOI

https://doi.org/10.1016/j.sysarc.2022.102806

CITACION

Sergio Barrachina, Adrián Castelló, Manuel F. Dolz, Tze Meng Low, Héctor Martínez, Enrique S. Quintana-Ortí, Upasana Sridhar, Andrés E. Tomás, Reformulating the direct convolution for high-performance deep learning inference on ARM processors, Journal of Systems Architecture, Volume 135, 2023, 102806, ISSN 1383-7621, https://doi.org/10.1016/j.sysarc.2022.102806. (https://www.sciencedirect.com/science/article/pii/S1383762122002910) Abstract: We present two high-performance implementations of the convolution operator via the direct algorithm that outperform the so-called lowering approach based on the im2col transform plus the gemm kernel on an ARMv8-based processor. One of our methods presents the additional advantage of zero-memory overhead while the other employs an additional yet rather moderate workspace, substantially smaller than that required by the im2col+gemm solution. In contrast with a previous implementation of a similar zero-memory overhead direct convolution, this work exhibits the key advantage of preserving the conventional NHWC data layout for the input/output activations of the convolution layers. Keywords: Convolution; Direct algorithm; Deep learning; High performance; ARMv8 architecture

LINK TO THE REPOSITORY

https://www.sciencedirect.com/science/article/pii/S1383762122002910

LINK TO THE PUBLICATION

https://www.sciencedirect.com/science/article/pii/S1383762122002910